bibtex_url
stringlengths 41
50
| bibtext
stringlengths 693
2.88k
| abstract
stringlengths 0
2k
| authors
listlengths 1
45
| title
stringlengths 21
206
| id
stringlengths 7
16
| type
stringclasses 2
values | arxiv_id
stringlengths 9
12
⌀ |
|---|---|---|---|---|---|---|---|
https://aclanthology.org/2024.findings-acl.468.bib | @inproceedings{du-etal-2024-uor,
title = "{UOR}: Universal Backdoor Attacks on Pre-trained Language Models",
author = "Du, Wei and
Li, Peixuan and
Zhao, Haodong and
Ju, Tianjie and
Ren, Ge and
Liu, Gongshen",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.468",
pages = "7865--7877",
abstract = "Task-agnostic and transferable backdoors implanted in pre-trained language models (PLMs) pose a severe security threat as they can be inherited to any downstream task. However, existing methods rely on manual selection of triggers and backdoor representations, hindering their effectiveness and universality across different PLMs or usage paradigms. In this paper, we propose a new backdoor attack method called UOR, which overcomes these limitations by turning manual selection into automatic optimization. Specifically, we design poisoned supervised contrastive learning, which can automatically learn more uniform and universal backdoor representations. This allows for more even coverage of the output space, thus hitting more labels in downstream tasks after fine-tuning. Furthermore, we utilize gradient search to select appropriate trigger words that can be adapted to different PLMs and vocabularies. Experiments show that UOR achieves better attack performance on various text classification tasks compared to manual methods. Moreover, we test on PLMs with different architectures, usage paradigms, and more challenging tasks, achieving higher scores for universality.",
}
| Task-agnostic and transferable backdoors implanted in pre-trained language models (PLMs) pose a severe security threat as they can be inherited to any downstream task. However, existing methods rely on manual selection of triggers and backdoor representations, hindering their effectiveness and universality across different PLMs or usage paradigms. In this paper, we propose a new backdoor attack method called UOR, which overcomes these limitations by turning manual selection into automatic optimization. Specifically, we design poisoned supervised contrastive learning, which can automatically learn more uniform and universal backdoor representations. This allows for more even coverage of the output space, thus hitting more labels in downstream tasks after fine-tuning. Furthermore, we utilize gradient search to select appropriate trigger words that can be adapted to different PLMs and vocabularies. Experiments show that UOR achieves better attack performance on various text classification tasks compared to manual methods. Moreover, we test on PLMs with different architectures, usage paradigms, and more challenging tasks, achieving higher scores for universality. | [
"Du, Wei",
"Li, Peixuan",
"Zhao, Haodong",
"Ju, Tianjie",
"Ren, Ge",
"Liu, Gongshen"
]
| {UOR}: Universal Backdoor Attacks on Pre-trained Language Models | findings-acl.468 | Poster | 2305.14910v3 |
https://aclanthology.org/2024.findings-acl.469.bib | @inproceedings{haller-etal-2024-language,
title = "Language models emulate certain cognitive profiles: An investigation of how predictability measures interact with individual differences",
author = {Haller, Patrick and
Bolliger, Lena and
J{\"a}ger, Lena},
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.469",
pages = "7878--7892",
abstract = "To date, most investigations on surprisal and entropy effects in reading have been conducted on the group level, disregarding individual differences. In this work, we revisit the predictive power (PP) of different LMs{'} surprisal and entropy measures on data of human reading times as a measure of processing effort by incorporating information of language users{'} cognitive capacities. To do so, we assess the PP of surprisal and entropy estimated from generative language models (LMs) on reading data obtained from individuals who also completed a wide range of psychometric tests.Specifically, we investigate if modulating surprisal and entropy relative to cognitive scores increases prediction accuracy of reading times, and we examine whether LMs exhibit systematic biases in the prediction of reading times for cognitively high- or low-performing groups, revealing what type of psycholinguistic subjects a given LM emulates.Our study finds that in most cases, incorporating cognitive capacities increases predictive power of surprisal and entropy on reading times, and that generally, high performance in the psychometric tests is associated with lower sensitivity to predictability effects. Finally, our results suggest that the analyzed LMs emulate readers with lower verbal intelligence, suggesting that for a given target group (i.e., individuals with high verbal intelligence), these LMs provide less accurate predictability effect estimates.",
}
| To date, most investigations on surprisal and entropy effects in reading have been conducted on the group level, disregarding individual differences. In this work, we revisit the predictive power (PP) of different LMs{'} surprisal and entropy measures on data of human reading times as a measure of processing effort by incorporating information of language users{'} cognitive capacities. To do so, we assess the PP of surprisal and entropy estimated from generative language models (LMs) on reading data obtained from individuals who also completed a wide range of psychometric tests.Specifically, we investigate if modulating surprisal and entropy relative to cognitive scores increases prediction accuracy of reading times, and we examine whether LMs exhibit systematic biases in the prediction of reading times for cognitively high- or low-performing groups, revealing what type of psycholinguistic subjects a given LM emulates.Our study finds that in most cases, incorporating cognitive capacities increases predictive power of surprisal and entropy on reading times, and that generally, high performance in the psychometric tests is associated with lower sensitivity to predictability effects. Finally, our results suggest that the analyzed LMs emulate readers with lower verbal intelligence, suggesting that for a given target group (i.e., individuals with high verbal intelligence), these LMs provide less accurate predictability effect estimates. | [
"Haller, Patrick",
"Bolliger, Lena",
"J{\\\"a}ger, Lena"
]
| Language models emulate certain cognitive profiles: An investigation of how predictability measures interact with individual differences | findings-acl.469 | Poster | 2406.04988v2 |
https://aclanthology.org/2024.findings-acl.470.bib | @inproceedings{cai-oconnor-2024-state,
title = "The State of Relation Extraction Data Quality: Is Bigger Always Better?",
author = "Cai, Erica and
O{'}Connor, Brendan",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.470",
pages = "7893--7906",
abstract = "Relation extraction (RE) extracts structured tuples of relationships (e.g. friend, enemy) between entities (e.g. Sherlock Holmes, John Watson) from text, with exciting potential applications. Hundreds of RE papers have been published in recent years; do their evaluation practices inform these goals? We review recent surveys and a sample of recent RE methods papers, compiling 38 datasets currently being used. Unfortunately, many have frequent label errors, and ones with known problems continue to be used. Many datasets focus on producing labels for a large number of relation types, often through error-prone annotation methods (e.g. distant supervision or crowdsourcing), and many recent papers rely exclusively on such datasets. We draw attention to a promising alternative: datasets with a \textit{small} number of relations, often in specific domains like chemistry, finance, or biomedicine, where it is possible to obtain high quality expert annotations; such data can more realistically evaluate RE performance. The research community should consider more often using such resources.",
}
| Relation extraction (RE) extracts structured tuples of relationships (e.g. friend, enemy) between entities (e.g. Sherlock Holmes, John Watson) from text, with exciting potential applications. Hundreds of RE papers have been published in recent years; do their evaluation practices inform these goals? We review recent surveys and a sample of recent RE methods papers, compiling 38 datasets currently being used. Unfortunately, many have frequent label errors, and ones with known problems continue to be used. Many datasets focus on producing labels for a large number of relation types, often through error-prone annotation methods (e.g. distant supervision or crowdsourcing), and many recent papers rely exclusively on such datasets. We draw attention to a promising alternative: datasets with a \textit{small} number of relations, often in specific domains like chemistry, finance, or biomedicine, where it is possible to obtain high quality expert annotations; such data can more realistically evaluate RE performance. The research community should consider more often using such resources. | [
"Cai, Erica",
"O{'}Connor, Brendan"
]
| The State of Relation Extraction Data Quality: Is Bigger Always Better? | findings-acl.470 | Poster | 2402.03375v3 |
https://aclanthology.org/2024.findings-acl.471.bib | @inproceedings{zhang-etal-2024-naturalcodebench,
title = "{N}atural{C}ode{B}ench: Examining Coding Performance Mismatch on {H}uman{E}val and Natural User Queries",
author = "Zhang, Shudan and
Zhao, Hanlin and
Liu, Xiao and
Zheng, Qinkai and
Qi, Zehan and
Gu, Xiaotao and
Dong, Yuxiao and
Tang, Jie",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.471",
pages = "7907--7928",
abstract = "Large language models (LLMs) have manifested strong ability to generate codes for productive activities. However, current benchmarks for code synthesis, such as HumanEval, MBPP, and DS-1000, are predominantly oriented towards introductory tasks on algorithm and data science, insufficiently satisfying challenging requirements prevalent in real-world coding. To fill this gap, we propose NaturalCodeBench (NCB), a challenging code benchmark designed to mirror the complexity and variety of scenarios in real coding tasks. NCB comprises 402 high-quality problems in Python and Java, meticulously selected from natural user queries from online coding services, covering 6 different domains. Noting the extraordinary difficulty in creating testing cases for real-world queries, we also introduce a semi-automated pipeline to enhance the efficiency of test case construction. Comparing with manual solutions, it achieves an efficiency increase of more than 4 times. Our systematic experiments on 39 LLMs find that performance gaps on NCB between models with close HumanEval scores could still be significant, indicating a lack of focus on practical code synthesis scenarios or over-specified optimization on HumanEval. On the other hand, even the best-performing GPT-4 is still far from satisfying on NCB. The evaluation toolkit and development set are available at https://github.com/THUDM/NaturalCodeBench.",
}
| Large language models (LLMs) have manifested strong ability to generate codes for productive activities. However, current benchmarks for code synthesis, such as HumanEval, MBPP, and DS-1000, are predominantly oriented towards introductory tasks on algorithm and data science, insufficiently satisfying challenging requirements prevalent in real-world coding. To fill this gap, we propose NaturalCodeBench (NCB), a challenging code benchmark designed to mirror the complexity and variety of scenarios in real coding tasks. NCB comprises 402 high-quality problems in Python and Java, meticulously selected from natural user queries from online coding services, covering 6 different domains. Noting the extraordinary difficulty in creating testing cases for real-world queries, we also introduce a semi-automated pipeline to enhance the efficiency of test case construction. Comparing with manual solutions, it achieves an efficiency increase of more than 4 times. Our systematic experiments on 39 LLMs find that performance gaps on NCB between models with close HumanEval scores could still be significant, indicating a lack of focus on practical code synthesis scenarios or over-specified optimization on HumanEval. On the other hand, even the best-performing GPT-4 is still far from satisfying on NCB. The evaluation toolkit and development set are available at https://github.com/THUDM/NaturalCodeBench. | [
"Zhang, Shudan",
"Zhao, Hanlin",
"Liu, Xiao",
"Zheng, Qinkai",
"Qi, Zehan",
"Gu, Xiaotao",
"Dong, Yuxiao",
"Tang, Jie"
]
| {N}atural{C}ode{B}ench: Examining Coding Performance Mismatch on {H}uman{E}val and Natural User Queries | findings-acl.471 | Poster | 2405.04520v1 |
https://aclanthology.org/2024.findings-acl.472.bib | @inproceedings{yuan-etal-2024-llmcrit,
title = "{LLMC}rit: Teaching Large Language Models to Use Criteria",
author = "Yuan, Weizhe and
Liu, Pengfei and
Gall{\'e}, Matthias",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.472",
pages = "7929--7960",
abstract = "Humans follow criteria when they execute tasks, and these criteria are directly used to assess the quality of task completion. Therefore, having models learn to use criteria to provide feedback can help humans or models to perform tasks better. However, current research in this area tends to consider only a limited number of criteria, or only a limited number of quality assessment aspects. To fill this gap, we propose a general framework that enables large language models (LLMs) to use comprehensive criteria for a task in delivering natural language feedback on task execution. In particular, we present a model-in-the-loop framework that semi-automatically derives criteria from collected guidelines for different writing tasks and constructs in-context demonstrations for each criterion. We choose three tasks from real-world scenarios to operationalize this idea: paper introduction writing, Python code writing, and Reddit post writing, and evaluate our feedback generation framework using different LLMs. The results reveal the fine-grained effects of adding criteria and demonstrations and provide valuable guidance on how to teach LLMs to use criteria more effectively.",
}
| Humans follow criteria when they execute tasks, and these criteria are directly used to assess the quality of task completion. Therefore, having models learn to use criteria to provide feedback can help humans or models to perform tasks better. However, current research in this area tends to consider only a limited number of criteria, or only a limited number of quality assessment aspects. To fill this gap, we propose a general framework that enables large language models (LLMs) to use comprehensive criteria for a task in delivering natural language feedback on task execution. In particular, we present a model-in-the-loop framework that semi-automatically derives criteria from collected guidelines for different writing tasks and constructs in-context demonstrations for each criterion. We choose three tasks from real-world scenarios to operationalize this idea: paper introduction writing, Python code writing, and Reddit post writing, and evaluate our feedback generation framework using different LLMs. The results reveal the fine-grained effects of adding criteria and demonstrations and provide valuable guidance on how to teach LLMs to use criteria more effectively. | [
"Yuan, Weizhe",
"Liu, Pengfei",
"Gall{\\'e}, Matthias"
]
| {LLMC}rit: Teaching Large Language Models to Use Criteria | findings-acl.472 | Poster | 2403.01069v2 |
https://aclanthology.org/2024.findings-acl.473.bib | @inproceedings{ranaldi-etal-2024-empowering,
title = "Empowering cross-lingual abilities of instruction-tuned large language models by translation-following demonstrations",
author = "Ranaldi, Leonardo and
Pucci, Giulia and
Freitas, Andre",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.473",
pages = "7961--7973",
abstract = "The language ability of Large Language Models (LLMs) is often unbalanced towards English because of the imbalance in the distribution of the pre-training data. This disparity is demanded in further fine-tuning and affecting the cross-lingual abilities of LLMs. In this paper, we propose to empower Instruction-tuned LLMs (It-LLMs) in languages other than English by building semantic alignment between them. Hence, we propose \textit{CrossAlpaca}, an It-LLM with cross-lingual Instruction-following and Translation-following demonstrations to improve semantic alignment between languages. We validate our approach on the multilingual Question Answering (QA) benchmarks XQUAD and MLQA and adapted versions of MMLU and BBH.Our models, tested over six different languages, outperform the It-LLMs tuned on monolingual data. The final results show that instruction tuning on non-English data is not enough and that semantic alignment can be further improved by Translation-following demonstrations.",
}
| The language ability of Large Language Models (LLMs) is often unbalanced towards English because of the imbalance in the distribution of the pre-training data. This disparity is demanded in further fine-tuning and affecting the cross-lingual abilities of LLMs. In this paper, we propose to empower Instruction-tuned LLMs (It-LLMs) in languages other than English by building semantic alignment between them. Hence, we propose \textit{CrossAlpaca}, an It-LLM with cross-lingual Instruction-following and Translation-following demonstrations to improve semantic alignment between languages. We validate our approach on the multilingual Question Answering (QA) benchmarks XQUAD and MLQA and adapted versions of MMLU and BBH.Our models, tested over six different languages, outperform the It-LLMs tuned on monolingual data. The final results show that instruction tuning on non-English data is not enough and that semantic alignment can be further improved by Translation-following demonstrations. | [
"Ranaldi, Leonardo",
"Pucci, Giulia",
"Freitas, Andre"
]
| Empowering cross-lingual abilities of instruction-tuned large language models by translation-following demonstrations | findings-acl.473 | Poster | 2308.14186v1 |
https://aclanthology.org/2024.findings-acl.474.bib | @inproceedings{kumar-etal-2024-ranking,
title = "Ranking Entities along Conceptual Space Dimensions with {LLM}s: An Analysis of Fine-Tuning Strategies",
author = "Kumar, Nitesh and
Chatterjee, Usashi and
Schockaert, Steven",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.474",
pages = "7974--7989",
abstract = "Conceptual spaces represent entities in terms of their primitive semantic features. Such representations are highly valuable but they are notoriously difficult to learn, especially when it comes to modelling perceptual and subjective features. Distilling conceptual spaces from Large Language Models (LLMs) has recently emerged as a promising strategy, but existing work has been limited to probing pre-trained LLMs using relatively simple zero-shot strategies. We focus in particular on the task of ranking entities according to a given conceptual space dimension. Unfortunately, we cannot directly fine-tune LLMs on this task, because ground truth rankings for conceptual space dimensions are rare. We therefore use more readily available features as training data and analyse whether the ranking capabilities of the resulting models transfer to perceptual and subjective features. We find that this is indeed the case, to some extent, but having at least some perceptual and subjective features in the training data seems essential for achieving the best results.",
}
| Conceptual spaces represent entities in terms of their primitive semantic features. Such representations are highly valuable but they are notoriously difficult to learn, especially when it comes to modelling perceptual and subjective features. Distilling conceptual spaces from Large Language Models (LLMs) has recently emerged as a promising strategy, but existing work has been limited to probing pre-trained LLMs using relatively simple zero-shot strategies. We focus in particular on the task of ranking entities according to a given conceptual space dimension. Unfortunately, we cannot directly fine-tune LLMs on this task, because ground truth rankings for conceptual space dimensions are rare. We therefore use more readily available features as training data and analyse whether the ranking capabilities of the resulting models transfer to perceptual and subjective features. We find that this is indeed the case, to some extent, but having at least some perceptual and subjective features in the training data seems essential for achieving the best results. | [
"Kumar, Nitesh",
"Chatterjee, Usashi",
"Schockaert, Steven"
]
| Ranking Entities along Conceptual Space Dimensions with {LLM}s: An Analysis of Fine-Tuning Strategies | findings-acl.474 | Poster | 2402.15337v2 |
https://aclanthology.org/2024.findings-acl.475.bib | @inproceedings{gao-etal-2024-efficient,
title = "Efficient $k$-Nearest-Neighbor Machine Translation with Dynamic Retrieval",
author = "Gao, Yan and
Cao, Zhiwei and
Miao, Zhongjian and
Yang, Baosong and
Liu, Shiyu and
Zhang, Min and
Su, Jinsong",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.475",
pages = "7990--8001",
abstract = "To achieve non-parametric NMT domain adaptation, $k$-Nearest-Neighbor Machine Translation ($k$NN-MT) constructs an external datastore to store domain-specific translation knowledge, which derives a $k$NN distribution to interpolate the prediction distribution of the NMT model via a linear interpolation coefficient $\lambda$. Despite its success, $k$NN retrieval at each timestep leads to substantial time overhead. To address this issue, dominant studies resort to $k$NN-MT with adaptive retrieval ($k$NN-MT-AR), which dynamically estimates $\lambda$ and skips $k$NN retrieval if $\lambda$ is less than a fixed threshold. Unfortunately, $k$NN-MT-AR does not yield satisfactory results. In this paper, we first conduct a preliminary study to reveal two key limitations of $k$NN-MT-AR: 1) the optimization gap leads to inaccurate estimation of $\lambda$ for determining $k$NN retrieval skipping, and 2) using a fixed threshold fails to accommodate the dynamic demands for $k$NN retrieval at different timesteps. To mitigate these limitations, we then propose $k$NN-MT with dynamic retrieval ($k$NN-MT-DR) that significantly extends vanilla $k$NN-MT in two aspects. Firstly, we equip $k$NN-MT with a MLP-based classifier for determining whether to skip $k$NN retrieval at each timestep. Particularly, we explore several carefully-designed scalar features to fully exert the potential of the classifier. Secondly, we propose a timestep-aware threshold adjustment method to dynamically generate the threshold, which further improves the efficiency of our model. Experimental results on the widely-used datasets demonstrate the effectiveness and generality of our model.",
}
| To achieve non-parametric NMT domain adaptation, $k$-Nearest-Neighbor Machine Translation ($k$NN-MT) constructs an external datastore to store domain-specific translation knowledge, which derives a $k$NN distribution to interpolate the prediction distribution of the NMT model via a linear interpolation coefficient $\lambda$. Despite its success, $k$NN retrieval at each timestep leads to substantial time overhead. To address this issue, dominant studies resort to $k$NN-MT with adaptive retrieval ($k$NN-MT-AR), which dynamically estimates $\lambda$ and skips $k$NN retrieval if $\lambda$ is less than a fixed threshold. Unfortunately, $k$NN-MT-AR does not yield satisfactory results. In this paper, we first conduct a preliminary study to reveal two key limitations of $k$NN-MT-AR: 1) the optimization gap leads to inaccurate estimation of $\lambda$ for determining $k$NN retrieval skipping, and 2) using a fixed threshold fails to accommodate the dynamic demands for $k$NN retrieval at different timesteps. To mitigate these limitations, we then propose $k$NN-MT with dynamic retrieval ($k$NN-MT-DR) that significantly extends vanilla $k$NN-MT in two aspects. Firstly, we equip $k$NN-MT with a MLP-based classifier for determining whether to skip $k$NN retrieval at each timestep. Particularly, we explore several carefully-designed scalar features to fully exert the potential of the classifier. Secondly, we propose a timestep-aware threshold adjustment method to dynamically generate the threshold, which further improves the efficiency of our model. Experimental results on the widely-used datasets demonstrate the effectiveness and generality of our model. | [
"Gao, Yan",
"Cao, Zhiwei",
"Miao, Zhongjian",
"Yang, Baosong",
"Liu, Shiyu",
"Zhang, Min",
"Su, Jinsong"
]
| Efficient $k$-Nearest-Neighbor Machine Translation with Dynamic Retrieval | findings-acl.475 | Poster | 2406.06073v1 |
https://aclanthology.org/2024.findings-acl.476.bib | @inproceedings{courtois-etal-2024-symmetric,
title = "Symmetric Dot-Product Attention for Efficient Training of {BERT} Language Models",
author = "Courtois, Martin and
Ostendorff, Malte and
Hennig, Leonhard and
Rehm, Georg",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.476",
pages = "8002--8011",
abstract = "Initially introduced as a machine translation model, the Transformer architecture has now become the foundation for modern deep learning architecture, with applications in a wide range of fields, from computer vision to natural language processing. Nowadays, to tackle increasingly more complex tasks, Transformer-based models are stretched to enormous sizes, requiring increasingly larger training datasets, and unsustainable amount of compute resources. The ubiquitous nature of the Transformer and its core component, the attention mechanism, are thus prime targets for efficiency research.In this work, we propose an alternative compatibility function for the self-attention mechanism introduced by the Transformer architecture. This compatibility function exploits an overlap in the learned representation of the traditional scaled dot-product attention, leading to a symmetric with pairwise coefficient dot-product attention. When applied to the pre-training of BERT-like models, this new symmetric attention mechanism reaches a score of 79.36 on the GLUE benchmark against 78.74 for the traditional implementation, leads to a reduction of 6{\%} in the number of trainable parameters, and reduces the number of training steps required before convergence by half.",
}
| Initially introduced as a machine translation model, the Transformer architecture has now become the foundation for modern deep learning architecture, with applications in a wide range of fields, from computer vision to natural language processing. Nowadays, to tackle increasingly more complex tasks, Transformer-based models are stretched to enormous sizes, requiring increasingly larger training datasets, and unsustainable amount of compute resources. The ubiquitous nature of the Transformer and its core component, the attention mechanism, are thus prime targets for efficiency research.In this work, we propose an alternative compatibility function for the self-attention mechanism introduced by the Transformer architecture. This compatibility function exploits an overlap in the learned representation of the traditional scaled dot-product attention, leading to a symmetric with pairwise coefficient dot-product attention. When applied to the pre-training of BERT-like models, this new symmetric attention mechanism reaches a score of 79.36 on the GLUE benchmark against 78.74 for the traditional implementation, leads to a reduction of 6{\%} in the number of trainable parameters, and reduces the number of training steps required before convergence by half. | [
"Courtois, Martin",
"Ostendorff, Malte",
"Hennig, Leonhard",
"Rehm, Georg"
]
| Symmetric Dot-Product Attention for Efficient Training of {BERT} Language Models | findings-acl.476 | Poster | 2406.06366v2 |
https://aclanthology.org/2024.findings-acl.477.bib | @inproceedings{wu-etal-2024-synthesizing,
title = "Synthesizing Conversations from Unlabeled Documents using Automatic Response Segmentation",
author = "Wu, Fanyou and
Xu, Weijie and
Reddy, Chandan and
Sengamedu, Srinivasan",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.477",
pages = "8012--8026",
abstract = "In this study, we tackle the challenge of inadequate and costly training data that has hindered the development of conversational question answering (ConvQA) systems. Enterprises have a large corpus of diverse internal documents. Instead of relying on a searching engine, a more compelling approach for people to comprehend these documents is to create a dialogue system. In this paper, we propose a robust dialog synthesising method. We learn the segmentation of data for the dialog task instead of using segmenting at sentence boundaries. The synthetic dataset generated by our proposed method achieves superior quality when compared to WikiDialog, as assessed through machine and human evaluations. By employing our inpainted data for ConvQA retrieval system pre-training, we observed a notable improvement in performance across OR-QuAC benchmarks.",
}
| In this study, we tackle the challenge of inadequate and costly training data that has hindered the development of conversational question answering (ConvQA) systems. Enterprises have a large corpus of diverse internal documents. Instead of relying on a searching engine, a more compelling approach for people to comprehend these documents is to create a dialogue system. In this paper, we propose a robust dialog synthesising method. We learn the segmentation of data for the dialog task instead of using segmenting at sentence boundaries. The synthetic dataset generated by our proposed method achieves superior quality when compared to WikiDialog, as assessed through machine and human evaluations. By employing our inpainted data for ConvQA retrieval system pre-training, we observed a notable improvement in performance across OR-QuAC benchmarks. | [
"Wu, Fanyou",
"Xu, Weijie",
"Reddy, Ch",
"an",
"Sengamedu, Srinivasan"
]
| Synthesizing Conversations from Unlabeled Documents using Automatic Response Segmentation | findings-acl.477 | Poster | 2406.03703v1 |
https://aclanthology.org/2024.findings-acl.478.bib | @inproceedings{fonseca-cohen-2024-large,
title = "Can Large Language Models Follow Concept Annotation Guidelines? A Case Study on Scientific and Financial Domains",
author = "Fonseca, Marcio and
Cohen, Shay",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.478",
pages = "8027--8042",
abstract = "Although large language models (LLMs) exhibit remarkable capacity to leverage in-context demonstrations, it is still unclear to what extent they can learn new facts or concept definitions via prompts. To address this question, we examine the capacity of instruction-tuned LLMs to follow in-context concept annotation guidelines for zero-shot sentence labeling tasks. We design guidelines that present different types of factual and counterfactual concept definitions, which are used as prompts for zero-shot sentence classification tasks. Our results show that although concept definitions consistently help in task performance, only the larger models (with 70B parameters or more) have limited ability to work under counterfactual contexts. Importantly, only proprietary models such as GPT-3.5 can recognize nonsensical guidelines, which we hypothesize is due to more sophisticated alignment methods. Finally, we find that Falcon-180B-chat is outperformed by Llama-2-70B-chat is most cases, which indicates that increasing model scale does not guarantee better adherence to guidelines. Altogether, our simple evaluation method reveals significant gaps in concept understanding between the most capable open-source language models and the leading proprietary APIs.",
}
| Although large language models (LLMs) exhibit remarkable capacity to leverage in-context demonstrations, it is still unclear to what extent they can learn new facts or concept definitions via prompts. To address this question, we examine the capacity of instruction-tuned LLMs to follow in-context concept annotation guidelines for zero-shot sentence labeling tasks. We design guidelines that present different types of factual and counterfactual concept definitions, which are used as prompts for zero-shot sentence classification tasks. Our results show that although concept definitions consistently help in task performance, only the larger models (with 70B parameters or more) have limited ability to work under counterfactual contexts. Importantly, only proprietary models such as GPT-3.5 can recognize nonsensical guidelines, which we hypothesize is due to more sophisticated alignment methods. Finally, we find that Falcon-180B-chat is outperformed by Llama-2-70B-chat is most cases, which indicates that increasing model scale does not guarantee better adherence to guidelines. Altogether, our simple evaluation method reveals significant gaps in concept understanding between the most capable open-source language models and the leading proprietary APIs. | [
"Fonseca, Marcio",
"Cohen, Shay"
]
| Can Large Language Models Follow Concept Annotation Guidelines? A Case Study on Scientific and Financial Domains | findings-acl.478 | Poster | 2311.08704v2 |
https://aclanthology.org/2024.findings-acl.479.bib | @inproceedings{sarhangzadeh-watanabe-2024-alignment,
title = "Alignment-Based Decoding Policy for Low-Latency and Anticipation-Free Neural {J}apanese Input Method Editors",
author = "Sarhangzadeh, Armin and
Watanabe, Taro",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.479",
pages = "8043--8054",
abstract = "Japanese input method editors (IMEs) are essential tools for inputting Japanese text using a limited set of characters such as the kana syllabary. However, despite their importance, the potential of newer attention-based encoder-decoder neural networks, such as Transformer, has not yet been fully explored for IMEs due to their high computational cost and low-quality intermediate output in simultaneous settings, leading to high latencies. In this work, we propose a simple decoding policy to enable the use of attention-based encoder-decoder networks for simultaneous kana-kanji conversion in the context of Japanese IMEs inspired by simultaneous machine translation (SimulMT). We demonstrate that simply decoding by explicitly considering the word boundaries achieves a fairly strong quality-latency trade-off, as it can be seen as equivalent to performing decoding on aligned prefixes and thus achieving an incremental anticipation-free conversion. We further show how such a policy can be applied in practice to achieve high-quality conversions with minimal computational overhead. Our experiments show that our approach can achieve a noticeably better quality-latency trade-off compared to the baselines, while also being a more practical approach due to its ability to directly handle streaming input. Our code is available at https://anonymous.4open.science/r/transformer{\_}ime-D327.",
}
| Japanese input method editors (IMEs) are essential tools for inputting Japanese text using a limited set of characters such as the kana syllabary. However, despite their importance, the potential of newer attention-based encoder-decoder neural networks, such as Transformer, has not yet been fully explored for IMEs due to their high computational cost and low-quality intermediate output in simultaneous settings, leading to high latencies. In this work, we propose a simple decoding policy to enable the use of attention-based encoder-decoder networks for simultaneous kana-kanji conversion in the context of Japanese IMEs inspired by simultaneous machine translation (SimulMT). We demonstrate that simply decoding by explicitly considering the word boundaries achieves a fairly strong quality-latency trade-off, as it can be seen as equivalent to performing decoding on aligned prefixes and thus achieving an incremental anticipation-free conversion. We further show how such a policy can be applied in practice to achieve high-quality conversions with minimal computational overhead. Our experiments show that our approach can achieve a noticeably better quality-latency trade-off compared to the baselines, while also being a more practical approach due to its ability to directly handle streaming input. Our code is available at https://anonymous.4open.science/r/transformer{\_}ime-D327. | [
"Sarhangzadeh, Armin",
"Watanabe, Taro"
]
| Alignment-Based Decoding Policy for Low-Latency and Anticipation-Free Neural {J}apanese Input Method Editors | findings-acl.479 | Poster | 2011.06868v2 |
https://aclanthology.org/2024.findings-acl.480.bib | @inproceedings{wang-etal-2024-ecok,
title = "{EC}o{K}: Emotional Commonsense Knowledge Graph for Mining Emotional Gold",
author = "Wang, Zhunheng and
Liu, Xiaoyi and
Hu, Mengting and
Ying, Rui and
Jiang, Ming and
Wu, Jianfeng and
Xie, Yalan and
Gao, Hang and
Cheng, Renhong",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.480",
pages = "8055--8074",
abstract = "The demand for understanding and expressing emotions in the field of natural language processing is growing rapidly. Knowledge graphs, as an important form of knowledge representation, have been widely utilized in various emotion-related tasks. However, existing knowledge graphs mainly focus on the representation and reasoning of general factual knowledge, while there are still significant deficiencies in the understanding and reasoning of emotional knowledge. In this work, we construct a comprehensive and accurate emotional commonsense knowledge graph, ECoK. We integrate cutting-edge theories from multiple disciplines such as psychology, cognitive science, and linguistics, and combine techniques such as large language models and natural language processing. By mining a large amount of text, dialogue, and sentiment analysis data, we construct rich emotional knowledge and establish the knowledge generation model COMET-ECoK. Experimental results show that ECoK contains high-quality emotional reasoning knowledge, and the performance of our knowledge generation model surpasses GPT-4-Turbo, which can help downstream tasks better understand and reason about emotions. Our data and code is available from https://github.com/ZornWang/ECoK.",
}
| The demand for understanding and expressing emotions in the field of natural language processing is growing rapidly. Knowledge graphs, as an important form of knowledge representation, have been widely utilized in various emotion-related tasks. However, existing knowledge graphs mainly focus on the representation and reasoning of general factual knowledge, while there are still significant deficiencies in the understanding and reasoning of emotional knowledge. In this work, we construct a comprehensive and accurate emotional commonsense knowledge graph, ECoK. We integrate cutting-edge theories from multiple disciplines such as psychology, cognitive science, and linguistics, and combine techniques such as large language models and natural language processing. By mining a large amount of text, dialogue, and sentiment analysis data, we construct rich emotional knowledge and establish the knowledge generation model COMET-ECoK. Experimental results show that ECoK contains high-quality emotional reasoning knowledge, and the performance of our knowledge generation model surpasses GPT-4-Turbo, which can help downstream tasks better understand and reason about emotions. Our data and code is available from https://github.com/ZornWang/ECoK. | [
"Wang, Zhunheng",
"Liu, Xiaoyi",
"Hu, Mengting",
"Ying, Rui",
"Jiang, Ming",
"Wu, Jianfeng",
"Xie, Yalan",
"Gao, Hang",
"Cheng, Renhong"
]
| {EC}o{K}: Emotional Commonsense Knowledge Graph for Mining Emotional Gold | findings-acl.480 | Poster | 1909.10681v2 |
https://aclanthology.org/2024.findings-acl.481.bib | @inproceedings{yao-etal-2024-deterministic,
title = "Deterministic Reversible Data Augmentation for Neural Machine Translation",
author = "Yao, Jiashu and
Huang, Heyan and
Liu, Zeming and
Guo, Yuhang",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.481",
pages = "8075--8089",
abstract = "Data augmentation is an effective way to diversify corpora in machine translation, but previous methods may introduce semantic inconsistency between original and augmented data because of irreversible operations and random subword sampling procedures. To generate both symbolically diverse and semantically consistent augmentation data, we propose Deterministic Reversible Data Augmentation (DRDA), a simple but effective data augmentation method for neural machine translation. DRDA adopts deterministic segmentations and reversible operations to generate multi-granularity subword representations and pulls them closer together with multi-view techniques. With no extra corpora or model changes required, DRDA outperforms strong baselines on several translation tasks with a clear margin (up to 4.3 BLEU gain over Transformer) and exhibits good robustness in noisy, low-resource, and cross-domain datasets.",
}
| Data augmentation is an effective way to diversify corpora in machine translation, but previous methods may introduce semantic inconsistency between original and augmented data because of irreversible operations and random subword sampling procedures. To generate both symbolically diverse and semantically consistent augmentation data, we propose Deterministic Reversible Data Augmentation (DRDA), a simple but effective data augmentation method for neural machine translation. DRDA adopts deterministic segmentations and reversible operations to generate multi-granularity subword representations and pulls them closer together with multi-view techniques. With no extra corpora or model changes required, DRDA outperforms strong baselines on several translation tasks with a clear margin (up to 4.3 BLEU gain over Transformer) and exhibits good robustness in noisy, low-resource, and cross-domain datasets. | [
"Yao, Jiashu",
"Huang, Heyan",
"Liu, Zeming",
"Guo, Yuhang"
]
| Deterministic Reversible Data Augmentation for Neural Machine Translation | findings-acl.481 | Poster | 2406.02517v1 |
https://aclanthology.org/2024.findings-acl.482.bib | @inproceedings{zhou-etal-2024-latent,
title = "Latent Learningscape Guided In-context Learning",
author = "Zhou, Anlai and
Jiang, Sunshine and
Liu, Yifei and
Wu, Yiquan and
Kuang, Kun and
Xiao, Jun",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.482",
pages = "8090--8101",
abstract = "The growing interest in leveraging large language models is driven by their exceptional imitation and reasoning capabilities. In-context learning (ICL), a streamlined method, has shown potential in boosting these models{'} performance without modifying their underlying parameters, especially when supplied with suitable demonstrations. However, existing methods mainly choose demonstrations by comparing surface-level semantic similarities (e.g., based on embedding) and fall short of identifying the most fitting ones. This paper introduces the concept of a {``}latent learningscape{''}, a more nuanced representation that describes the characteristic of the demonstrations. Building on this concept, we develop a results-driven approach to characterize the latent learningscape features of demonstrations, which then inform the creation of more effective prompts. Through comprehensive testing across datasets in arithmetic, commonsense, and symbolic reasoning tasks, our approach outperforms leading models, showing an average increase in scores by 7.4 percentage points.",
}
| The growing interest in leveraging large language models is driven by their exceptional imitation and reasoning capabilities. In-context learning (ICL), a streamlined method, has shown potential in boosting these models{'} performance without modifying their underlying parameters, especially when supplied with suitable demonstrations. However, existing methods mainly choose demonstrations by comparing surface-level semantic similarities (e.g., based on embedding) and fall short of identifying the most fitting ones. This paper introduces the concept of a {``}latent learningscape{''}, a more nuanced representation that describes the characteristic of the demonstrations. Building on this concept, we develop a results-driven approach to characterize the latent learningscape features of demonstrations, which then inform the creation of more effective prompts. Through comprehensive testing across datasets in arithmetic, commonsense, and symbolic reasoning tasks, our approach outperforms leading models, showing an average increase in scores by 7.4 percentage points. | [
"Zhou, Anlai",
"Jiang, Sunshine",
"Liu, Yifei",
"Wu, Yiquan",
"Kuang, Kun",
"Xiao, Jun"
]
| Latent Learningscape Guided In-context Learning | findings-acl.482 | Poster | 2004.01255v1 |
https://aclanthology.org/2024.findings-acl.483.bib | @inproceedings{qi-etal-2024-smr,
title = "{SMR}: State Memory Replay for Long Sequence Modeling",
author = "Qi, Biqing and
Gao, Junqi and
Zhang, Kaiyan and
Li, Dong and
Liu, Jianxing and
Wu, Ligang and
Zhou, Bowen",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.483",
pages = "8102--8116",
abstract = "Despite the promising performance of state space models (SSMs) in long sequence modeling, limitations still exist. Advanced SSMs like S5 and S6 (Mamba) in addressing non-uniform sampling, their recursive structures impede efficient SSM computation via convolution. To overcome compatibility limitations in parallel convolutional computation, this paper proposes a novel non-recursive non-uniform sample processing strategy. Theoretical analysis of SSMs through the lens of Event-Triggered Control (ETC) theory reveals the Non-Stable State (NSS) problem, where deviations from sampling point requirements lead to error transmission and accumulation, causing the divergence of the SSM{'}s hidden state. Our analysis further reveals that adjustments of input sequences with early memories can mitigate the NSS problem, achieving Sampling Step Adaptation (SSA).Building on this insight, we introduce a simple yet effective plug-and-play mechanism, State Memory Replay (SMR), which utilizes learnable memories to adjust the current state with multi-step information for generalization at sampling points different from those in the training data. This enables SSMs to stably model varying sampling points. Experiments on long-range modeling tasks in autoregressive language modeling and Long Range Arena demonstrate the general effectiveness of the SMR mechanism for a series of SSM models.",
}
| Despite the promising performance of state space models (SSMs) in long sequence modeling, limitations still exist. Advanced SSMs like S5 and S6 (Mamba) in addressing non-uniform sampling, their recursive structures impede efficient SSM computation via convolution. To overcome compatibility limitations in parallel convolutional computation, this paper proposes a novel non-recursive non-uniform sample processing strategy. Theoretical analysis of SSMs through the lens of Event-Triggered Control (ETC) theory reveals the Non-Stable State (NSS) problem, where deviations from sampling point requirements lead to error transmission and accumulation, causing the divergence of the SSM{'}s hidden state. Our analysis further reveals that adjustments of input sequences with early memories can mitigate the NSS problem, achieving Sampling Step Adaptation (SSA).Building on this insight, we introduce a simple yet effective plug-and-play mechanism, State Memory Replay (SMR), which utilizes learnable memories to adjust the current state with multi-step information for generalization at sampling points different from those in the training data. This enables SSMs to stably model varying sampling points. Experiments on long-range modeling tasks in autoregressive language modeling and Long Range Arena demonstrate the general effectiveness of the SMR mechanism for a series of SSM models. | [
"Qi, Biqing",
"Gao, Junqi",
"Zhang, Kaiyan",
"Li, Dong",
"Liu, Jianxing",
"Wu, Ligang",
"Zhou, Bowen"
]
| {SMR}: State Memory Replay for Long Sequence Modeling | findings-acl.483 | Poster | 2303.01081v1 |
https://aclanthology.org/2024.findings-acl.484.bib | @inproceedings{mishra-etal-2024-characterizing,
title = "Characterizing Large Language Models as Rationalizers of Knowledge-intensive Tasks",
author = "Mishra, Aditi and
Rahman, Sajjadur and
Mitra, Kushan and
Kim, Hannah and
Hruschka, Estevam",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.484",
pages = "8117--8139",
abstract = "Large language models (LLMs) are proficient at generating fluent text with minimal task-specific supervision. However, their ability to generate rationales for knowledge-intensive tasks (KITs) remains under-explored. Generating rationales for KIT solutions, such as commonsense multiple-choice QA, requires external knowledge to support predictions and refute alternate options. In this work, we consider the task of generating retrieval-augmented rationalization of KIT model predictions via external knowledge guidance within a few-shot setting. Surprisingly, crowd-workers preferred LLM-generated rationales over existing crowd-sourced rationales, generated in a similar knowledge-guided setting, on aspects such as factuality, sufficiency, and convincingness. However, fine-grained evaluation of such rationales highlights the need for further improvements in conciseness, novelty, and domain invariance. Additionally, through an expert-sourced study evaluating the reliability of the rationales, we demonstrate that humans{'} trust in LLM-generated rationales erodes when communicated faithfully, i.e., without taking model prediction accuracy into account. We find that even instrumenting simple guardrails can be effective for reliable rationalization.",
}
| Large language models (LLMs) are proficient at generating fluent text with minimal task-specific supervision. However, their ability to generate rationales for knowledge-intensive tasks (KITs) remains under-explored. Generating rationales for KIT solutions, such as commonsense multiple-choice QA, requires external knowledge to support predictions and refute alternate options. In this work, we consider the task of generating retrieval-augmented rationalization of KIT model predictions via external knowledge guidance within a few-shot setting. Surprisingly, crowd-workers preferred LLM-generated rationales over existing crowd-sourced rationales, generated in a similar knowledge-guided setting, on aspects such as factuality, sufficiency, and convincingness. However, fine-grained evaluation of such rationales highlights the need for further improvements in conciseness, novelty, and domain invariance. Additionally, through an expert-sourced study evaluating the reliability of the rationales, we demonstrate that humans{'} trust in LLM-generated rationales erodes when communicated faithfully, i.e., without taking model prediction accuracy into account. We find that even instrumenting simple guardrails can be effective for reliable rationalization. | [
"Mishra, Aditi",
"Rahman, Sajjadur",
"Mitra, Kushan",
"Kim, Hannah",
"Hruschka, Estevam"
]
| Characterizing Large Language Models as Rationalizers of Knowledge-intensive Tasks | findings-acl.484 | Poster | 2311.05085v2 |
https://aclanthology.org/2024.findings-acl.485.bib | @inproceedings{li-etal-2024-challenging,
title = "Challenging Large Language Models with New Tasks: A Study on their Adaptability and Robustness",
author = "Li, Chenxi and
Tian, Yuanhe and
Zerong, Zhaxi and
Song, Yan and
Xia, Fei",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.485",
pages = "8140--8162",
abstract = "Recent progress in large language models (LLMs) has marked a notable milestone in the field of artificial intelligence. The conventional evaluation of LLMs primarily relies on existing tasks and benchmarks, raising concerns about test set contamination and the genuine comprehension abilities of LLMs. To address these concerns, we propose to evaluate LLMs by designing new tasks, automatically generating evaluation datasets for the tasks, and conducting detailed error analyses to scrutinize LLMs{'} adaptability to new tasks, their sensitivity to prompt variations, and their error tendencies. We investigate the capacity of LLMs to adapt to new but simple tasks, especially when they diverge from the models{'} pre-existing knowledge. Our methodology emphasizes the creation of straightforward tasks, facilitating a precise error analysis to uncover the underlying causes of LLM failures. This strategic approach also aims to uncover effective strategies for enhancing LLM performance based on the detailed error analysis of system output.",
}
| Recent progress in large language models (LLMs) has marked a notable milestone in the field of artificial intelligence. The conventional evaluation of LLMs primarily relies on existing tasks and benchmarks, raising concerns about test set contamination and the genuine comprehension abilities of LLMs. To address these concerns, we propose to evaluate LLMs by designing new tasks, automatically generating evaluation datasets for the tasks, and conducting detailed error analyses to scrutinize LLMs{'} adaptability to new tasks, their sensitivity to prompt variations, and their error tendencies. We investigate the capacity of LLMs to adapt to new but simple tasks, especially when they diverge from the models{'} pre-existing knowledge. Our methodology emphasizes the creation of straightforward tasks, facilitating a precise error analysis to uncover the underlying causes of LLM failures. This strategic approach also aims to uncover effective strategies for enhancing LLM performance based on the detailed error analysis of system output. | [
"Li, Chenxi",
"Tian, Yuanhe",
"Zerong, Zhaxi",
"Song, Yan",
"Xia, Fei"
]
| Challenging Large Language Models with New Tasks: A Study on their Adaptability and Robustness | findings-acl.485 | Poster | 2403.16913v1 |
https://aclanthology.org/2024.findings-acl.486.bib | @inproceedings{vasilyev-etal-2024-linear,
title = "Linear Cross-Lingual Mapping of Sentence Embeddings",
author = "Vasilyev, Oleg and
Isono, Fumika and
Bohannon, John",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.486",
pages = "8163--8171",
abstract = "Semantics of a sentence is defined with much less ambiguity than semantics of a single word, and we assume that it should be better preserved by translation to another language. If multilingual sentence embeddings intend to represent sentence semantics, then the similarity between embeddings of any two sentences must be invariant with respect to translation. Based on this suggestion, we consider a simple linear cross-lingual mapping as a possible improvement of the multilingual embeddings. We also consider deviation from orthogonality conditions as a measure of deficiency of the embeddings.",
}
| Semantics of a sentence is defined with much less ambiguity than semantics of a single word, and we assume that it should be better preserved by translation to another language. If multilingual sentence embeddings intend to represent sentence semantics, then the similarity between embeddings of any two sentences must be invariant with respect to translation. Based on this suggestion, we consider a simple linear cross-lingual mapping as a possible improvement of the multilingual embeddings. We also consider deviation from orthogonality conditions as a measure of deficiency of the embeddings. | [
"Vasilyev, Oleg",
"Isono, Fumika",
"Bohannon, John"
]
| Linear Cross-Lingual Mapping of Sentence Embeddings | findings-acl.486 | Poster | 2305.14256v2 |
https://aclanthology.org/2024.findings-acl.487.bib | @inproceedings{zhang-etal-2024-ultra,
title = "{ULTRA}: Unleash {LLM}s{'} Potential for Event Argument Extraction through Hierarchical Modeling and Pair-wise Self-Refinement",
author = "Zhang, Xinliang Frederick and
Blum, Carter and
Choji, Temma and
Shah, Shalin and
Vempala, Alakananda",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.487",
pages = "8172--8185",
abstract = "Structural extraction of events within discourse is critical since it avails a deeper understanding of communication patterns and behavior trends. Event argument extraction (EAE), at the core of event-centric understanding, is the task of identifying role-specific text spans (i.e., arguments) for a given event. Document-level EAE (DocEAE) focuses on arguments that are scattered across an entire document. In this work, we explore open-source Large Language Models (LLMs) for DocEAE, and propose ULTRA, a hierarchical framework that extracts event arguments more cost-effectively. Further, it alleviates the positional bias issue intrinsic to LLMs. ULTRA sequentially reads text chunks of a document to generate a candidate argument set, upon which non-pertinent candidates are dropped through self-refinement. We introduce LEAFER to address the challenge LLMs face in locating the exact boundary of an argument. ULTRA outperforms strong baselines, including strong supervised models and ChatGPT, by 9.8{\%} when evaluated by Exact Match (EM).",
}
| Structural extraction of events within discourse is critical since it avails a deeper understanding of communication patterns and behavior trends. Event argument extraction (EAE), at the core of event-centric understanding, is the task of identifying role-specific text spans (i.e., arguments) for a given event. Document-level EAE (DocEAE) focuses on arguments that are scattered across an entire document. In this work, we explore open-source Large Language Models (LLMs) for DocEAE, and propose ULTRA, a hierarchical framework that extracts event arguments more cost-effectively. Further, it alleviates the positional bias issue intrinsic to LLMs. ULTRA sequentially reads text chunks of a document to generate a candidate argument set, upon which non-pertinent candidates are dropped through self-refinement. We introduce LEAFER to address the challenge LLMs face in locating the exact boundary of an argument. ULTRA outperforms strong baselines, including strong supervised models and ChatGPT, by 9.8{\%} when evaluated by Exact Match (EM). | [
"Zhang, Xinliang Frederick",
"Blum, Carter",
"Choji, Temma",
"Shah, Shalin",
"Vempala, Alakan",
"a"
]
| {ULTRA}: Unleash {LLM}s{'} Potential for Event Argument Extraction through Hierarchical Modeling and Pair-wise Self-Refinement | findings-acl.487 | Poster | 2401.13218v1 |
https://aclanthology.org/2024.findings-acl.488.bib | @inproceedings{lai-etal-2024-llms,
title = "{LLM}s Beyond {E}nglish: Scaling the Multilingual Capability of {LLM}s with Cross-Lingual Feedback",
author = "Lai, Wen and
Mesgar, Mohsen and
Fraser, Alexander",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.488",
pages = "8186--8213",
abstract = "To democratize large language models (LLMs) to most natural languages, it is imperative to make these models capable of understanding and generating texts in many languages, in particular low-resource ones. While recent multilingual LLMs demonstrate remarkable performance in such capabilities, these LLMs still support a limited number of human languages due to the lack of training data for low resource languages. Moreover, these LLMs are not yet aligned with human preference for downstream tasks, which is crucial for the success of LLMs in English. In this paper, we introduce xLLaMA-100 and xBLOOM-100 (collectively xLLMs-100), which scale the multilingual capabilities of LLaMA and BLOOM to 100 languages. To do so, we construct two datasets: a multilingual instruction dataset including 100 languages, which represents the largest language coverage to date, and a cross-lingual human feedback dataset encompassing 30 languages. We perform multilingual instruction tuning on the constructed instruction data and further align the LLMs with human feedback using the DPO algorithm on our cross-lingual human feedback dataset. We evaluate the multilingual understanding and generating capabilities of xLLMs-100 on five multilingual benchmarks. Experimental results show that xLLMs-100 consistently outperforms its peers across the benchmarks by considerable margins, defining a new state-of-the-art multilingual LLM that supports 100 languages.",
}
| To democratize large language models (LLMs) to most natural languages, it is imperative to make these models capable of understanding and generating texts in many languages, in particular low-resource ones. While recent multilingual LLMs demonstrate remarkable performance in such capabilities, these LLMs still support a limited number of human languages due to the lack of training data for low resource languages. Moreover, these LLMs are not yet aligned with human preference for downstream tasks, which is crucial for the success of LLMs in English. In this paper, we introduce xLLaMA-100 and xBLOOM-100 (collectively xLLMs-100), which scale the multilingual capabilities of LLaMA and BLOOM to 100 languages. To do so, we construct two datasets: a multilingual instruction dataset including 100 languages, which represents the largest language coverage to date, and a cross-lingual human feedback dataset encompassing 30 languages. We perform multilingual instruction tuning on the constructed instruction data and further align the LLMs with human feedback using the DPO algorithm on our cross-lingual human feedback dataset. We evaluate the multilingual understanding and generating capabilities of xLLMs-100 on five multilingual benchmarks. Experimental results show that xLLMs-100 consistently outperforms its peers across the benchmarks by considerable margins, defining a new state-of-the-art multilingual LLM that supports 100 languages. | [
"Lai, Wen",
"Mesgar, Mohsen",
"Fraser, Alex",
"er"
]
| {LLM}s Beyond {E}nglish: Scaling the Multilingual Capability of {LLM}s with Cross-Lingual Feedback | findings-acl.488 | Poster | 2406.01771v1 |
https://aclanthology.org/2024.findings-acl.489.bib | @inproceedings{qian-etal-2024-bass,
title = "{BASS}: Batched Attention-optimized Speculative Sampling",
author = "Qian, Haifeng and
Gonugondla, Sujan Kumar and
Ha, Sungsoo and
Shang, Mingyue and
Gouda, Sanjay Krishna and
Nallapati, Ramesh and
Sengupta, Sudipta and
Ma, Xiaofei and
Deoras, Anoop",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.489",
pages = "8214--8224",
abstract = "Speculative decoding has emerged as a powerful method to improve latency and throughput in hosting large language models. However, most existing implementations focus on generating a single sequence. Real-world generative AI applications often require multiple responses and how to perform speculative decoding in a batched setting while preserving its latency benefits poses non-trivial challenges. This paper describes a system of batched speculative decoding that sets a new state of the art in multi-sequence generation latency and that demonstrates superior GPU utilization as well as quality of generations within a time budget. For example, for a 7.8B-size model on a single A100 GPU and with a batch size of 8, each sequence is generated at an average speed of 5.8ms per token, the overall throughput being 1.1K tokens per second. These results represent state-of-the-art latency and a 2.15$\times$ speed-up over optimized regular decoding. Within a time budget that regular decoding does not finish, our system is able to generate sequences with HumanEval Pass@First of 43{\%} and Pass@All of 61{\%}, far exceeding what{'}s feasible with single-sequence speculative decoding. Our peak GPU utilization during decoding reaches as high as 15.8{\%}, more than 3$\times$ the highest of that of regular decoding and around 10$\times$ of single-sequence speculative decoding.",
}
| Speculative decoding has emerged as a powerful method to improve latency and throughput in hosting large language models. However, most existing implementations focus on generating a single sequence. Real-world generative AI applications often require multiple responses and how to perform speculative decoding in a batched setting while preserving its latency benefits poses non-trivial challenges. This paper describes a system of batched speculative decoding that sets a new state of the art in multi-sequence generation latency and that demonstrates superior GPU utilization as well as quality of generations within a time budget. For example, for a 7.8B-size model on a single A100 GPU and with a batch size of 8, each sequence is generated at an average speed of 5.8ms per token, the overall throughput being 1.1K tokens per second. These results represent state-of-the-art latency and a 2.15$\times$ speed-up over optimized regular decoding. Within a time budget that regular decoding does not finish, our system is able to generate sequences with HumanEval Pass@First of 43{\%} and Pass@All of 61{\%}, far exceeding what{'}s feasible with single-sequence speculative decoding. Our peak GPU utilization during decoding reaches as high as 15.8{\%}, more than 3$\times$ the highest of that of regular decoding and around 10$\times$ of single-sequence speculative decoding. | [
"Qian, Haifeng",
"Gonugondla, Sujan Kumar",
"Ha, Sungsoo",
"Shang, Mingyue",
"Gouda, Sanjay Krishna",
"Nallapati, Ramesh",
"Sengupta, Sudipta",
"Ma, Xiaofei",
"Deoras, Anoop"
]
| {BASS}: Batched Attention-optimized Speculative Sampling | findings-acl.489 | Poster | 2311.04951v2 |
https://aclanthology.org/2024.findings-acl.490.bib | @inproceedings{wu-etal-2024-deciphering,
title = "Deciphering Digital Detectives: Understanding {LLM} Behaviors and Capabilities in Multi-Agent Mystery Games",
author = "Wu, Dekun and
Shi, Haochen and
Sun, Zhiyuan and
Liu, Bang",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.490",
pages = "8225--8291",
abstract = "In this study, we explore the application of Large Language Models (LLMs) in Jubensha, a Chinese detective role-playing game and a novel area in Artificial Intelligence (AI) driven gaming. We introduce the first dataset specifically for Jubensha, including character scripts and game rules, to foster AI agent development in this complex narrative environment. Our work also presents a unique multi-agent interaction framework using LLMs, allowing AI agents to autonomously engage in Jubensha games. To evaluate the gaming performance of these AI agents, we developed novel methods measuring their mastery of case information and reasoning skills. Furthermore, we incorporated the latest advancements in prompting engineering to enhance the agents{'} performance in information gathering, murderer identification, and logical reasoning. The experimental results validate the effectiveness of our proposed methods. This work aims to offer a novel perspective on understanding LLM capabilities and establish a new benchmark for evaluating large language model-based agents.",
}
| In this study, we explore the application of Large Language Models (LLMs) in Jubensha, a Chinese detective role-playing game and a novel area in Artificial Intelligence (AI) driven gaming. We introduce the first dataset specifically for Jubensha, including character scripts and game rules, to foster AI agent development in this complex narrative environment. Our work also presents a unique multi-agent interaction framework using LLMs, allowing AI agents to autonomously engage in Jubensha games. To evaluate the gaming performance of these AI agents, we developed novel methods measuring their mastery of case information and reasoning skills. Furthermore, we incorporated the latest advancements in prompting engineering to enhance the agents{'} performance in information gathering, murderer identification, and logical reasoning. The experimental results validate the effectiveness of our proposed methods. This work aims to offer a novel perspective on understanding LLM capabilities and establish a new benchmark for evaluating large language model-based agents. | [
"Wu, Dekun",
"Shi, Haochen",
"Sun, Zhiyuan",
"Liu, Bang"
]
| Deciphering Digital Detectives: Understanding {LLM} Behaviors and Capabilities in Multi-Agent Mystery Games | findings-acl.490 | Poster | 2312.00746v2 |
https://aclanthology.org/2024.findings-acl.491.bib | @inproceedings{shaier-etal-2024-say,
title = "It Is Not About What You Say, It Is About How You Say It: A Surprisingly Simple Approach for Improving Reading Comprehension",
author = "Shaier, Sagi and
Hunter, Lawrence and
Wense, Katharina",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.491",
pages = "8292--8305",
abstract = "Natural language processing has seen rapid progress over the past decade. Due to the speed of developments, some practices get established without proper evaluation. Considering one such case and focusing on reading comprehension, we ask our first research question: 1) How does the order of inputs {--} i.e., question and context {--} affect model performance? Additionally, given recent advancements in input emphasis, we ask a second research question: 2) Does emphasizing either the question, the context, or both enhance performance? Experimenting with 9 large language models across 3 datasets, we find that presenting the context before the question improves model performance, with an accuracy increase of up to 31{\%}. Furthermore, emphasizing the context yields superior results compared to question emphasis, and in general, emphasizing parts of the input is particularly effective for addressing questions that models lack the parametric knowledge to answer. Experimenting with both prompt-based and attention-based emphasis methods, we additionally find that the best method is surprisingly simple: it only requires concatenating a few tokens to the input and results in an ac- curacy improvement of up to 36{\%}, allowing smaller models to outperform their significantly larger counterparts.",
}
| Natural language processing has seen rapid progress over the past decade. Due to the speed of developments, some practices get established without proper evaluation. Considering one such case and focusing on reading comprehension, we ask our first research question: 1) How does the order of inputs {--} i.e., question and context {--} affect model performance? Additionally, given recent advancements in input emphasis, we ask a second research question: 2) Does emphasizing either the question, the context, or both enhance performance? Experimenting with 9 large language models across 3 datasets, we find that presenting the context before the question improves model performance, with an accuracy increase of up to 31{\%}. Furthermore, emphasizing the context yields superior results compared to question emphasis, and in general, emphasizing parts of the input is particularly effective for addressing questions that models lack the parametric knowledge to answer. Experimenting with both prompt-based and attention-based emphasis methods, we additionally find that the best method is surprisingly simple: it only requires concatenating a few tokens to the input and results in an ac- curacy improvement of up to 36{\%}, allowing smaller models to outperform their significantly larger counterparts. | [
"Shaier, Sagi",
"Hunter, Lawrence",
"Wense, Katharina"
]
| It Is Not About What You Say, It Is About How You Say It: A Surprisingly Simple Approach for Improving Reading Comprehension | findings-acl.491 | Poster | 2406.16779v1 |
https://aclanthology.org/2024.findings-acl.492.bib | @inproceedings{lo-etal-2024-large,
title = "Large Language Models Relearn Removed Concepts",
author = "Lo, Michelle and
Barez, Fazl and
Cohen, Shay",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.492",
pages = "8306--8323",
abstract = "Advances in model editing through neuron pruning hold promise for removing undesirable concepts from large language models. However, it remains unclear whether models have the capacity to reacquire pruned concepts after editing. To investigate this, we evaluate concept relearning in models by tracking concept saliency and similarity in pruned neurons during retraining for named entity recognition tasks. Our findings reveal that models can quickly regain performance post-pruning by relocating advanced concepts to earlier layers and reallocating pruned concepts to primed neurons with similar semantics. This suggests that models exhibit polysemantic capacities and can blend old and new concepts in individual neurons. While neuron pruning provides interpretability into model concepts, our results highlight the challenges of permanent concept removal for improved model *safety*. Monitoring concept reemergence and developing techniques to mitigate relearning of unsafe concepts will be important directions for more robust model editing. Overall, our work strongly demonstrates the resilience and fluidity of concept representations in LLMs post concept removal.",
}
| Advances in model editing through neuron pruning hold promise for removing undesirable concepts from large language models. However, it remains unclear whether models have the capacity to reacquire pruned concepts after editing. To investigate this, we evaluate concept relearning in models by tracking concept saliency and similarity in pruned neurons during retraining for named entity recognition tasks. Our findings reveal that models can quickly regain performance post-pruning by relocating advanced concepts to earlier layers and reallocating pruned concepts to primed neurons with similar semantics. This suggests that models exhibit polysemantic capacities and can blend old and new concepts in individual neurons. While neuron pruning provides interpretability into model concepts, our results highlight the challenges of permanent concept removal for improved model *safety*. Monitoring concept reemergence and developing techniques to mitigate relearning of unsafe concepts will be important directions for more robust model editing. Overall, our work strongly demonstrates the resilience and fluidity of concept representations in LLMs post concept removal. | [
"Lo, Michelle",
"Barez, Fazl",
"Cohen, Shay"
]
| Large Language Models Relearn Removed Concepts | findings-acl.492 | Poster | 2401.01814v1 |
https://aclanthology.org/2024.findings-acl.493.bib | @inproceedings{wang-etal-2024-towards-unified,
title = "Towards Unified Task Embeddings Across Multiple Models: Bridging the Gap for Prompt-Based Large Language Models and Beyond",
author = "Wang, Xinyu and
Xu, Hainiu and
Gui, Lin and
He, Yulan",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.493",
pages = "8324--8340",
abstract = "Task embedding, a meta-learning technique that captures task-specific information, has gained popularity, especially in areas such as multi-task learning, model editing, and interpretability. However, it faces challenges with the emergence of prompt-guided Large Language Models (LLMs) operating in a gradient-free manner. Existing task embedding methods rely on fine-tuned, task-specific language models, which hinders the adaptability of task embeddings across diverse models, especially prompt-based LLMs. To hardness the potential of task embeddings in the era of LLMs, we propose a framework for unified task embeddings (FUTE), harmonizing task embeddings from various models, including smaller language models and LLMs with varied prompts, within a single vector space. Such uniformity enables comparison and analysis of similarities amongst different models, broadening the scope and utility of existing task embedding methods in multi-model scenarios, while maintaining their performance comparable to architecture-specific methods.",
}
| Task embedding, a meta-learning technique that captures task-specific information, has gained popularity, especially in areas such as multi-task learning, model editing, and interpretability. However, it faces challenges with the emergence of prompt-guided Large Language Models (LLMs) operating in a gradient-free manner. Existing task embedding methods rely on fine-tuned, task-specific language models, which hinders the adaptability of task embeddings across diverse models, especially prompt-based LLMs. To hardness the potential of task embeddings in the era of LLMs, we propose a framework for unified task embeddings (FUTE), harmonizing task embeddings from various models, including smaller language models and LLMs with varied prompts, within a single vector space. Such uniformity enables comparison and analysis of similarities amongst different models, broadening the scope and utility of existing task embedding methods in multi-model scenarios, while maintaining their performance comparable to architecture-specific methods. | [
"Wang, Xinyu",
"Xu, Hainiu",
"Gui, Lin",
"He, Yulan"
]
| Towards Unified Task Embeddings Across Multiple Models: Bridging the Gap for Prompt-Based Large Language Models and Beyond | findings-acl.493 | Poster | 2402.14522v2 |
https://aclanthology.org/2024.findings-acl.494.bib | @inproceedings{liu-etal-2024-toad,
title = "{TOAD}: Task-Oriented Automatic Dialogs with Diverse Response Styles",
author = "Liu, Yinhong and
Fang, Yimai and
Vandyke, David and
Collier, Nigel",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.494",
pages = "8341--8356",
abstract = "In light of recent advances in large language models (LLMs), the expectations for the next generation of virtual assistants include enhanced naturalness and adaptability across diverse usage scenarios. However, the creation of high-quality annotated data for Task-Oriented Dialog (TOD) is recognized to be slow and costly. To address these challenges, we introduce Task-Oriented Automatic Dialogs (TOAD), a novel and scalable TOD dataset along with its automatic generation pipeline. The TOAD dataset simulates realistic app context interaction and provide a variety of system response style options. Two aspects of system response styles are considered, verbosity level and users{'} expression mirroring. We benchmark TOAD on two response generation tasks, and the results show that modeling more verbose responses or responses without user expression mirroring is more challenging.",
}
| In light of recent advances in large language models (LLMs), the expectations for the next generation of virtual assistants include enhanced naturalness and adaptability across diverse usage scenarios. However, the creation of high-quality annotated data for Task-Oriented Dialog (TOD) is recognized to be slow and costly. To address these challenges, we introduce Task-Oriented Automatic Dialogs (TOAD), a novel and scalable TOD dataset along with its automatic generation pipeline. The TOAD dataset simulates realistic app context interaction and provide a variety of system response style options. Two aspects of system response styles are considered, verbosity level and users{'} expression mirroring. We benchmark TOAD on two response generation tasks, and the results show that modeling more verbose responses or responses without user expression mirroring is more challenging. | [
"Liu, Yinhong",
"Fang, Yimai",
"V",
"yke, David",
"Collier, Nigel"
]
| {TOAD}: Task-Oriented Automatic Dialogs with Diverse Response Styles | findings-acl.494 | Poster | 2402.10137v3 |
https://aclanthology.org/2024.findings-acl.495.bib | @inproceedings{zhang-etal-2024-machine,
title = "Machine-Generated Text Localization",
author = "Zhang, Zhongping and
Qin, Wenda and
Plummer, Bryan",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.495",
pages = "8357--8371",
abstract = "Machine-Generated Text (MGT) detection aims to identify a piece of text as machine or human written. Prior work has primarily formulated MGT detection as a binary classification task over an entire document, with limited work exploring cases where only part of a document is machine generated. This paper provides the first in-depth study of MGT that localizes the portions of a document that were machine generated. Thus, if a bad actor were to change a key portion of a news article to spread misinformation, whole document MGT detection may fail since the vast majority is human written, but our approach can succeed due to its granular approach. A key challenge in our MGT localization task is that short spans of text, *e.g.*, a single sentence, provides little information indicating if it is machine generated due to its short length. To address this, we leverage contextual information, where we predict whether multiple sentences are machine or human written at once. This enables our approach to identify changes in style or content to boost performance. A gain of 4-13{\%} mean Average Precision (mAP) over prior work demonstrates the effectiveness of approach on five diverse datasets: GoodNews, VisualNews, WikiText, Essay, and WP. We release our implementation at https://github.com/Zhongping-Zhang/MGT{\_}Localization.",
}
| Machine-Generated Text (MGT) detection aims to identify a piece of text as machine or human written. Prior work has primarily formulated MGT detection as a binary classification task over an entire document, with limited work exploring cases where only part of a document is machine generated. This paper provides the first in-depth study of MGT that localizes the portions of a document that were machine generated. Thus, if a bad actor were to change a key portion of a news article to spread misinformation, whole document MGT detection may fail since the vast majority is human written, but our approach can succeed due to its granular approach. A key challenge in our MGT localization task is that short spans of text, *e.g.*, a single sentence, provides little information indicating if it is machine generated due to its short length. To address this, we leverage contextual information, where we predict whether multiple sentences are machine or human written at once. This enables our approach to identify changes in style or content to boost performance. A gain of 4-13{\%} mean Average Precision (mAP) over prior work demonstrates the effectiveness of approach on five diverse datasets: GoodNews, VisualNews, WikiText, Essay, and WP. We release our implementation at https://github.com/Zhongping-Zhang/MGT{\_}Localization. | [
"Zhang, Zhongping",
"Qin, Wenda",
"Plummer, Bryan"
]
| Machine-Generated Text Localization | findings-acl.495 | Poster | 2112.07513v1 |
https://aclanthology.org/2024.findings-acl.496.bib | @inproceedings{lamarche-langlais-2024-benchie,
title = "{B}ench{IE}{\textasciicircum}{FL}: A Manually Re-Annotated Fact-Based Open Information Extraction Benchmark",
author = "Lamarche, Fabrice and
Langlais, Philippe",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.496",
pages = "8372--8394",
abstract = "Open Information Extraction (OIE) is a field of natural language processing that aims to present textual information in a format that allows it to be organized, analyzed and reflected upon. Numerous OIE systems are developed, claiming ever-increasing performance, marking the need for objective benchmarks. BenchIE is the latest reference we know of. Despite being very well thought out, we noticed a number of issues we believe are limiting. Therefore, we propose BenchIE{\textasciicircum}FL, a new OIE benchmark which fully enforces the principles of BenchIE while containing fewer errors, omissions and shortcomings when candidate facts are matched towards reference ones. BenchIE{\textasciicircum}FL allows insightful conclusions to be drawn on the actual performance of OIE extractors.",
}
| Open Information Extraction (OIE) is a field of natural language processing that aims to present textual information in a format that allows it to be organized, analyzed and reflected upon. Numerous OIE systems are developed, claiming ever-increasing performance, marking the need for objective benchmarks. BenchIE is the latest reference we know of. Despite being very well thought out, we noticed a number of issues we believe are limiting. Therefore, we propose BenchIE{\textasciicircum}FL, a new OIE benchmark which fully enforces the principles of BenchIE while containing fewer errors, omissions and shortcomings when candidate facts are matched towards reference ones. BenchIE{\textasciicircum}FL allows insightful conclusions to be drawn on the actual performance of OIE extractors. | [
"Lamarche, Fabrice",
"Langlais, Philippe"
]
| {B}ench{IE}{\textasciicircum}{FL}: A Manually Re-Annotated Fact-Based Open Information Extraction Benchmark | findings-acl.496 | Poster | 2407.16860v1 |
https://aclanthology.org/2024.findings-acl.497.bib | @inproceedings{agrawal-etal-2024-causalcite,
title = "{C}ausal{C}ite: A Causal Formulation of Paper Citations",
author = {Agrawal, Ishan and
Jin, Zhijing and
Mokhtarian, Ehsan and
Guo, Siyuan and
Chen, Yuen and
Sachan, Mrinmaya and
Sch{\"o}lkopf, Bernhard},
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.497",
pages = "8395--8410",
abstract = "Citation count of a paper is a commonly used proxy for evaluating the significance of a paper in the scientific community. Yet citation measures are widely criticized for failing to accurately reflect the true impact of a paper. Thus, we propose CausalCite, a new way to measure the significance of a paper by assessing the causal impact of the paper on its follow-up papers. CausalCite is based on a novel causal inference method, TextMatch, which adapts the traditional matching framework to high-dimensional text embeddings. TextMatch encodes each paper using text embeddings from large language models (LLMs), extracts similar samples by cosine similarity, and synthesizes a counterfactual sample as the weighted average of similar papers according to their similarity values. We demonstrate the effectiveness of CausalCite on various criteria, such as high correlation with paper impact as reported by scientific experts on a previous dataset of 1K papers, (test-of-time) awards for past papers, and its stability across various subfields of AI. We also provide a set of findings that can serve as suggested ways for future researchers to use our metric for a better understanding of the quality of a paper. Our code is available at https://github.com/causalNLP/causal-cite.",
}
| Citation count of a paper is a commonly used proxy for evaluating the significance of a paper in the scientific community. Yet citation measures are widely criticized for failing to accurately reflect the true impact of a paper. Thus, we propose CausalCite, a new way to measure the significance of a paper by assessing the causal impact of the paper on its follow-up papers. CausalCite is based on a novel causal inference method, TextMatch, which adapts the traditional matching framework to high-dimensional text embeddings. TextMatch encodes each paper using text embeddings from large language models (LLMs), extracts similar samples by cosine similarity, and synthesizes a counterfactual sample as the weighted average of similar papers according to their similarity values. We demonstrate the effectiveness of CausalCite on various criteria, such as high correlation with paper impact as reported by scientific experts on a previous dataset of 1K papers, (test-of-time) awards for past papers, and its stability across various subfields of AI. We also provide a set of findings that can serve as suggested ways for future researchers to use our metric for a better understanding of the quality of a paper. Our code is available at https://github.com/causalNLP/causal-cite. | [
"Agrawal, Ishan",
"Jin, Zhijing",
"Mokhtarian, Ehsan",
"Guo, Siyuan",
"Chen, Yuen",
"Sachan, Mrinmaya",
"Sch{\\\"o}lkopf, Bernhard"
]
| {C}ausal{C}ite: A Causal Formulation of Paper Citations | findings-acl.497 | Poster | 2311.02790v3 |
https://aclanthology.org/2024.findings-acl.498.bib | @inproceedings{zhu-etal-2024-question,
title = "Question Translation Training for Better Multilingual Reasoning",
author = "Zhu, Wenhao and
Huang, Shujian and
Yuan, Fei and
She, Shuaijie and
Chen, Jiajun and
Birch, Alexandra",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.498",
pages = "8411--8423",
abstract = "Large language models show compelling performance on reasoning tasks but they tend to perform much worse in languages other than English. This is unsurprising given that their training data largely consists of English text and instructions. A typical solution is to translate instruction data into all languages of interest, and then train on the resulting multilingual data, which is called translate-training. This approach not only incurs high cost, but also results in poorly translated data due to the non-standard formatting of mathematical chain-of-thought. In this paper, we explore the benefits of question alignment, where we train the model to translate reasoning questions into English by finetuning on X-English parallel question data. In this way we perform targeted, in-domain language alignment which makes best use of English instruction data to unlock the LLMs{'} multilingual reasoning abilities. Experimental results on LLaMA2-13B show that question alignment leads to consistent improvements over the translate-training approach: an average improvement of 11.3{\%} and 16.1{\%} accuracy across ten languages on the MGSM and MSVAMP multilingual reasoning benchmarks.",
}
| Large language models show compelling performance on reasoning tasks but they tend to perform much worse in languages other than English. This is unsurprising given that their training data largely consists of English text and instructions. A typical solution is to translate instruction data into all languages of interest, and then train on the resulting multilingual data, which is called translate-training. This approach not only incurs high cost, but also results in poorly translated data due to the non-standard formatting of mathematical chain-of-thought. In this paper, we explore the benefits of question alignment, where we train the model to translate reasoning questions into English by finetuning on X-English parallel question data. In this way we perform targeted, in-domain language alignment which makes best use of English instruction data to unlock the LLMs{'} multilingual reasoning abilities. Experimental results on LLaMA2-13B show that question alignment leads to consistent improvements over the translate-training approach: an average improvement of 11.3{\%} and 16.1{\%} accuracy across ten languages on the MGSM and MSVAMP multilingual reasoning benchmarks. | [
"Zhu, Wenhao",
"Huang, Shujian",
"Yuan, Fei",
"She, Shuaijie",
"Chen, Jiajun",
"Birch, Alex",
"ra"
]
| Question Translation Training for Better Multilingual Reasoning | findings-acl.498 | Poster | 2401.07817v3 |
https://aclanthology.org/2024.findings-acl.499.bib | @inproceedings{wang-etal-2024-improving,
title = "Improving {LLM} Generations via Fine-Grained Self-Endorsement",
author = "Wang, Ante and
Song, Linfeng and
Peng, Baolin and
Jin, Lifeng and
Tian, Ye and
Mi, Haitao and
Su, Jinsong and
Yu, Dong",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.499",
pages = "8424--8436",
abstract = "This work studies mitigating fact-conflicting hallucinations for large language model (LLM) at inference time.Particularly, we propose a self-endorsement framework that leverages the fine-grained fact-level comparisons across multiple sampled responses.Compared with prior ensemble methods (e.g., self-consistency) that perform response-level selection, our approach can better alleviate hallucinations for knowledge-intensive tasks.Our approach can broadly benefit smaller and open-source LLMs as it mainly conducts simple content-based comparisons.Experiments on Biographies show that our method can effectively improve the factuality of generations with simple and intuitive prompts across different scales of LLMs.Besides, comprehensive analyses on TriviaQA and GSM8K demonstrate the potential of self-endorsement for broader application.",
}
| This work studies mitigating fact-conflicting hallucinations for large language model (LLM) at inference time.Particularly, we propose a self-endorsement framework that leverages the fine-grained fact-level comparisons across multiple sampled responses.Compared with prior ensemble methods (e.g., self-consistency) that perform response-level selection, our approach can better alleviate hallucinations for knowledge-intensive tasks.Our approach can broadly benefit smaller and open-source LLMs as it mainly conducts simple content-based comparisons.Experiments on Biographies show that our method can effectively improve the factuality of generations with simple and intuitive prompts across different scales of LLMs.Besides, comprehensive analyses on TriviaQA and GSM8K demonstrate the potential of self-endorsement for broader application. | [
"Wang, Ante",
"Song, Linfeng",
"Peng, Baolin",
"Jin, Lifeng",
"Tian, Ye",
"Mi, Haitao",
"Su, Jinsong",
"Yu, Dong"
]
| Improving {LLM} Generations via Fine-Grained Self-Endorsement | findings-acl.499 | Poster | 2402.10614v2 |
https://aclanthology.org/2024.findings-acl.500.bib | @inproceedings{long-etal-2024-multi,
title = "Multi-Label Classification for Implicit Discourse Relation Recognition",
author = "Long, Wanqiu and
N, Siddharth and
Webber, Bonnie",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.500",
pages = "8437--8451",
abstract = "Discourse relations play a pivotal role in establishing coherence within textual content, uniting sentences and clauses into a cohesive narrative. The Penn Discourse Treebank (PDTB) stands as one of the most extensively utilized datasets in this domain. In PDTB-3, the annotators can assign multiple labels to an example, when they believe the simultaneous presence of multiple relations. Prior research in discourse relation recognition has treated these instances as separate examples during training, with a gold-standard prediction matching one of the labels considered correct at test time. However, this approach is inadequate, as it fails to account for the interdependence of labels in real-world contexts and to distinguish between cases where only one sense relation holds and cases where multiple relations hold simultaneously. In our work, we address this challenge by exploring various multi-label classification frameworks to handle implicit discourse relation recognition. We show that the methods for multi-label prediction don{'}t depress performance for single-label prediction. Additionally, we give comprehensive analysis of results and data. Our work contributes to advancing the understanding and application of discourse relations and provide a foundation for the future study.",
}
| Discourse relations play a pivotal role in establishing coherence within textual content, uniting sentences and clauses into a cohesive narrative. The Penn Discourse Treebank (PDTB) stands as one of the most extensively utilized datasets in this domain. In PDTB-3, the annotators can assign multiple labels to an example, when they believe the simultaneous presence of multiple relations. Prior research in discourse relation recognition has treated these instances as separate examples during training, with a gold-standard prediction matching one of the labels considered correct at test time. However, this approach is inadequate, as it fails to account for the interdependence of labels in real-world contexts and to distinguish between cases where only one sense relation holds and cases where multiple relations hold simultaneously. In our work, we address this challenge by exploring various multi-label classification frameworks to handle implicit discourse relation recognition. We show that the methods for multi-label prediction don{'}t depress performance for single-label prediction. Additionally, we give comprehensive analysis of results and data. Our work contributes to advancing the understanding and application of discourse relations and provide a foundation for the future study. | [
"Long, Wanqiu",
"N, Siddharth",
"Webber, Bonnie"
]
| Multi-Label Classification for Implicit Discourse Relation Recognition | findings-acl.500 | Poster | 2406.04461v1 |
https://aclanthology.org/2024.findings-acl.501.bib | @inproceedings{babe-etal-2024-studenteval,
title = "{S}tudent{E}val: A Benchmark of Student-Written Prompts for Large Language Models of Code",
author = "Babe, Hannah and
Nguyen, Sydney and
Zi, Yangtian and
Guha, Arjun and
Feldman, Molly and
Anderson, Carolyn",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.501",
pages = "8452--8474",
abstract = "Code LLMs have the potential to make it easier for non-experts to understand and write code. However, current CodeLLM benchmarks rely on a single expert-written prompt per problem, making it hard to generalize their success to non-expert users. In this paper, we present a new natural-language-to-code benchmark of prompts written by a key population of non-experts: beginning programmers. StudentEval contains 1,749 prompts written by 80 students who have only completed one introductory Python course. StudentEval contains numerous non-expert prompts describing the same problem, enabling exploration of key factors in prompt success. We use StudentEval to evaluate 12 Code LLMs and find that StudentEval is a better discriminator of model performance than existing benchmarks. Our analysis of student prompting strategies reveals that nondeterministic LLM sampling can mislead students about the quality of their descriptions, a finding with key implications for Code LLMs in education.",
}
| Code LLMs have the potential to make it easier for non-experts to understand and write code. However, current CodeLLM benchmarks rely on a single expert-written prompt per problem, making it hard to generalize their success to non-expert users. In this paper, we present a new natural-language-to-code benchmark of prompts written by a key population of non-experts: beginning programmers. StudentEval contains 1,749 prompts written by 80 students who have only completed one introductory Python course. StudentEval contains numerous non-expert prompts describing the same problem, enabling exploration of key factors in prompt success. We use StudentEval to evaluate 12 Code LLMs and find that StudentEval is a better discriminator of model performance than existing benchmarks. Our analysis of student prompting strategies reveals that nondeterministic LLM sampling can mislead students about the quality of their descriptions, a finding with key implications for Code LLMs in education. | [
"Babe, Hannah",
"Nguyen, Sydney",
"Zi, Yangtian",
"Guha, Arjun",
"Feldman, Molly",
"Anderson, Carolyn"
]
| {S}tudent{E}val: A Benchmark of Student-Written Prompts for Large Language Models of Code | findings-acl.501 | Poster | 2305.18507v2 |
https://aclanthology.org/2024.findings-acl.502.bib | @inproceedings{zhang-etal-2024-prolex,
title = "{P}ro{L}ex: A Benchmark for Language Proficiency-oriented Lexical Substitution",
author = "Zhang, Xuanming and
Chen, Zixun and
Yu, Zhou",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.502",
pages = "8475--8493",
abstract = "Lexical Substitution discovers appropriate substitutes for a given target word in a context sentence. However, the task fails to consider substitutes that are of equal or higher proficiency than the target, an aspect that could be beneficial for language learners looking to improve their writing. To bridge this gap, we propose a new task {---} language proficiency-oriented lexical substitution. We also introduce ProLex, a novel benchmark designed to assess systems{'} ability to generate not only appropriate substitutes but also substitutes that demonstrate better language proficiency. Besides the benchmark, we propose models that can automatically perform the new task. We show that our best model, a Llama2-13B model fine-tuned with task-specific synthetic data, outperforms ChatGPT by an average of 3.2{\%} in F-score and achieves comparable results with GPT-4 on ProLex.",
}
| Lexical Substitution discovers appropriate substitutes for a given target word in a context sentence. However, the task fails to consider substitutes that are of equal or higher proficiency than the target, an aspect that could be beneficial for language learners looking to improve their writing. To bridge this gap, we propose a new task {---} language proficiency-oriented lexical substitution. We also introduce ProLex, a novel benchmark designed to assess systems{'} ability to generate not only appropriate substitutes but also substitutes that demonstrate better language proficiency. Besides the benchmark, we propose models that can automatically perform the new task. We show that our best model, a Llama2-13B model fine-tuned with task-specific synthetic data, outperforms ChatGPT by an average of 3.2{\%} in F-score and achieves comparable results with GPT-4 on ProLex. | [
"Zhang, Xuanming",
"Chen, Zixun",
"Yu, Zhou"
]
| {P}ro{L}ex: A Benchmark for Language Proficiency-oriented Lexical Substitution | findings-acl.502 | Poster | 2106.04102v2 |
https://aclanthology.org/2024.findings-acl.503.bib | @inproceedings{jinnai-etal-2024-generating,
title = "Generating Diverse and High-Quality Texts by Minimum {B}ayes Risk Decoding",
author = "Jinnai, Yuu and
Honda, Ukyo and
Morimura, Tetsuro and
Zhang, Peinan",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.503",
pages = "8494--8525",
abstract = "One of the most important challenges in text generation systems is to produce outputs that are not only correct but also diverse.Recently, Minimum Bayes-Risk (MBR) decoding has gained prominence for generating sentences of the highest quality among the decoding algorithms. However, existing algorithms proposed to generate diverse outputs are predominantly based on beam search or random sampling, thus their output quality is capped by these underlying decoding algorithms. In this paper, we investigate an alternative approach {--} we develop diversity-promoting decoding algorithms by enforcing diversity objectives to MBR decoding.We propose two variants of MBR; (i) Diverse MBR (DMBR) that adds a diversity penalty to the decoding objective and (ii) $k$-medoids MBR (KMBR) that reformulates the decoding task as a clustering problem.We evaluate DMBR and KMBR on a variety of directed text generation tasks using encoder-decoder models and a language model with prompting. The experimental results show that the proposed method achieves a better trade-off than the diverse beam search and sampling algorithms overall.",
}
| One of the most important challenges in text generation systems is to produce outputs that are not only correct but also diverse.Recently, Minimum Bayes-Risk (MBR) decoding has gained prominence for generating sentences of the highest quality among the decoding algorithms. However, existing algorithms proposed to generate diverse outputs are predominantly based on beam search or random sampling, thus their output quality is capped by these underlying decoding algorithms. In this paper, we investigate an alternative approach {--} we develop diversity-promoting decoding algorithms by enforcing diversity objectives to MBR decoding.We propose two variants of MBR; (i) Diverse MBR (DMBR) that adds a diversity penalty to the decoding objective and (ii) $k$-medoids MBR (KMBR) that reformulates the decoding task as a clustering problem.We evaluate DMBR and KMBR on a variety of directed text generation tasks using encoder-decoder models and a language model with prompting. The experimental results show that the proposed method achieves a better trade-off than the diverse beam search and sampling algorithms overall. | [
"Jinnai, Yuu",
"Honda, Ukyo",
"Morimura, Tetsuro",
"Zhang, Peinan"
]
| Generating Diverse and High-Quality Texts by Minimum {B}ayes Risk Decoding | findings-acl.503 | Poster | 2401.05054v2 |
https://aclanthology.org/2024.findings-acl.504.bib | @inproceedings{rarrick-etal-2024-gate,
title = "{GATE} {X}-{E} : A Challenge Set for Gender-Fair Translations from Weakly-Gendered Languages",
author = "Rarrick, Spencer and
Naik, Ranjita and
Poudel, Sundar and
Chowdhary, Vishal",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.504",
pages = "8526--8546",
abstract = "Neural Machine Translation (NMT) continues to improve in quality and adoption, yet the in advertent perpetuation of gender bias remains a significant concern. Despite numerous studies on gender bias in translations into English from weakly gendered-languages, there are no benchmarks for evaluating this phenomenon or for assessing mitigation strategies. To address this gap, we introduce GATE X-E, an extension to the GATE (Rarrick et al., 2023) corpus, that consists of human translations from Turkish, Hungarian, Finnish, and Persian into English. Each translation is accompanied by feminine, masculine, and neutral variants. The dataset, which contains between 1250 and 1850 instances for each of the four language pairs, features natural sentences with a wide range of sentence lengths and domains, challenging translation rewriters on various linguistic phenomena. Additionally, we present a translation gender rewriting solution built with GPT-4 and use GATE X-E to evaluate it. We open source our contributions to encourage further research on gender debiasing.",
}
| Neural Machine Translation (NMT) continues to improve in quality and adoption, yet the in advertent perpetuation of gender bias remains a significant concern. Despite numerous studies on gender bias in translations into English from weakly gendered-languages, there are no benchmarks for evaluating this phenomenon or for assessing mitigation strategies. To address this gap, we introduce GATE X-E, an extension to the GATE (Rarrick et al., 2023) corpus, that consists of human translations from Turkish, Hungarian, Finnish, and Persian into English. Each translation is accompanied by feminine, masculine, and neutral variants. The dataset, which contains between 1250 and 1850 instances for each of the four language pairs, features natural sentences with a wide range of sentence lengths and domains, challenging translation rewriters on various linguistic phenomena. Additionally, we present a translation gender rewriting solution built with GPT-4 and use GATE X-E to evaluate it. We open source our contributions to encourage further research on gender debiasing. | [
"Rarrick, Spencer",
"Naik, Ranjita",
"Poudel, Sundar",
"Chowdhary, Vishal"
]
| {GATE} {X}-{E} : A Challenge Set for Gender-Fair Translations from Weakly-Gendered Languages | findings-acl.504 | Poster | 2402.14277v1 |
https://aclanthology.org/2024.findings-acl.505.bib | @inproceedings{jinnai-ariu-2024-hyperparameter,
title = "Hyperparameter-Free Approach for Faster Minimum {B}ayes Risk Decoding",
author = "Jinnai, Yuu and
Ariu, Kaito",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.505",
pages = "8547--8566",
abstract = "Minimum Bayes-Risk (MBR) decoding is shown to be a powerful alternative to beam search decoding for a wide range of text generation tasks. However, MBR requires a huge amount of time for inference to compute the MBR objective, which makes the method infeasible in many situations where response time is critical. Confidence-based pruning (CBP) (Cheng and Vlachos, 2023) has recently been proposed to reduce the inference time in machine translation tasks. Although it is shown to significantly reduce the amount of computation, it requires hyperparameter tuning using a development set to be effective. To this end, we propose Adaptive Minimum Bayes-Risk (AMBR) decoding, a hyperparameter-free method to run MBR decoding efficiently. AMBR is derived from the observation that the problem of computing the sample-based MBR objective is the medoid identification problem. AMBR uses the Correlated Sequential Halving (CSH) algorithm (Baharav and Tse, 2019), the algorithm with the best performance guarantee to date for the medoid identification problem, to compute the sample-based MBR objective. We evaluate AMBR on machine translation, text summarization, and image captioning tasks. The results show that AMBR achieves on par with CBP, with CBP selecting hyperparameters through an Oracle for each given computation budget.",
}
| Minimum Bayes-Risk (MBR) decoding is shown to be a powerful alternative to beam search decoding for a wide range of text generation tasks. However, MBR requires a huge amount of time for inference to compute the MBR objective, which makes the method infeasible in many situations where response time is critical. Confidence-based pruning (CBP) (Cheng and Vlachos, 2023) has recently been proposed to reduce the inference time in machine translation tasks. Although it is shown to significantly reduce the amount of computation, it requires hyperparameter tuning using a development set to be effective. To this end, we propose Adaptive Minimum Bayes-Risk (AMBR) decoding, a hyperparameter-free method to run MBR decoding efficiently. AMBR is derived from the observation that the problem of computing the sample-based MBR objective is the medoid identification problem. AMBR uses the Correlated Sequential Halving (CSH) algorithm (Baharav and Tse, 2019), the algorithm with the best performance guarantee to date for the medoid identification problem, to compute the sample-based MBR objective. We evaluate AMBR on machine translation, text summarization, and image captioning tasks. The results show that AMBR achieves on par with CBP, with CBP selecting hyperparameters through an Oracle for each given computation budget. | [
"Jinnai, Yuu",
"Ariu, Kaito"
]
| Hyperparameter-Free Approach for Faster Minimum {B}ayes Risk Decoding | findings-acl.505 | Poster | 2401.02749v2 |
https://aclanthology.org/2024.findings-acl.506.bib | @inproceedings{oshika-etal-2024-simplifying,
title = "Simplifying Translations for Children: Iterative Simplification Considering Age of Acquisition with {LLM}s",
author = "Oshika, Masashi and
Morishita, Makoto and
Hirao, Tsutomu and
Sasano, Ryohei and
Takeda, Koichi",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.506",
pages = "8567--8577",
abstract = "In recent years, neural machine translation (NMT) has become widely used in everyday life. However, the current NMT lacks a mechanism to adjust the difficulty level of translations to match the user{'}s language level. Additionally, due to the bias in the training data for NMT, translations of simple source sentences are often produced with complex words. In particular, this could pose a problem for children, who may not be able to understand the meaning of the translations correctly. In this study, we propose a method that replaces high Age of Acquisitions (AoA) words in translations with simpler words to match the translations to the user{'}s level. We achieve this by using large language models (LLMs), providing a triple of a source sentence, a translation, and a target word to be replaced. We create a benchmark dataset using back-translation on Simple English Wikipedia. The experimental results obtained from the dataset show that our method effectively replaces high-AoA words with lower-AoA words and, moreover, can iteratively replace most of the high-AoA words while still maintaining high BLEU and COMET scores.",
}
| In recent years, neural machine translation (NMT) has become widely used in everyday life. However, the current NMT lacks a mechanism to adjust the difficulty level of translations to match the user{'}s language level. Additionally, due to the bias in the training data for NMT, translations of simple source sentences are often produced with complex words. In particular, this could pose a problem for children, who may not be able to understand the meaning of the translations correctly. In this study, we propose a method that replaces high Age of Acquisitions (AoA) words in translations with simpler words to match the translations to the user{'}s level. We achieve this by using large language models (LLMs), providing a triple of a source sentence, a translation, and a target word to be replaced. We create a benchmark dataset using back-translation on Simple English Wikipedia. The experimental results obtained from the dataset show that our method effectively replaces high-AoA words with lower-AoA words and, moreover, can iteratively replace most of the high-AoA words while still maintaining high BLEU and COMET scores. | [
"Oshika, Masashi",
"Morishita, Makoto",
"Hirao, Tsutomu",
"Sasano, Ryohei",
"Takeda, Koichi"
]
| Simplifying Translations for Children: Iterative Simplification Considering Age of Acquisition with {LLM}s | findings-acl.506 | Poster | 2408.04217v1 |
https://aclanthology.org/2024.findings-acl.507.bib | @inproceedings{liu-etal-2024-bi,
title = "Bi-Chainer: Automated Large Language Models Reasoning with Bidirectional Chaining",
author = "Liu, Shuqi and
He, Bowei and
Song, Linqi",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.507",
pages = "8578--8598",
abstract = "Large Language Models (LLMs) have shown human-like reasoning abilities but still face challenges in solving complex logical problems. Existing unidirectional chaining methods, such as forward chaining and backward chaining, suffer from issues like low prediction accuracy and efficiency. To address these, we propose a bidirectional chaining method, Bi-Chainer, which dynamically switches to depth-first reasoning in the opposite reasoning direction when it encounters multiple branching options within the current direction. Thus, the intermediate reasoning results can be utilized as guidance to facilitate the reasoning process. We show that Bi-Chainer achieves sizable accuracy boots over unidirectional chaining frameworks on four challenging logical reasoning datasets. Moreover, Bi-Chainer enhances the accuracy of intermediate proof steps and reduces the average number of inference calls, resulting in more efficient and accurate reasoning.",
}
| Large Language Models (LLMs) have shown human-like reasoning abilities but still face challenges in solving complex logical problems. Existing unidirectional chaining methods, such as forward chaining and backward chaining, suffer from issues like low prediction accuracy and efficiency. To address these, we propose a bidirectional chaining method, Bi-Chainer, which dynamically switches to depth-first reasoning in the opposite reasoning direction when it encounters multiple branching options within the current direction. Thus, the intermediate reasoning results can be utilized as guidance to facilitate the reasoning process. We show that Bi-Chainer achieves sizable accuracy boots over unidirectional chaining frameworks on four challenging logical reasoning datasets. Moreover, Bi-Chainer enhances the accuracy of intermediate proof steps and reduces the average number of inference calls, resulting in more efficient and accurate reasoning. | [
"Liu, Shuqi",
"He, Bowei",
"Song, Linqi"
]
| Bi-Chainer: Automated Large Language Models Reasoning with Bidirectional Chaining | findings-acl.507 | Poster | 2406.06586v1 |
https://aclanthology.org/2024.findings-acl.508.bib | @inproceedings{fonseca-cohen-2024-large-language,
title = "Can Large Language Model Summarizers Adapt to Diverse Scientific Communication Goals?",
author = "Fonseca, Marcio and
Cohen, Shay",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.508",
pages = "8599--8618",
abstract = "In this work, we investigate the controllability of large language models (LLMs) on scientific summarization tasks. We identify key stylistic and content coverage factors that characterize different types of summaries such as paper reviews, abstracts, and lay summaries. By controlling stylistic features, we find that non-fine-tuned LLMs outperform humans in the MuP review generation task, both in terms of similarity to reference summaries and human preferences. Also, we show that we can improve the controllability of LLMs with keyword-based classifier-free guidance (CFG) while achieving lexical overlap comparable to strong fine-tuned baselines on arXiv and PubMed. However, our results also indicate that LLMs cannot consistently generate long summaries with more than 8 sentences. Furthermore, these models exhibit limited capacity to produce highly abstractive lay summaries. Although LLMs demonstrate strong generic summarization competency, sophisticated content control without costly fine-tuning remains an open problem for domain-specific applications.",
}
| In this work, we investigate the controllability of large language models (LLMs) on scientific summarization tasks. We identify key stylistic and content coverage factors that characterize different types of summaries such as paper reviews, abstracts, and lay summaries. By controlling stylistic features, we find that non-fine-tuned LLMs outperform humans in the MuP review generation task, both in terms of similarity to reference summaries and human preferences. Also, we show that we can improve the controllability of LLMs with keyword-based classifier-free guidance (CFG) while achieving lexical overlap comparable to strong fine-tuned baselines on arXiv and PubMed. However, our results also indicate that LLMs cannot consistently generate long summaries with more than 8 sentences. Furthermore, these models exhibit limited capacity to produce highly abstractive lay summaries. Although LLMs demonstrate strong generic summarization competency, sophisticated content control without costly fine-tuning remains an open problem for domain-specific applications. | [
"Fonseca, Marcio",
"Cohen, Shay"
]
| Can Large Language Model Summarizers Adapt to Diverse Scientific Communication Goals? | findings-acl.508 | Poster | 2401.10415v2 |
https://aclanthology.org/2024.findings-acl.509.bib | @inproceedings{yang-etal-2024-knowledge,
title = "Knowledge Context Modeling with Pre-trained Language Models for Contrastive Knowledge Graph Completion",
author = "Yang, Guangqian and
Liu, Yi and
Zhang, Lei and
Zhang, Licheng and
Xie, Hongtao and
Mao, Zhendong",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.509",
pages = "8619--8630",
abstract = "Text-based knowledge graph completion (KGC) methods utilize pre-trained language models for triple encoding and further fine-tune the model to achieve completion. Despite their excellent performance, they neglect the knowledge context in inferring process. Intuitively, knowledge contexts, which refer to the neighboring triples around the target triples, are important information for triple inferring, since they provide additional detailed information about the entities. To this end, we propose a novel framework named KnowC, which models the knowledge context as additional prompts with pre-trained language models for knowledge graph completion. Given the substantial number of neighbors typically associated with entities, along with the constrained input token capacity of language models, we further devise several strategies to sample the neighbors. We conduct extensive experiments on common datasets FB15k-237, WN18RR and Wikidata5M, experiments show that KnowC achieves state-of-the-art performance.",
}
| Text-based knowledge graph completion (KGC) methods utilize pre-trained language models for triple encoding and further fine-tune the model to achieve completion. Despite their excellent performance, they neglect the knowledge context in inferring process. Intuitively, knowledge contexts, which refer to the neighboring triples around the target triples, are important information for triple inferring, since they provide additional detailed information about the entities. To this end, we propose a novel framework named KnowC, which models the knowledge context as additional prompts with pre-trained language models for knowledge graph completion. Given the substantial number of neighbors typically associated with entities, along with the constrained input token capacity of language models, we further devise several strategies to sample the neighbors. We conduct extensive experiments on common datasets FB15k-237, WN18RR and Wikidata5M, experiments show that KnowC achieves state-of-the-art performance. | [
"Yang, Guangqian",
"Liu, Yi",
"Zhang, Lei",
"Zhang, Licheng",
"Xie, Hongtao",
"Mao, Zhendong"
]
| Knowledge Context Modeling with Pre-trained Language Models for Contrastive Knowledge Graph Completion | findings-acl.509 | Poster | 2010.00309v1 |
https://aclanthology.org/2024.findings-acl.510.bib | @inproceedings{zhang-etal-2024-stronger,
title = "Stronger, Lighter, Better: Towards Life-Long Attribute Value Extraction for {E}-Commerce Products",
author = "Zhang, Tao and
Zhang, Chenwei and
Li, Xian and
Shang, Jingbo and
Nguyen, Hoang and
Yu, Philip",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.510",
pages = "8631--8643",
abstract = "Attribute value extraction involves identifying the value spans of predetermined attributes in product texts. This area of research has traditionally operated under a closed-world assumption, focusing on products from a static set of categories and their associated attributes. However, products in e-commerce stores are ever-increasing and evolving, calling for life-long learning. If continuously trained on the fast-increasing products and attributes, most existing solutions not only struggle for parameter efficiency but also endure foreseeable defects due to data contamination, catastrophic forgetting, etc. As a remedy, we propose and study a new task, which aims to effectively maintain a strong single model for many domains in a life-long learning fashion, without jeopardizing the model performance and parameter efficiency. We introduce factorization into the model and make it domain-aware by decoupling the modeling of product type and attribute, as a way to promote de-contamination and parameter efficiency while scaling up. Tuning the model with distillation prevents forgetting historical knowledge and enables continuous learning from emerging domains. Experiments on hundreds of domains showed that our model attains the near state-of-the-art performance with affordable parameter size, the least historical knowledge forgetting, and the greatest robustness against noises, whilst adding only a few parameters per domain when compared with competitive baselines.",
}
| Attribute value extraction involves identifying the value spans of predetermined attributes in product texts. This area of research has traditionally operated under a closed-world assumption, focusing on products from a static set of categories and their associated attributes. However, products in e-commerce stores are ever-increasing and evolving, calling for life-long learning. If continuously trained on the fast-increasing products and attributes, most existing solutions not only struggle for parameter efficiency but also endure foreseeable defects due to data contamination, catastrophic forgetting, etc. As a remedy, we propose and study a new task, which aims to effectively maintain a strong single model for many domains in a life-long learning fashion, without jeopardizing the model performance and parameter efficiency. We introduce factorization into the model and make it domain-aware by decoupling the modeling of product type and attribute, as a way to promote de-contamination and parameter efficiency while scaling up. Tuning the model with distillation prevents forgetting historical knowledge and enables continuous learning from emerging domains. Experiments on hundreds of domains showed that our model attains the near state-of-the-art performance with affordable parameter size, the least historical knowledge forgetting, and the greatest robustness against noises, whilst adding only a few parameters per domain when compared with competitive baselines. | [
"Zhang, Tao",
"Zhang, Chenwei",
"Li, Xian",
"Shang, Jingbo",
"Nguyen, Hoang",
"Yu, Philip"
]
| Stronger, Lighter, Better: Towards Life-Long Attribute Value Extraction for {E}-Commerce Products | findings-acl.510 | Poster | 2312.06684v1 |
https://aclanthology.org/2024.findings-acl.511.bib | @inproceedings{namgoong-etal-2024-exploring,
title = "Exploring Domain Robust Lightweight Reward Models based on Router Mechanism",
author = "Namgoong, Hyuk and
Jung, Jeesu and
Jung, Sangkeun and
Roh, YoonHyung",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.511",
pages = "8644--8652",
abstract = "Recent advancements in large language models have heavily relied on the large reward model from reinforcement learning from human feedback for fine-tuning. However, the use of a single reward model across various domains may not always be optimal, often requiring retraining from scratch when new domain data is introduced. To address these challenges, we explore the utilization of small language models operating in a domain-specific manner based on router mechanisms. Our three approaches are: 1) utilize mixture of experts to form a single reward model by modularizing an internal router and experts, 2) employing external router to select the appropriate reward model from multiple domain-specific models, and 3) the framework reduces parameter size by loading reward models and router adapters onto a single small language model using adapters. Experimental validation underscores the effectiveness of our approach, demonstrating performance comparable to baseline methods while also reducing the total parameter size.",
}
| Recent advancements in large language models have heavily relied on the large reward model from reinforcement learning from human feedback for fine-tuning. However, the use of a single reward model across various domains may not always be optimal, often requiring retraining from scratch when new domain data is introduced. To address these challenges, we explore the utilization of small language models operating in a domain-specific manner based on router mechanisms. Our three approaches are: 1) utilize mixture of experts to form a single reward model by modularizing an internal router and experts, 2) employing external router to select the appropriate reward model from multiple domain-specific models, and 3) the framework reduces parameter size by loading reward models and router adapters onto a single small language model using adapters. Experimental validation underscores the effectiveness of our approach, demonstrating performance comparable to baseline methods while also reducing the total parameter size. | [
"Namgoong, Hyuk",
"Jung, Jeesu",
"Jung, Sangkeun",
"Roh, YoonHyung"
]
| Exploring Domain Robust Lightweight Reward Models based on Router Mechanism | findings-acl.511 | Poster | 2407.17546v1 |
https://aclanthology.org/2024.findings-acl.512.bib | @inproceedings{an-etal-2024-generalized,
title = "Generalized Category Discovery with Large Language Models in the Loop",
author = "An, Wenbin and
Shi, Wenkai and
Tian, Feng and
Lin, Haonan and
Wang, QianYing and
Wu, Yaqiang and
Cai, Mingxiang and
Wang, Luyan and
Chen, Yan and
Zhu, Haiping and
Chen, Ping",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.512",
pages = "8653--8665",
abstract = "Generalized Category Discovery (GCD) is a crucial task that aims to recognize both known and novel categories from a set of unlabeled data by utilizing a few labeled data with only known categories. Due to the lack of supervision and category information, current methods usually perform poorly on novel categories and struggle to reveal semantic meanings of the discovered clusters, which limits their applications in the real world. To mitigate the above issues, we propose Loop, an end-to-end active-learning framework that introduces Large Language Models (LLMs) into the training loop, which can boost model performance and generate category names without relying on any human efforts. Specifically, we first propose Local Inconsistent Sampling (LIS) to select samples that have a higher probability of falling to wrong clusters, based on neighborhood prediction consistency and entropy of cluster assignment probabilities. Then we propose a Scalable Query strategy to allow LLMs to choose true neighbors of the selected samples from multiple candidate samples. Based on the feedback from LLMs, we perform Refined Neighborhood Contrastive Learning (RNCL) to pull samples and their neighbors closer to learn clustering-friendly representations. Finally, we select representative samples from clusters corresponding to novel categories to allow LLMs to generate category names for them. Extensive experiments on three benchmark datasets show that Loop outperforms SOTA models by a large margin and generates accurate category names for the discovered clusters. Code and data are available at https://github.com/Lackel/LOOP.",
}
| Generalized Category Discovery (GCD) is a crucial task that aims to recognize both known and novel categories from a set of unlabeled data by utilizing a few labeled data with only known categories. Due to the lack of supervision and category information, current methods usually perform poorly on novel categories and struggle to reveal semantic meanings of the discovered clusters, which limits their applications in the real world. To mitigate the above issues, we propose Loop, an end-to-end active-learning framework that introduces Large Language Models (LLMs) into the training loop, which can boost model performance and generate category names without relying on any human efforts. Specifically, we first propose Local Inconsistent Sampling (LIS) to select samples that have a higher probability of falling to wrong clusters, based on neighborhood prediction consistency and entropy of cluster assignment probabilities. Then we propose a Scalable Query strategy to allow LLMs to choose true neighbors of the selected samples from multiple candidate samples. Based on the feedback from LLMs, we perform Refined Neighborhood Contrastive Learning (RNCL) to pull samples and their neighbors closer to learn clustering-friendly representations. Finally, we select representative samples from clusters corresponding to novel categories to allow LLMs to generate category names for them. Extensive experiments on three benchmark datasets show that Loop outperforms SOTA models by a large margin and generates accurate category names for the discovered clusters. Code and data are available at https://github.com/Lackel/LOOP. | [
"An, Wenbin",
"Shi, Wenkai",
"Tian, Feng",
"Lin, Haonan",
"Wang, QianYing",
"Wu, Yaqiang",
"Cai, Mingxiang",
"Wang, Luyan",
"Chen, Yan",
"Zhu, Haiping",
"Chen, Ping"
]
| Generalized Category Discovery with Large Language Models in the Loop | findings-acl.512 | Poster | 2312.10897v2 |
https://aclanthology.org/2024.findings-acl.513.bib | @inproceedings{wang-etal-2024-vaegpt,
title = "{VAEGPT}-Sim: Improving Sentence Representation with Limited Corpus Using Gradually-Denoising {VAE}",
author = "Wang, Zhenyi and
Ning, Haiyan and
Ling, Qing and
Wang, Dan",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.513",
pages = "8666--8681",
abstract = "Text embedding requires a highly efficient method for training domain-specific models on limited data, as general models trained on large corpora lack universal applicability in highly specific fields. Therefore, we have introduced VAEGPT-Sim, an innovative model for generating synonyms that combines a denoising variational autoencoder with a target-specific discriminator to generate synonymous sentences that closely resemble human language. Even when trained with completely unsupervised settings, it maintains a harmonious balance between semantic similarity and lexical diversity, as shown by a comprehensive evaluation metric system with the highest average scores compared to other generative models. When VAEGPT-Sim is utilized as a module for contrastive learning in text representation, it delivers state-of-the-art results in small-dataset training on STS benchmarks, surpassing ConSERT by 2.8 points. This approach optimizes the effectiveness of text representation despite a limited corpus, signifying an advancement in domain-specific embedding technology.",
}
| Text embedding requires a highly efficient method for training domain-specific models on limited data, as general models trained on large corpora lack universal applicability in highly specific fields. Therefore, we have introduced VAEGPT-Sim, an innovative model for generating synonyms that combines a denoising variational autoencoder with a target-specific discriminator to generate synonymous sentences that closely resemble human language. Even when trained with completely unsupervised settings, it maintains a harmonious balance between semantic similarity and lexical diversity, as shown by a comprehensive evaluation metric system with the highest average scores compared to other generative models. When VAEGPT-Sim is utilized as a module for contrastive learning in text representation, it delivers state-of-the-art results in small-dataset training on STS benchmarks, surpassing ConSERT by 2.8 points. This approach optimizes the effectiveness of text representation despite a limited corpus, signifying an advancement in domain-specific embedding technology. | [
"Wang, Zhenyi",
"Ning, Haiyan",
"Ling, Qing",
"Wang, Dan"
]
| {VAEGPT}-Sim: Improving Sentence Representation with Limited Corpus Using Gradually-Denoising {VAE} | findings-acl.513 | Poster | 2209.04333v3 |
https://aclanthology.org/2024.findings-acl.514.bib | @inproceedings{guo-etal-2024-pptc,
title = "{PPTC} Benchmark: Evaluating Large Language Models for {P}ower{P}oint Task Completion",
author = "Guo, Yiduo and
Zhang, Zekai and
Liang, Yaobo and
Zhao, Dongyan and
Duan, Nan",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.514",
pages = "8682--8701",
abstract = "Recent evaluations of Large Language Models (LLMs) have centered around testing their zero-shot/few-shot capabilities for basic natural language tasks and their ability to translate instructions into tool APIs. However, the evaluation of LLMs utilizing complex tools to finish multi-turn, multi-modal instructions in a complex multi-modal environment has not been investigated. To address this gap, we introduce the PowerPoint Task Completion (PPTC) benchmark to assess LLMs{'} ability to create and edit PPT files based on user instructions. It contains 279 multi-turn sessions covering diverse topics and hundreds of instructions involving multi-modal operations. We also propose the PPTX-Match Evaluation System that evaluates if LLMs finish the instruction based on the prediction file rather than the label API sequence, thus it supports various LLM-generated API sequences. We measure 3 closed LLMs and 6 open-source LLMs. The results show that GPT-4 outperforms other LLMs with 75.1{\%} accuracy in single-turn dialogue testing but faces challenges in completing entire sessions, achieving just 6{\%} session accuracy. We find three main error causes in our benchmark: error accumulation in the multi-turn session, long PPT template processing, and multi-modality perception. These pose great challenges for future LLM and agent systems .",
}
| Recent evaluations of Large Language Models (LLMs) have centered around testing their zero-shot/few-shot capabilities for basic natural language tasks and their ability to translate instructions into tool APIs. However, the evaluation of LLMs utilizing complex tools to finish multi-turn, multi-modal instructions in a complex multi-modal environment has not been investigated. To address this gap, we introduce the PowerPoint Task Completion (PPTC) benchmark to assess LLMs{'} ability to create and edit PPT files based on user instructions. It contains 279 multi-turn sessions covering diverse topics and hundreds of instructions involving multi-modal operations. We also propose the PPTX-Match Evaluation System that evaluates if LLMs finish the instruction based on the prediction file rather than the label API sequence, thus it supports various LLM-generated API sequences. We measure 3 closed LLMs and 6 open-source LLMs. The results show that GPT-4 outperforms other LLMs with 75.1{\%} accuracy in single-turn dialogue testing but faces challenges in completing entire sessions, achieving just 6{\%} session accuracy. We find three main error causes in our benchmark: error accumulation in the multi-turn session, long PPT template processing, and multi-modality perception. These pose great challenges for future LLM and agent systems . | [
"Guo, Yiduo",
"Zhang, Zekai",
"Liang, Yaobo",
"Zhao, Dongyan",
"Duan, Nan"
]
| {PPTC} Benchmark: Evaluating Large Language Models for {P}ower{P}oint Task Completion | findings-acl.514 | Poster | 2311.01767v2 |
https://aclanthology.org/2024.findings-acl.515.bib | @inproceedings{zhao-etal-2024-fact,
title = "Fact-and-Reflection ({F}a{R}) Improves Confidence Calibration of Large Language Models",
author = "Zhao, Xinran and
Zhang, Hongming and
Pan, Xiaoman and
Yao, Wenlin and
Yu, Dong and
Wu, Tongshuang and
Chen, Jianshu",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.515",
pages = "8702--8718",
abstract = "For a LLM to be trustworthy, its confidence level should be well-calibrated with its actual performance. While it is now common sense that LLM performances are greatly impacted by prompts, the confidence calibration in prompting LLMs has yet to be thoroughly explored.In this paper, we explore how different prompting strategies influence LLM confidence calibration and how it could be improved. We conduct extensive experiments on six prompting methods in the question-answering context and we observe that, while these methods help improve the expected LLM calibration, they also trigger LLMs to be over-confident when responding to some instances.Inspired by human cognition, we propose Fact-and-Reflection (FaR) prompting, which improves the LLM calibration in two steps. First, FaR elicits the known {``}facts{''} that are relevant to the input prompt from the LLM. And then it asks the model to {``}reflect{''} over them to generate the final answer.Experiments show that FaR prompting achieves significantly better calibration; it lowers the Expected Calibration Error by 23.5{\%} on our multi-purpose QA tasks. Notably, FaR prompting even elicits the capability of verbally expressing concerns in less confident scenarios, which helps trigger retrieval augmentation for solving these harder instances.",
}
| For a LLM to be trustworthy, its confidence level should be well-calibrated with its actual performance. While it is now common sense that LLM performances are greatly impacted by prompts, the confidence calibration in prompting LLMs has yet to be thoroughly explored.In this paper, we explore how different prompting strategies influence LLM confidence calibration and how it could be improved. We conduct extensive experiments on six prompting methods in the question-answering context and we observe that, while these methods help improve the expected LLM calibration, they also trigger LLMs to be over-confident when responding to some instances.Inspired by human cognition, we propose Fact-and-Reflection (FaR) prompting, which improves the LLM calibration in two steps. First, FaR elicits the known {``}facts{''} that are relevant to the input prompt from the LLM. And then it asks the model to {``}reflect{''} over them to generate the final answer.Experiments show that FaR prompting achieves significantly better calibration; it lowers the Expected Calibration Error by 23.5{\%} on our multi-purpose QA tasks. Notably, FaR prompting even elicits the capability of verbally expressing concerns in less confident scenarios, which helps trigger retrieval augmentation for solving these harder instances. | [
"Zhao, Xinran",
"Zhang, Hongming",
"Pan, Xiaoman",
"Yao, Wenlin",
"Yu, Dong",
"Wu, Tongshuang",
"Chen, Jianshu"
]
| Fact-and-Reflection ({F}a{R}) Improves Confidence Calibration of Large Language Models | findings-acl.515 | Poster | 2402.17124v1 |
https://aclanthology.org/2024.findings-acl.516.bib | @inproceedings{chen-etal-2024-db,
title = "{DB}-{LLM}: Accurate Dual-Binarization for Efficient {LLM}s",
author = "Chen, Hong and
Lv, Chengtao and
Ding, Liang and
Qin, Haotong and
Zhou, Xiabin and
Ding, Yifu and
Liu, Xuebo and
Zhang, Min and
Guo, Jinyang and
Liu, Xianglong and
Tao, Dacheng",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.516",
pages = "8719--8730",
abstract = "Large language models (LLMs) have significantly advanced the field of natural language processing, while the expensive memory and computation consumption impede their practical deployment. Quantization emerges as one of the most effective methods for improving the computational efficiency of LLMs. However, existing ultra-low-bit quantization always causes severe accuracy drops. In this paper, we empirically investigate the micro and macro characteristics of ultra-low bit quantization and present a novel \textbf{D}ual-\textbf{B}inarization method for \textbf{LLM}s, namely \textbf{DB-LLM}. For the micro-level, we take both the accuracy advantage of 2-bit-width and the efficiency advantage of binarization into account, introducing \textit{Flexible Dual Binarization} (\textbf{FDB}). By splitting 2-bit quantized weights into two independent sets of binaries, FDB ensures the accuracy of representations and introduces flexibility, utilizing the efficient bitwise operations of binarization while retaining the inherent high sparsity of ultra-low bit quantization. For the macro-level, we find the distortion that exists in the prediction of LLM after quantization, which is specified as the deviations related to the ambiguity of samples. We propose the \textit{Deviation-Aware Distillation} (\textbf{DAD}) method, enabling the model to focus differently on various samples. Comprehensive experiments show that our DB-LLM not only significantly surpasses the current State-of-The-Art (SoTA) in ultra-low bit quantization (, perplexity decreased from 9.64 to 7.23), but also achieves an additional 20{\%} reduction in computational consumption compared to the SOTA method under the same bit-width. Our code is available at https://github.com/Hon-Chen/DB-LLM.",
}
| Large language models (LLMs) have significantly advanced the field of natural language processing, while the expensive memory and computation consumption impede their practical deployment. Quantization emerges as one of the most effective methods for improving the computational efficiency of LLMs. However, existing ultra-low-bit quantization always causes severe accuracy drops. In this paper, we empirically investigate the micro and macro characteristics of ultra-low bit quantization and present a novel \textbf{D}ual-\textbf{B}inarization method for \textbf{LLM}s, namely \textbf{DB-LLM}. For the micro-level, we take both the accuracy advantage of 2-bit-width and the efficiency advantage of binarization into account, introducing \textit{Flexible Dual Binarization} (\textbf{FDB}). By splitting 2-bit quantized weights into two independent sets of binaries, FDB ensures the accuracy of representations and introduces flexibility, utilizing the efficient bitwise operations of binarization while retaining the inherent high sparsity of ultra-low bit quantization. For the macro-level, we find the distortion that exists in the prediction of LLM after quantization, which is specified as the deviations related to the ambiguity of samples. We propose the \textit{Deviation-Aware Distillation} (\textbf{DAD}) method, enabling the model to focus differently on various samples. Comprehensive experiments show that our DB-LLM not only significantly surpasses the current State-of-The-Art (SoTA) in ultra-low bit quantization (, perplexity decreased from 9.64 to 7.23), but also achieves an additional 20{\%} reduction in computational consumption compared to the SOTA method under the same bit-width. Our code is available at https://github.com/Hon-Chen/DB-LLM. | [
"Chen, Hong",
"Lv, Chengtao",
"Ding, Liang",
"Qin, Haotong",
"Zhou, Xiabin",
"Ding, Yifu",
"Liu, Xuebo",
"Zhang, Min",
"Guo, Jinyang",
"Liu, Xianglong",
"Tao, Dacheng"
]
| {DB}-{LLM}: Accurate Dual-Binarization for Efficient {LLM}s | findings-acl.516 | Poster | 2405.19888v1 |
https://aclanthology.org/2024.findings-acl.517.bib | @inproceedings{liu-etal-2024-tempcompass,
title = "{T}emp{C}ompass: Do Video {LLM}s Really Understand Videos?",
author = "Liu, Yuanxin and
Li, Shicheng and
Liu, Yi and
Wang, Yuxiang and
Ren, Shuhuai and
Li, Lei and
Chen, Sishuo and
Sun, Xu and
Hou, Lu",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.517",
pages = "8731--8772",
abstract = "Recently, there is a surge in interest surrounding video large language models (Video LLMs). However, existing benchmarks fail to provide a comprehensive feedback on the temporal perception ability of Video LLMs. On the one hand, most of them are unable to distinguish between different temporal aspects (e.g., speed, direction) and thus cannot reflect the nuanced performance on these specific aspects. On the other hand, they are limited in the diversity of task formats (e.g., only multi-choice QA), which hinders the understanding of how temporal perception performance may vary across different types of tasks. Motivated by these two problems, we propose the \textbf{TempCompass} benchmark, which introduces a diversity of temporal aspects and task formats. To collect high-quality test data, we devise two novel strategies: (1) In video collection, we construct conflicting videos that share the same static content but differ in a specific temporal aspect, which prevents Video LLMs from leveraging single-frame bias or language priors. (2) To collect the task instructions, we propose a paradigm where humans first annotate meta-information for a video and then an LLM generates the instruction. We also design an LLM-based approach to automatically and accurately evaluate the responses from Video LLMs. Based on TempCompass, we comprehensively evaluate 9 state-of-the-art (SOTA) Video LLMs and 3 Image LLMs, and reveal the discerning fact that these models exhibit notably poor temporal perception ability.",
}
| Recently, there is a surge in interest surrounding video large language models (Video LLMs). However, existing benchmarks fail to provide a comprehensive feedback on the temporal perception ability of Video LLMs. On the one hand, most of them are unable to distinguish between different temporal aspects (e.g., speed, direction) and thus cannot reflect the nuanced performance on these specific aspects. On the other hand, they are limited in the diversity of task formats (e.g., only multi-choice QA), which hinders the understanding of how temporal perception performance may vary across different types of tasks. Motivated by these two problems, we propose the \textbf{TempCompass} benchmark, which introduces a diversity of temporal aspects and task formats. To collect high-quality test data, we devise two novel strategies: (1) In video collection, we construct conflicting videos that share the same static content but differ in a specific temporal aspect, which prevents Video LLMs from leveraging single-frame bias or language priors. (2) To collect the task instructions, we propose a paradigm where humans first annotate meta-information for a video and then an LLM generates the instruction. We also design an LLM-based approach to automatically and accurately evaluate the responses from Video LLMs. Based on TempCompass, we comprehensively evaluate 9 state-of-the-art (SOTA) Video LLMs and 3 Image LLMs, and reveal the discerning fact that these models exhibit notably poor temporal perception ability. | [
"Liu, Yuanxin",
"Li, Shicheng",
"Liu, Yi",
"Wang, Yuxiang",
"Ren, Shuhuai",
"Li, Lei",
"Chen, Sishuo",
"Sun, Xu",
"Hou, Lu"
]
| {T}emp{C}ompass: Do Video {LLM}s Really Understand Videos? | findings-acl.517 | Poster | 2403.00476v3 |
https://aclanthology.org/2024.findings-acl.518.bib | @inproceedings{zhu-rzeszotarski-2024-get,
title = "{``}Get Their Hands Dirty, Not Mine{''}: On Researcher-Annotator Collaboration and the Agency of Annotators",
author = "Zhu, Shengqi and
Rzeszotarski, Jeffrey",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.518",
pages = "8773--8782",
abstract = "Annotation quality is often framed as post-hoc cleanup of annotator-caused issues. This position paper discusses whether, how, and why this narrative limits the scope of improving annotation. We call to consider annotation as a procedural collaboration, outlining three points in this direction:(1) An issue can be either annotator- or researcher-oriented, where one party is accountable and the other party may lack ability to fix it; (2) yet, they can co-occur or have similar consequences, and thus any specific problem we encounter may be a combination;(3) therefore, we need a new language to capture the nuance and holistically describe the full procedure to resolve these issues.To that end, we propose to study how agency is manifested in annotation and picture how this perspective benefits the community more broadly.",
}
| Annotation quality is often framed as post-hoc cleanup of annotator-caused issues. This position paper discusses whether, how, and why this narrative limits the scope of improving annotation. We call to consider annotation as a procedural collaboration, outlining three points in this direction:(1) An issue can be either annotator- or researcher-oriented, where one party is accountable and the other party may lack ability to fix it; (2) yet, they can co-occur or have similar consequences, and thus any specific problem we encounter may be a combination;(3) therefore, we need a new language to capture the nuance and holistically describe the full procedure to resolve these issues.To that end, we propose to study how agency is manifested in annotation and picture how this perspective benefits the community more broadly. | [
"Zhu, Shengqi",
"Rzeszotarski, Jeffrey"
]
| {``}Get Their Hands Dirty, Not Mine{''}: On Researcher-Annotator Collaboration and the Agency of Annotators | findings-acl.518 | Poster | 2305.12815v2 |
https://aclanthology.org/2024.findings-acl.519.bib | @inproceedings{zhang-etal-2024-teaching,
title = "Teaching Large Language Models an Unseen Language on the Fly",
author = "Zhang, Chen and
Liu, Xiao and
Lin, Jiuheng and
Feng, Yansong",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.519",
pages = "8783--8800",
abstract = "Existing large language models struggle to support numerous low-resource languages, particularly the extremely low-resource ones, for which there is minimal training data available for effective parameter updating. We thus investigate whether LLMs can learn a new language on the fly solely through prompting. To study this question, we collect a research suite for Zhuang, a language supported by no LLMs currently. We introduce DiPMT++, a framework for adapting LLMs to unseen languages by in-context learning. Using a dictionary and 5K parallel sentences only, DiPMT++ significantly enhances the performance of GPT-4 from 0 to 16 BLEU for Chinese-to-Zhuang translation and achieves 32 BLEU for Zhuang-to-Chinese translation. We also validate the effectiveness of our framework on Kalamang, another unseen language. Furthermore, we demonstrate the practical utility of DiPMT++ in aiding humans in translating completely unseen languages, which could contribute to the preservation of linguistic diversity.",
}
| Existing large language models struggle to support numerous low-resource languages, particularly the extremely low-resource ones, for which there is minimal training data available for effective parameter updating. We thus investigate whether LLMs can learn a new language on the fly solely through prompting. To study this question, we collect a research suite for Zhuang, a language supported by no LLMs currently. We introduce DiPMT++, a framework for adapting LLMs to unseen languages by in-context learning. Using a dictionary and 5K parallel sentences only, DiPMT++ significantly enhances the performance of GPT-4 from 0 to 16 BLEU for Chinese-to-Zhuang translation and achieves 32 BLEU for Zhuang-to-Chinese translation. We also validate the effectiveness of our framework on Kalamang, another unseen language. Furthermore, we demonstrate the practical utility of DiPMT++ in aiding humans in translating completely unseen languages, which could contribute to the preservation of linguistic diversity. | [
"Zhang, Chen",
"Liu, Xiao",
"Lin, Jiuheng",
"Feng, Yansong"
]
| Teaching Large Language Models an Unseen Language on the Fly | findings-acl.519 | Poster | 2402.19167v2 |
https://aclanthology.org/2024.findings-acl.520.bib | @inproceedings{lu-etal-2024-error,
title = "Error Analysis Prompting Enables Human-Like Translation Evaluation in Large Language Models",
author = "Lu, Qingyu and
Qiu, Baopu and
Ding, Liang and
Zhang, Kanjian and
Kocmi, Tom and
Tao, Dacheng",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.520",
pages = "8801--8816",
abstract = "Generative large language models (LLMs), e.g., ChatGPT, have demonstrated remarkable proficiency across several NLP tasks, such as machine translation, text summarization. Recent research (Kocmi and Federmann, 2023) has shown that utilizing LLMs for assessing the quality of machine translation (MT) achieves state-of-the-art performance at the system level but performs poorly at the segment level. To further improve the performance of LLMs on MT quality assessment, we conduct an investigation into several prompting designs, and propose a new prompting method called Error Analysis Prompting (EAPrompt) by combining Chain-of-Thoughts (Wei et al., 2022) and Error Analysis (Lu et al., 2023). This technique emulates the commonly accepted human evaluation framework - Multidimensional Quality Metrics (MQM, Freitag et al., (2021)) and produces explainable and reliable MT evaluations at both the system and segment level. Experimental Results from WMT22 metrics shared task validate the effectiveness of EAPrompt on various LLMs, with different structures. Further analysis confirms that EAPrompt effectively distinguishes major errors from minor ones, while also sharing a similar distribution of the number of errors with MQM. These findings highlight the potential of EAPrompt as a human-like evaluator prompting technique for MT evaluation. We will release our code and scripts to facilitate the community.",
}
| Generative large language models (LLMs), e.g., ChatGPT, have demonstrated remarkable proficiency across several NLP tasks, such as machine translation, text summarization. Recent research (Kocmi and Federmann, 2023) has shown that utilizing LLMs for assessing the quality of machine translation (MT) achieves state-of-the-art performance at the system level but performs poorly at the segment level. To further improve the performance of LLMs on MT quality assessment, we conduct an investigation into several prompting designs, and propose a new prompting method called Error Analysis Prompting (EAPrompt) by combining Chain-of-Thoughts (Wei et al., 2022) and Error Analysis (Lu et al., 2023). This technique emulates the commonly accepted human evaluation framework - Multidimensional Quality Metrics (MQM, Freitag et al., (2021)) and produces explainable and reliable MT evaluations at both the system and segment level. Experimental Results from WMT22 metrics shared task validate the effectiveness of EAPrompt on various LLMs, with different structures. Further analysis confirms that EAPrompt effectively distinguishes major errors from minor ones, while also sharing a similar distribution of the number of errors with MQM. These findings highlight the potential of EAPrompt as a human-like evaluator prompting technique for MT evaluation. We will release our code and scripts to facilitate the community. | [
"Lu, Qingyu",
"Qiu, Baopu",
"Ding, Liang",
"Zhang, Kanjian",
"Kocmi, Tom",
"Tao, Dacheng"
]
| Error Analysis Prompting Enables Human-Like Translation Evaluation in Large Language Models | findings-acl.520 | Poster | 2303.13809v4 |
https://aclanthology.org/2024.findings-acl.521.bib | @inproceedings{zong-qiu-2024-gaokao,
title = "{GAOKAO}-{MM}: A {C}hinese Human-Level Benchmark for Multimodal Models Evaluation",
author = "Zong, Yi and
Qiu, Xipeng",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.521",
pages = "8817--8825",
abstract = "The Large Vision-Language Models (LVLMs) have demonstrated great abilities in image perception and language understanding. However, existing datasets either focus solely on primary perception abilities and commonsense knowledge, or have a low level of text comprehension difficulty, which are insufficient to reflect the comprehensive capabilities of LVLMs, particularly in terms of Chinese language proficiency. We propose GAOKAO-MM, a multimodal benchmark based on the Chinese College Entrance Examination (GAOKAO), comprising of 8 subjects and 12 types of images, such as diagrams, function graphs, maps and photos. GAOKAO-MM derives from native Chinese context and sets human-level requirements for the model{'}s abilities, including perception, understanding, knowledge and reasoning. We evaluate 10 LVLMs and find that the accuracies of all of them are lower than 50{\%}, with GPT-4-Vision (48.1{\%}), Qwen-VL-Plus (41.2{\%}) and Gemini-Pro-Vision (35.1{\%}) ranking in the top three positions. The results of our multi-dimension analysis indicate that LVLMs have moderate distance towards Artificial General Intelligence (AGI) and provide insights facilitating the development of multilingual LVLMs. The dataset and evaluation code are available through: https://github.com/OpenMOSS/GAOKAO-MM",
}
| The Large Vision-Language Models (LVLMs) have demonstrated great abilities in image perception and language understanding. However, existing datasets either focus solely on primary perception abilities and commonsense knowledge, or have a low level of text comprehension difficulty, which are insufficient to reflect the comprehensive capabilities of LVLMs, particularly in terms of Chinese language proficiency. We propose GAOKAO-MM, a multimodal benchmark based on the Chinese College Entrance Examination (GAOKAO), comprising of 8 subjects and 12 types of images, such as diagrams, function graphs, maps and photos. GAOKAO-MM derives from native Chinese context and sets human-level requirements for the model{'}s abilities, including perception, understanding, knowledge and reasoning. We evaluate 10 LVLMs and find that the accuracies of all of them are lower than 50{\%}, with GPT-4-Vision (48.1{\%}), Qwen-VL-Plus (41.2{\%}) and Gemini-Pro-Vision (35.1{\%}) ranking in the top three positions. The results of our multi-dimension analysis indicate that LVLMs have moderate distance towards Artificial General Intelligence (AGI) and provide insights facilitating the development of multilingual LVLMs. The dataset and evaluation code are available through: https://github.com/OpenMOSS/GAOKAO-MM | [
"Zong, Yi",
"Qiu, Xipeng"
]
| {GAOKAO}-{MM}: A {C}hinese Human-Level Benchmark for Multimodal Models Evaluation | findings-acl.521 | Poster | 2407.12023v1 |
https://aclanthology.org/2024.findings-acl.522.bib | @inproceedings{wang-etal-2024-diffchat,
title = "{D}iff{C}hat: Learning to Chat with Text-to-Image Synthesis Models for Interactive Image Creation",
author = "Wang, Jiapeng and
Wang, Chengyu and
Cao, Tingfeng and
Huang, Jun and
Jin, Lianwen",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.522",
pages = "8826--8840",
abstract = "We present DiffChat, a novel method to align Large Language Models (LLMs) to {``}chat{''} with prompt-as-input Text-to-Image Synthesis (TIS)models (e.g., Stable Diffusion) for interactive image creation. Given a raw prompt/image and a user-specified instruction, DiffChat can effectively make appropriate modifications and generate the target prompt, which can be leveraged to create the target image of high quality. To achieve this, we first collect an instruction-following prompt engineering dataset named InstructPE for the supervised training of DiffChat.Next, we propose a reinforcement learning framework with the feedback of three core criteria for image creation, i.e., aesthetics, user preference and content integrity. It involves an action-space dynamic modification technique to obtain more relevant positive samples and harder negative samples during the off-policy sampling. Content integrity is also introduced into the value estimation function for further improvement of produced images. Our method can exhibit superior performance than baseline models and strong competitors based on both automatic and human evaluations, which fully demonstrates its effectiveness.",
}
| We present DiffChat, a novel method to align Large Language Models (LLMs) to {``}chat{''} with prompt-as-input Text-to-Image Synthesis (TIS)models (e.g., Stable Diffusion) for interactive image creation. Given a raw prompt/image and a user-specified instruction, DiffChat can effectively make appropriate modifications and generate the target prompt, which can be leveraged to create the target image of high quality. To achieve this, we first collect an instruction-following prompt engineering dataset named InstructPE for the supervised training of DiffChat.Next, we propose a reinforcement learning framework with the feedback of three core criteria for image creation, i.e., aesthetics, user preference and content integrity. It involves an action-space dynamic modification technique to obtain more relevant positive samples and harder negative samples during the off-policy sampling. Content integrity is also introduced into the value estimation function for further improvement of produced images. Our method can exhibit superior performance than baseline models and strong competitors based on both automatic and human evaluations, which fully demonstrates its effectiveness. | [
"Wang, Jiapeng",
"Wang, Chengyu",
"Cao, Tingfeng",
"Huang, Jun",
"Jin, Lianwen"
]
| {D}iff{C}hat: Learning to Chat with Text-to-Image Synthesis Models for Interactive Image Creation | findings-acl.522 | Poster | 2403.04997v1 |
https://aclanthology.org/2024.findings-acl.523.bib | @inproceedings{yang-etal-2024-revisiting,
title = "Revisiting Parallel Context Windows: A Frustratingly Simple Alternative and Chain-of-Thought Deterioration",
author = "Yang, Kejuan and
Liu, Xiao and
Men, Kaiwen and
Zeng, Aohan and
Dong, Yuxiao and
Tang, Jie",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.523",
pages = "8841--8852",
abstract = "We identify two crucial limitations in the evaluation of recent parallel-integrated method Parallel Context Windows (PCW), which extends the maximum context lengths of language models, e.g., 2048 for LLaMA, by harnessing window-wise attention and positional embedding techniques. We first show that a simple yet strong baseline, weighted sum ensemble, is missing for the in-context few-shot classification. Moreover, on more challenging Chain-of-Thought (CoT) reasoning (e.g., HotpotQA), PCW would present unexpected deterioration regarding question miscomprehension and false inference. Based on our findings, we suggest that the existing PCW design may not guarantee sufficient improvement and practicality in handling lengthy documents in real-world applications. More community efforts on enabling language models{'} long context understanding ability should be paid.",
}
| We identify two crucial limitations in the evaluation of recent parallel-integrated method Parallel Context Windows (PCW), which extends the maximum context lengths of language models, e.g., 2048 for LLaMA, by harnessing window-wise attention and positional embedding techniques. We first show that a simple yet strong baseline, weighted sum ensemble, is missing for the in-context few-shot classification. Moreover, on more challenging Chain-of-Thought (CoT) reasoning (e.g., HotpotQA), PCW would present unexpected deterioration regarding question miscomprehension and false inference. Based on our findings, we suggest that the existing PCW design may not guarantee sufficient improvement and practicality in handling lengthy documents in real-world applications. More community efforts on enabling language models{'} long context understanding ability should be paid. | [
"Yang, Kejuan",
"Liu, Xiao",
"Men, Kaiwen",
"Zeng, Aohan",
"Dong, Yuxiao",
"Tang, Jie"
]
| Revisiting Parallel Context Windows: A Frustratingly Simple Alternative and Chain-of-Thought Deterioration | findings-acl.523 | Poster | 2305.15262v1 |
https://aclanthology.org/2024.findings-acl.524.bib | @inproceedings{zhang-etal-2024-rationales,
title = "Rationales for Answers to Simple Math Word Problems Confuse Large Language Models",
author = "Zhang, Yidan and
Xue, Mingfeng and
Liu, Dayiheng and
He, Zhenan",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.524",
pages = "8853--8869",
abstract = "Recently, large language models (LLMs) have demonstrated breakthrough mathematical problem-solving capabilities in grade school math word problems (MWP). For example, on the MWP benchmark GSM8K, the accuracy of GPT-3.5-Turbo and MetaMath-70B reaches 80.80{\%} and 82.30{\%}, respectively. One question arises, does it mean that LLMs have truly mastered related mathematical problem-solving abilities? In this paper, by presenting two types of benchmarks, where MCGSM8K aims at selecting one correct solution from four solutions, while GSM8K-Judgement judges whether a solution to a given question is true or false, we demonstrate that the ability of most LLMs to evaluate the mathematical reasoning process of MWP is far from sufficient. To compensate for this issue, we propose hybrid supervised fine-tuning data from the training data of GSM8K, MCGSM8K, and GSM8K-Judgement, which significantly improves performance on the proposed reasoning process evaluation benchmarks. For example, fine-tuning improves the performance of LLaMA-2-13B from 33.51{\%} to 70.89{\%} on MCGSM8K. In conclusion, we experimentally demonstrate that most LLMs have limited ability to evaluate the mathematical reasoning process of MWP, which can be enhanced through fine-tuning.",
}
| Recently, large language models (LLMs) have demonstrated breakthrough mathematical problem-solving capabilities in grade school math word problems (MWP). For example, on the MWP benchmark GSM8K, the accuracy of GPT-3.5-Turbo and MetaMath-70B reaches 80.80{\%} and 82.30{\%}, respectively. One question arises, does it mean that LLMs have truly mastered related mathematical problem-solving abilities? In this paper, by presenting two types of benchmarks, where MCGSM8K aims at selecting one correct solution from four solutions, while GSM8K-Judgement judges whether a solution to a given question is true or false, we demonstrate that the ability of most LLMs to evaluate the mathematical reasoning process of MWP is far from sufficient. To compensate for this issue, we propose hybrid supervised fine-tuning data from the training data of GSM8K, MCGSM8K, and GSM8K-Judgement, which significantly improves performance on the proposed reasoning process evaluation benchmarks. For example, fine-tuning improves the performance of LLaMA-2-13B from 33.51{\%} to 70.89{\%} on MCGSM8K. In conclusion, we experimentally demonstrate that most LLMs have limited ability to evaluate the mathematical reasoning process of MWP, which can be enhanced through fine-tuning. | [
"Zhang, Yidan",
"Xue, Mingfeng",
"Liu, Dayiheng",
"He, Zhenan"
]
| Rationales for Answers to Simple Math Word Problems Confuse Large Language Models | findings-acl.524 | Poster | 2312.04684v1 |
https://aclanthology.org/2024.findings-acl.525.bib | @inproceedings{shi-etal-2024-reslora,
title = "{R}es{L}o{RA}: Identity Residual Mapping in Low-Rank Adaption",
author = "Shi, Shuhua and
Huang, Shaohan and
Song, Minghui and
Li, Zhoujun and
Zhang, Zihan and
Huang, Haizhen and
Wei, Furu and
Deng, Weiwei and
Sun, Feng and
Zhang, Qi",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.525",
pages = "8870--8884",
abstract = "As one of the most popular parameter-efficient fine-tuning (PEFT) methods, low-rank adaptation (LoRA) is commonly applied to fine-tune large language models (LLMs). However, updating the weights of LoRA blocks effectively and expeditiously is challenging due to the long calculation path in the original model. To address this, we propose ResLoRA, an improved framework of LoRA. By adding residual paths during training and using merging approaches to eliminate these extra paths during inference, our method can achieve better results in fewer training steps without any extra trainable parameters or inference cost compared to LoRA. The experiments on NLG, NLU, and text-to-image tasks demonstrate the effectiveness of our method. To the best of our knowledge, ResLoRA is the first work that combines the residual path with LoRA. The code of our method is available at [this url](https://github.com/microsoft/LMOps/tree/main/reslora).",
}
| As one of the most popular parameter-efficient fine-tuning (PEFT) methods, low-rank adaptation (LoRA) is commonly applied to fine-tune large language models (LLMs). However, updating the weights of LoRA blocks effectively and expeditiously is challenging due to the long calculation path in the original model. To address this, we propose ResLoRA, an improved framework of LoRA. By adding residual paths during training and using merging approaches to eliminate these extra paths during inference, our method can achieve better results in fewer training steps without any extra trainable parameters or inference cost compared to LoRA. The experiments on NLG, NLU, and text-to-image tasks demonstrate the effectiveness of our method. To the best of our knowledge, ResLoRA is the first work that combines the residual path with LoRA. The code of our method is available at [this url](https://github.com/microsoft/LMOps/tree/main/reslora). | [
"Shi, Shuhua",
"Huang, Shaohan",
"Song, Minghui",
"Li, Zhoujun",
"Zhang, Zihan",
"Huang, Haizhen",
"Wei, Furu",
"Deng, Weiwei",
"Sun, Feng",
"Zhang, Qi"
]
| {R}es{L}o{RA}: Identity Residual Mapping in Low-Rank Adaption | findings-acl.525 | Poster | 2205.15127v1 |
https://aclanthology.org/2024.findings-acl.526.bib | @inproceedings{wang-etal-2024-towards-objectively,
title = "Towards Objectively Benchmarking Social Intelligence of Language Agents at the Action Level",
author = "Wang, Chenxu and
Dai, Bin and
Liu, Huaping and
Wang, Baoyuan",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.526",
pages = "8885--8897",
abstract = "Prominent large language models have exhibited human-level performance in many domains, even enabling the derived agents to simulate human and social interactions. While practical works have substantiated the practicability of grounding language agents in sandbox simulation or embodied simulators, current social intelligence benchmarks either stay at the language level or use subjective metrics. In pursuit of a more realistic and objective evaluation, we introduce the Social Tasks in Sandbox Simulation (STSS) benchmark, which assesses language agents objectively at the action level by scrutinizing the goal achievements within the multi-agent simulation.Additionally, we sample conversation scenarios to build a language-level benchmark to provide an economically prudent preliminary evaluation and align with prevailing benchmarks. To gauge the significance of agent architecture, we implement a target-driven planning (TDP) module as an adjunct to the existing agent. Our evaluative findings highlight that the STSS benchmark is challenging for state-of-the-art language agents. Furthermore, it effectively discriminates between distinct language agents, suggesting its usefulness as a benchmark for evaluating both language models and agent architectures. Our code is available at https://github.com/wcx21/Social-Tasks-in-Sandbox-Simulation.",
}
| Prominent large language models have exhibited human-level performance in many domains, even enabling the derived agents to simulate human and social interactions. While practical works have substantiated the practicability of grounding language agents in sandbox simulation or embodied simulators, current social intelligence benchmarks either stay at the language level or use subjective metrics. In pursuit of a more realistic and objective evaluation, we introduce the Social Tasks in Sandbox Simulation (STSS) benchmark, which assesses language agents objectively at the action level by scrutinizing the goal achievements within the multi-agent simulation.Additionally, we sample conversation scenarios to build a language-level benchmark to provide an economically prudent preliminary evaluation and align with prevailing benchmarks. To gauge the significance of agent architecture, we implement a target-driven planning (TDP) module as an adjunct to the existing agent. Our evaluative findings highlight that the STSS benchmark is challenging for state-of-the-art language agents. Furthermore, it effectively discriminates between distinct language agents, suggesting its usefulness as a benchmark for evaluating both language models and agent architectures. Our code is available at https://github.com/wcx21/Social-Tasks-in-Sandbox-Simulation. | [
"Wang, Chenxu",
"Dai, Bin",
"Liu, Huaping",
"Wang, Baoyuan"
]
| Towards Objectively Benchmarking Social Intelligence of Language Agents at the Action Level | findings-acl.526 | Poster | 2404.05337v1 |
https://aclanthology.org/2024.findings-acl.527.bib | @inproceedings{chen-etal-2024-semantic,
title = "Semantic Role Labeling from {C}hinese Speech via End-to-End Learning",
author = "Chen, Huiyao and
Li, Xinxin and
Zhang, Meishan and
Zhang, Min",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.527",
pages = "8898--8911",
abstract = "Semantic Role Labeling (SRL), crucial for understanding semantic relationships in sentences, has traditionally focused on text-based input. However, the increasing use of voice assistants and the need for hands-free interaction have highlighted the importance of SRL from speech.SRL from speech can be accomplished via a two-step pipeline directly: transcribing speech to text via Automatic Speech Recognition (ASR) and then applying text-based SRL, which could lead to error propagation and loss of useful acoustic features.Addressing these challenges, we present the first end-to-end approach for SRL from speech, integrating ASR and SRL in a joint-learning framework, focusing on the Chinese language. By employing a Stright-Through Gumbel-Softmax module for connecting ASR and SRL models, it enables gradient back-propagation and joint optimization, enhancing robustness and effectiveness.Experiments on the Chinese Proposition Bank 1.0 (CPB1.0) and a newly annotated dataset AS-SRL based on AISHELL-1 demonstrate the superiority of the end-to-end model over traditional pipelines, with significantly improved performance.",
}
| Semantic Role Labeling (SRL), crucial for understanding semantic relationships in sentences, has traditionally focused on text-based input. However, the increasing use of voice assistants and the need for hands-free interaction have highlighted the importance of SRL from speech.SRL from speech can be accomplished via a two-step pipeline directly: transcribing speech to text via Automatic Speech Recognition (ASR) and then applying text-based SRL, which could lead to error propagation and loss of useful acoustic features.Addressing these challenges, we present the first end-to-end approach for SRL from speech, integrating ASR and SRL in a joint-learning framework, focusing on the Chinese language. By employing a Stright-Through Gumbel-Softmax module for connecting ASR and SRL models, it enables gradient back-propagation and joint optimization, enhancing robustness and effectiveness.Experiments on the Chinese Proposition Bank 1.0 (CPB1.0) and a newly annotated dataset AS-SRL based on AISHELL-1 demonstrate the superiority of the end-to-end model over traditional pipelines, with significantly improved performance. | [
"Chen, Huiyao",
"Li, Xinxin",
"Zhang, Meishan",
"Zhang, Min"
]
| Semantic Role Labeling from {C}hinese Speech via End-to-End Learning | findings-acl.527 | Poster | 1911.04641v1 |
https://aclanthology.org/2024.findings-acl.528.bib | @inproceedings{tao-etal-2024-meel,
title = "{MEEL}: Multi-Modal Event Evolution Learning",
author = "Tao, Zhengwei and
Jin, Zhi and
Huang, Junqiang and
Chen, Xiancai and
Bai, Xiaoying and
Zhang, Yifan and
Tao, Chongyang",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.528",
pages = "8912--8925",
abstract = "Multi-modal Event Reasoning (MMER) endeavors to endow machines with the ability to comprehend intricate event relations across diverse data modalities. MMER is fundamental and underlies a wide broad of applications. Despite extensive instruction fine-tuning, current multi-modal large language models still fall short in such ability. The disparity stems from that existing models are insufficient to capture underlying principles governing event evolution in various scenarios. In this paper, we introduce Multi-Modal Event Evolution Learning (MEEL) to enable the model to grasp the event evolution mechanism yielding advanced MMER ability. Specifically, we commence with the design of event diversification to gather seed events from a rich spectrum of scenarios. Subsequently, we employ ChatGPT to generate evolving graphs for these seed events. We propose an instruction encapsulation process that formulates the evolving graphs into instruction-tuning data, aligning the comprehension of event reasoning to humans. Finally, we observe that models trained in this way are still struggling to fully comprehend event evolution. In such a case, we propose the guiding discrimination strategy, in which models are trained to discriminate the improper evolution direction. We collect and curate a benchmark M-EV2 for MMER. Extensive experiments on M-EV2 validate the effectiveness of our approach, showcasing competitive performance in open-source multi-modal LLMs.",
}
| Multi-modal Event Reasoning (MMER) endeavors to endow machines with the ability to comprehend intricate event relations across diverse data modalities. MMER is fundamental and underlies a wide broad of applications. Despite extensive instruction fine-tuning, current multi-modal large language models still fall short in such ability. The disparity stems from that existing models are insufficient to capture underlying principles governing event evolution in various scenarios. In this paper, we introduce Multi-Modal Event Evolution Learning (MEEL) to enable the model to grasp the event evolution mechanism yielding advanced MMER ability. Specifically, we commence with the design of event diversification to gather seed events from a rich spectrum of scenarios. Subsequently, we employ ChatGPT to generate evolving graphs for these seed events. We propose an instruction encapsulation process that formulates the evolving graphs into instruction-tuning data, aligning the comprehension of event reasoning to humans. Finally, we observe that models trained in this way are still struggling to fully comprehend event evolution. In such a case, we propose the guiding discrimination strategy, in which models are trained to discriminate the improper evolution direction. We collect and curate a benchmark M-EV2 for MMER. Extensive experiments on M-EV2 validate the effectiveness of our approach, showcasing competitive performance in open-source multi-modal LLMs. | [
"Tao, Zhengwei",
"Jin, Zhi",
"Huang, Junqiang",
"Chen, Xiancai",
"Bai, Xiaoying",
"Zhang, Yifan",
"Tao, Chongyang"
]
| {MEEL}: Multi-Modal Event Evolution Learning | findings-acl.528 | Poster | 2404.10429v1 |
https://aclanthology.org/2024.findings-acl.529.bib | @inproceedings{liang-etal-2024-llm,
title = "{LLM}-{REDIAL}: A Large-Scale Dataset for Conversational Recommender Systems Created from User Behaviors with {LLM}s",
author = "Liang, Tingting and
Jin, Chenxin and
Wang, Lingzhi and
Fan, Wenqi and
Xia, Congying and
Chen, Kai and
Yin, Yuyu",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.529",
pages = "8926--8939",
abstract = "The large-scale conversational recommendation dataset is pivotal for the development of conversational recommender systems (CRS). Most existing CRS datasets suffers from the problems of data inextensibility and semantic inconsistency. To tackle these limitations and establish a benchmark in the conversational recommendation scenario, in this paper, we introduce the LLM-REDIAL dataset to facilitate the research in CRS. LLM-REDIAL is constructed by leveraging large language models (LLMs) to generate the high-quality dialogues. To provide the LLMs with detailed guidance, we integrate historical user behavior data with dialogue templates that are carefully designed through the combination of multiple pre-defined goals. LLM-REDIAL has two main advantages. First, it is the largest multi-domain CRS dataset which consists of 47.6k multi-turn dialogues with 482.6k utterances across 4 domains. Second, dialogue semantics and the users{'} historical interaction information is highly consistent. Human evaluation are conducted to verify the quality of LLM-REDIAL. In addition, we evaluate the usability of advanced LLM-based models on LLM-REDIAL.",
}
| The large-scale conversational recommendation dataset is pivotal for the development of conversational recommender systems (CRS). Most existing CRS datasets suffers from the problems of data inextensibility and semantic inconsistency. To tackle these limitations and establish a benchmark in the conversational recommendation scenario, in this paper, we introduce the LLM-REDIAL dataset to facilitate the research in CRS. LLM-REDIAL is constructed by leveraging large language models (LLMs) to generate the high-quality dialogues. To provide the LLMs with detailed guidance, we integrate historical user behavior data with dialogue templates that are carefully designed through the combination of multiple pre-defined goals. LLM-REDIAL has two main advantages. First, it is the largest multi-domain CRS dataset which consists of 47.6k multi-turn dialogues with 482.6k utterances across 4 domains. Second, dialogue semantics and the users{'} historical interaction information is highly consistent. Human evaluation are conducted to verify the quality of LLM-REDIAL. In addition, we evaluate the usability of advanced LLM-based models on LLM-REDIAL. | [
"Liang, Tingting",
"Jin, Chenxin",
"Wang, Lingzhi",
"Fan, Wenqi",
"Xia, Congying",
"Chen, Kai",
"Yin, Yuyu"
]
| {LLM}-{REDIAL}: A Large-Scale Dataset for Conversational Recommender Systems Created from User Behaviors with {LLM}s | findings-acl.529 | Poster | 2308.16505v3 |
https://aclanthology.org/2024.findings-acl.530.bib | @inproceedings{kamruzzaman-etal-2024-investigating,
title = "Investigating Subtler Biases in {LLM}s: Ageism, Beauty, Institutional, and Nationality Bias in Generative Models",
author = "Kamruzzaman, Mahammed and
Shovon, Md. and
Kim, Gene",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.530",
pages = "8940--8965",
abstract = "LLMs are increasingly powerful and widely used to assist users in a variety of tasks. This use risks introducing LLM biases into consequential decisions such as job hiring, human performance evaluation, and criminal sentencing. Bias in NLP systems along the lines of gender and ethnicity has been widely studied, especially for specific stereotypes (e.g., Asians are good at math). In this paper, we investigate bias along less-studied but still consequential, dimensions, such as age and beauty, measuring subtler correlated decisions that LLMs make between social groups and unrelated positive and negative attributes. Although these subtler biases are understudied they follow people as much as gender and ethnicity do. So, we want to see whether they also follow one with LLMs.We introduce a template-generated dataset of sentence completion tasks that asks the model to select the most appropriate attribute to complete an evaluative statement about a person described as a member of a specific social group. We also reverse the completion task to select the social group based on an attribute. We report the correlations that we find for 4 cutting-edge LLMs. This dataset can be used as a benchmark to evaluate progress in more generalized biases and the templating technique can be used to expand the benchmark with minimal additional human annotation.",
}
| LLMs are increasingly powerful and widely used to assist users in a variety of tasks. This use risks introducing LLM biases into consequential decisions such as job hiring, human performance evaluation, and criminal sentencing. Bias in NLP systems along the lines of gender and ethnicity has been widely studied, especially for specific stereotypes (e.g., Asians are good at math). In this paper, we investigate bias along less-studied but still consequential, dimensions, such as age and beauty, measuring subtler correlated decisions that LLMs make between social groups and unrelated positive and negative attributes. Although these subtler biases are understudied they follow people as much as gender and ethnicity do. So, we want to see whether they also follow one with LLMs.We introduce a template-generated dataset of sentence completion tasks that asks the model to select the most appropriate attribute to complete an evaluative statement about a person described as a member of a specific social group. We also reverse the completion task to select the social group based on an attribute. We report the correlations that we find for 4 cutting-edge LLMs. This dataset can be used as a benchmark to evaluate progress in more generalized biases and the templating technique can be used to expand the benchmark with minimal additional human annotation. | [
"Kamruzzaman, Mahammed",
"Shovon, Md.",
"Kim, Gene"
]
| Investigating Subtler Biases in {LLM}s: Ageism, Beauty, Institutional, and Nationality Bias in Generative Models | findings-acl.530 | Poster | 2309.08902v3 |
https://aclanthology.org/2024.findings-acl.531.bib | @inproceedings{tao-etal-2024-evit,
title = "{EVIT}: Event-Oriented Instruction Tuning for Event Reasoning",
author = "Tao, Zhengwei and
Chen, Xiancai and
Jin, Zhi and
Bai, Xiaoying and
Zhao, Haiyan and
Lou, Yiwei",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.531",
pages = "8966--8979",
abstract = "Events refer to specific occurrences, incidents, or happenings that take place under a particular background. Event reasoning aims to infer events according to certain relations and predict future events. The cutting-edge techniques for event reasoning play a crucial role in various natural language processing applications. Large language models (LLMs) have made significant advancements in event reasoning owing to their wealth of knowledge and reasoning capabilities. However, smaller instruction-tuned models currently in use do not consistently demonstrate exceptional proficiency in managing these tasks. This discrepancy arises from the absence of explicit modeling of events and the interconnections of them within their instruction data. Consequently, these models face challenges in comprehending event structures and semantics while struggling to bridge the gap between their interpretations and human understanding of events. Additionally, their limitations in grasping event relations lead to constrained event reasoning abilities to effectively deduce and incorporate pertinent event knowledge. In this paper, we propose Event-Oriented Instruction Tuning to train our large language model named EvIT specializing in event reasoning tasks. Specifically, we first propose a novel structure named event quadruple which contains the structure and semantics of events and is complete in the event representation. We then design event-relation learning based on the structures. We encapsulate the learning into the instruction-tuning formulation to better stimulate the event reasoning capacity of our model. To implement our training, we design a heuristic unsupervised method to mine event quadruple from a large-scale corpus. At last, we finetune a Llama model on our Event-Oriented Instruction Tuning. We conduct extensive experiments on event reasoning tasks on several datasets. Automatic and human evaluations demonstrate EvIT achieves competitive performances on event reasoning.",
}
| Events refer to specific occurrences, incidents, or happenings that take place under a particular background. Event reasoning aims to infer events according to certain relations and predict future events. The cutting-edge techniques for event reasoning play a crucial role in various natural language processing applications. Large language models (LLMs) have made significant advancements in event reasoning owing to their wealth of knowledge and reasoning capabilities. However, smaller instruction-tuned models currently in use do not consistently demonstrate exceptional proficiency in managing these tasks. This discrepancy arises from the absence of explicit modeling of events and the interconnections of them within their instruction data. Consequently, these models face challenges in comprehending event structures and semantics while struggling to bridge the gap between their interpretations and human understanding of events. Additionally, their limitations in grasping event relations lead to constrained event reasoning abilities to effectively deduce and incorporate pertinent event knowledge. In this paper, we propose Event-Oriented Instruction Tuning to train our large language model named EvIT specializing in event reasoning tasks. Specifically, we first propose a novel structure named event quadruple which contains the structure and semantics of events and is complete in the event representation. We then design event-relation learning based on the structures. We encapsulate the learning into the instruction-tuning formulation to better stimulate the event reasoning capacity of our model. To implement our training, we design a heuristic unsupervised method to mine event quadruple from a large-scale corpus. At last, we finetune a Llama model on our Event-Oriented Instruction Tuning. We conduct extensive experiments on event reasoning tasks on several datasets. Automatic and human evaluations demonstrate EvIT achieves competitive performances on event reasoning. | [
"Tao, Zhengwei",
"Chen, Xiancai",
"Jin, Zhi",
"Bai, Xiaoying",
"Zhao, Haiyan",
"Lou, Yiwei"
]
| {EVIT}: Event-Oriented Instruction Tuning for Event Reasoning | findings-acl.531 | Poster | 2404.11978v1 |
https://aclanthology.org/2024.findings-acl.532.bib | @inproceedings{juseon-do-etal-2024-instructcmp,
title = "{I}nstruct{CMP}: Length Control in Sentence Compression through Instruction-based Large Language Models",
author = "Juseon-Do, Juseon-Do and
Kamigaito, Hidetaka and
Okumura, Manabu and
Kwon, Jingun",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.532",
pages = "8980--8996",
abstract = "Extractive summarization can produce faithful summaries but often requires additional constraints such as a desired summary length. Traditional sentence compression models do not typically consider the constraints because of their restricted model abilities, which require model modifications for coping with them. To bridge this gap, we propose Instruction-based Compression (InstructCMP), an approach to the sentence compression task that can consider the length constraint through instructions by leveraging the zero-shot task-solving abilities of Large Language Models (LLMs). For this purpose, we created new evaluation datasets by transforming traditional sentence compression datasets into an instruction format. By using the datasets, we first reveal that the current LLMs still face challenges in accurately controlling the length for a compressed text. To address this issue, we propose an approach named length priming, that incorporates additional length information into the instructions without external resources. While the length priming effectively works in a zero-shot setting, a training dataset with the instructions would further improve the ability of length control. Thus, we additionally created a training dataset in an instruction format to fine-tune the model on it. Experimental results and analysis show that applying the length priming significantly improves performances of InstructCMP in both zero-shot and fine-tuning settings without the need of any model modifications.",
}
| Extractive summarization can produce faithful summaries but often requires additional constraints such as a desired summary length. Traditional sentence compression models do not typically consider the constraints because of their restricted model abilities, which require model modifications for coping with them. To bridge this gap, we propose Instruction-based Compression (InstructCMP), an approach to the sentence compression task that can consider the length constraint through instructions by leveraging the zero-shot task-solving abilities of Large Language Models (LLMs). For this purpose, we created new evaluation datasets by transforming traditional sentence compression datasets into an instruction format. By using the datasets, we first reveal that the current LLMs still face challenges in accurately controlling the length for a compressed text. To address this issue, we propose an approach named length priming, that incorporates additional length information into the instructions without external resources. While the length priming effectively works in a zero-shot setting, a training dataset with the instructions would further improve the ability of length control. Thus, we additionally created a training dataset in an instruction format to fine-tune the model on it. Experimental results and analysis show that applying the length priming significantly improves performances of InstructCMP in both zero-shot and fine-tuning settings without the need of any model modifications. | [
"Juseon-Do, Juseon-Do",
"Kamigaito, Hidetaka",
"Okumura, Manabu",
"Kwon, Jingun"
]
| {I}nstruct{CMP}: Length Control in Sentence Compression through Instruction-based Large Language Models | findings-acl.532 | Poster | 2406.11097v2 |
https://aclanthology.org/2024.findings-acl.533.bib | @inproceedings{goyal-etal-2024-symtax,
title = "{S}ym{T}ax: Symbiotic Relationship and Taxonomy Fusion for Effective Citation Recommendation",
author = "Goyal, Karan and
Goel, Mayank and
Goyal, Vikram and
Mohania, Mukesh",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.533",
pages = "8997--9008",
abstract = "Citing pertinent literature is pivotal to writing and reviewing a scientific document. Existing techniques mainly focus on the local context or the global context for recommending citations but fail to consider the actual human citation behaviour. We propose SymTax, a three-stage recommendation architecture that considers both the local and the global context, and additionally the taxonomical representations of query-candidate tuples and the Symbiosis prevailing amongst them. SymTax learns to embed the infused taxonomies in the hyperbolic space and uses hyperbolic separation as a latent feature to compute query-candidate similarity. We build a novel and large dataset ArSyTa containing 8.27 million citation contexts and describe the creation process in detail. We conduct extensive experiments and ablation studies to demonstrate the effectiveness and design choice of each module in our framework. Also, combinatorial analysis from our experiments shed light on the choice of language models (LMs) and fusion embedding, and the inclusion of section heading as a signal. Our proposed module that captures the symbiotic relationship solely leads to performance gains of 26.66{\%} and 39.25{\%} in Recall@5 w.r.t. SOTA on ACL-200 and RefSeer datasets, respectively. The complete framework yields a gain of 22.56{\%} in Recall@5 wrt SOTA on our proposed dataset. The code and dataset are available at https://github.com/goyalkaraniit/SymTax.",
}
| Citing pertinent literature is pivotal to writing and reviewing a scientific document. Existing techniques mainly focus on the local context or the global context for recommending citations but fail to consider the actual human citation behaviour. We propose SymTax, a three-stage recommendation architecture that considers both the local and the global context, and additionally the taxonomical representations of query-candidate tuples and the Symbiosis prevailing amongst them. SymTax learns to embed the infused taxonomies in the hyperbolic space and uses hyperbolic separation as a latent feature to compute query-candidate similarity. We build a novel and large dataset ArSyTa containing 8.27 million citation contexts and describe the creation process in detail. We conduct extensive experiments and ablation studies to demonstrate the effectiveness and design choice of each module in our framework. Also, combinatorial analysis from our experiments shed light on the choice of language models (LMs) and fusion embedding, and the inclusion of section heading as a signal. Our proposed module that captures the symbiotic relationship solely leads to performance gains of 26.66{\%} and 39.25{\%} in Recall@5 w.r.t. SOTA on ACL-200 and RefSeer datasets, respectively. The complete framework yields a gain of 22.56{\%} in Recall@5 wrt SOTA on our proposed dataset. The code and dataset are available at https://github.com/goyalkaraniit/SymTax. | [
"Goyal, Karan",
"Goel, Mayank",
"Goyal, Vikram",
"Mohania, Mukesh"
]
| {S}ym{T}ax: Symbiotic Relationship and Taxonomy Fusion for Effective Citation Recommendation | findings-acl.533 | Poster | 2406.01606v1 |
https://aclanthology.org/2024.findings-acl.534.bib | @inproceedings{yoon-etal-2024-assessing,
title = "Assessing News Thumbnail Representativeness: Counterfactual text can enhance the cross-modal matching ability",
author = "Yoon, Yejun and
Yoon, Seunghyun and
Park, Kunwoo",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.534",
pages = "9009--9024",
abstract = "This paper addresses the critical challenge of assessing the representativeness of news thumbnail images, which often serve as the first visual engagement for readers when an article is disseminated on social media. We focus on whether a news image represents the actors discussed in the news text. To serve the challenge, we introduce NewsTT, a manually annotated dataset of 1000 news thumbnail images and text pairs. We found that the pretrained vision and language models, such as BLIP-2, struggle with this task. Since news subjects frequently involve named entities or proper nouns, the pretrained models could have a limited capability to match news actors{'} visual and textual appearances. We hypothesize that learning to contrast news text with its counterfactual, of which named entities are replaced, can enhance the cross-modal matching ability of vision and language models. We propose CFT-CLIP, a contrastive learning framework that updates vision and language bi-encoders according to the hypothesis. We found that our simple method can boost the performance for assessing news thumbnail representativeness, supporting our assumption. Code and data can be accessed at https://github.com/ssu-humane/news-images-acl24.",
}
| This paper addresses the critical challenge of assessing the representativeness of news thumbnail images, which often serve as the first visual engagement for readers when an article is disseminated on social media. We focus on whether a news image represents the actors discussed in the news text. To serve the challenge, we introduce NewsTT, a manually annotated dataset of 1000 news thumbnail images and text pairs. We found that the pretrained vision and language models, such as BLIP-2, struggle with this task. Since news subjects frequently involve named entities or proper nouns, the pretrained models could have a limited capability to match news actors{'} visual and textual appearances. We hypothesize that learning to contrast news text with its counterfactual, of which named entities are replaced, can enhance the cross-modal matching ability of vision and language models. We propose CFT-CLIP, a contrastive learning framework that updates vision and language bi-encoders according to the hypothesis. We found that our simple method can boost the performance for assessing news thumbnail representativeness, supporting our assumption. Code and data can be accessed at https://github.com/ssu-humane/news-images-acl24. | [
"Yoon, Yejun",
"Yoon, Seunghyun",
"Park, Kunwoo"
]
| Assessing News Thumbnail Representativeness: Counterfactual text can enhance the cross-modal matching ability | findings-acl.534 | Poster | 2402.11159v3 |
https://aclanthology.org/2024.findings-acl.535.bib | @inproceedings{hong-liu-2024-towards,
title = "Towards Better Question Generation in {QA}-based Event Extraction",
author = "Hong, Zijin and
Liu, Jian",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.535",
pages = "9025--9038",
abstract = "Event Extraction (EE) is an essential information extraction task that aims to extract event-related information from unstructured texts.The paradigm of this task has shifted from conventional classification-based methods to more contemporary question-answering-based (QA-based) approaches. However, in QA-based EE, the quality of the questions dramatically affects the extraction accuracy, and how to generate high-quality questions for QA-based EE remains a challenge. In this work, to tackle this challenge, we suggest four criteria to evaluate the quality of a question and propose a reinforcement learning method, RLQG, for QA-based EE that can generate generalizable, high-quality, and context-dependent questions and provides clear guidance to QA models. The extensive experiments conducted on ACE and RAMS datasets have strongly validated our approach{'}s effectiveness, which also demonstrates its robustness in scenarios with limited training data. The corresponding code of RLQG is released for further research.",
}
| Event Extraction (EE) is an essential information extraction task that aims to extract event-related information from unstructured texts.The paradigm of this task has shifted from conventional classification-based methods to more contemporary question-answering-based (QA-based) approaches. However, in QA-based EE, the quality of the questions dramatically affects the extraction accuracy, and how to generate high-quality questions for QA-based EE remains a challenge. In this work, to tackle this challenge, we suggest four criteria to evaluate the quality of a question and propose a reinforcement learning method, RLQG, for QA-based EE that can generate generalizable, high-quality, and context-dependent questions and provides clear guidance to QA models. The extensive experiments conducted on ACE and RAMS datasets have strongly validated our approach{'}s effectiveness, which also demonstrates its robustness in scenarios with limited training data. The corresponding code of RLQG is released for further research. | [
"Hong, Zijin",
"Liu, Jian"
]
| Towards Better Question Generation in {QA}-based Event Extraction | findings-acl.535 | Poster | 2405.10517v3 |
https://aclanthology.org/2024.findings-acl.536.bib | @inproceedings{zheng-etal-2024-budget,
title = "Budget-Constrained Tool Learning with Planning",
author = "Zheng, Yuanhang and
Li, Peng and
Yan, Ming and
Zhang, Ji and
Huang, Fei and
Liu, Yang",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.536",
pages = "9039--9052",
abstract = "Despite intensive efforts devoted to tool learning, the problem of budget-constrained tool learning, which focuses on resolving user queries within a specific budget constraint, has been widely overlooked. This paper proposes a novel method for budget-constrained tool learning. Our approach involves creating a preferable plan under the budget constraint before utilizing the tools. This plan outlines the feasible tools and the maximum number of times they can be employed, offering a comprehensive overview of the tool learning process for large language models. This allows them to allocate the budget from a broader perspective. To devise the plan without incurring significant extra costs, we suggest initially estimating the usefulness of the candidate tools based on past experience. Subsequently, we employ dynamic programming to formulate the plan. Experimental results demonstrate that our method can be integrated with various tool learning methods, significantly enhancing their effectiveness under strict budget constraints.",
}
| Despite intensive efforts devoted to tool learning, the problem of budget-constrained tool learning, which focuses on resolving user queries within a specific budget constraint, has been widely overlooked. This paper proposes a novel method for budget-constrained tool learning. Our approach involves creating a preferable plan under the budget constraint before utilizing the tools. This plan outlines the feasible tools and the maximum number of times they can be employed, offering a comprehensive overview of the tool learning process for large language models. This allows them to allocate the budget from a broader perspective. To devise the plan without incurring significant extra costs, we suggest initially estimating the usefulness of the candidate tools based on past experience. Subsequently, we employ dynamic programming to formulate the plan. Experimental results demonstrate that our method can be integrated with various tool learning methods, significantly enhancing their effectiveness under strict budget constraints. | [
"Zheng, Yuanhang",
"Li, Peng",
"Yan, Ming",
"Zhang, Ji",
"Huang, Fei",
"Liu, Yang"
]
| Budget-Constrained Tool Learning with Planning | findings-acl.536 | Poster | 2402.15960v2 |
https://aclanthology.org/2024.findings-acl.537.bib | @inproceedings{li-etal-2024-textbind,
title = "{T}ext{B}ind: Multi-turn Interleaved Multimodal Instruction-following in the Wild",
author = "Li, Huayang and
Li, Siheng and
Cai, Deng and
Wang, Longyue and
Liu, Lemao and
Watanabe, Taro and
Yang, Yujiu and
Shi, Shuming",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.537",
pages = "9053--9076",
abstract = "Large language models with instruction-following abilities have revolutionized the field of artificial intelligence. These models show exceptional generalizability to tackle various real-world tasks through their natural language interfaces. However, their performance heavily relies on high-quality exemplar data, which is often difficult to obtain. This challenge is further exacerbated when it comes to multimodal instruction following. We introduce TextBind, an almost annotation-free framework for empowering LLMs with multi-turn interleaved multimodal instruction-following capabilities. Our approach requires only image-caption pairs and generates multi-turn multimodal instruction-response conversations from a language model. To accommodate interleaved image-text inputs and outputs, we devise MIM, a language model-centric architecture that seamlessly integrates image encoder and decoder models. Extensive quantitative and qualitative experiments demonstrate that MIM trained on TextBind achieves remarkable generation capability in multimodal conversations compared to recent baselines.",
}
| Large language models with instruction-following abilities have revolutionized the field of artificial intelligence. These models show exceptional generalizability to tackle various real-world tasks through their natural language interfaces. However, their performance heavily relies on high-quality exemplar data, which is often difficult to obtain. This challenge is further exacerbated when it comes to multimodal instruction following. We introduce TextBind, an almost annotation-free framework for empowering LLMs with multi-turn interleaved multimodal instruction-following capabilities. Our approach requires only image-caption pairs and generates multi-turn multimodal instruction-response conversations from a language model. To accommodate interleaved image-text inputs and outputs, we devise MIM, a language model-centric architecture that seamlessly integrates image encoder and decoder models. Extensive quantitative and qualitative experiments demonstrate that MIM trained on TextBind achieves remarkable generation capability in multimodal conversations compared to recent baselines. | [
"Li, Huayang",
"Li, Siheng",
"Cai, Deng",
"Wang, Longyue",
"Liu, Lemao",
"Watanabe, Taro",
"Yang, Yujiu",
"Shi, Shuming"
]
| {T}ext{B}ind: Multi-turn Interleaved Multimodal Instruction-following in the Wild | findings-acl.537 | Poster | 2309.08637v5 |
https://aclanthology.org/2024.findings-acl.538.bib | @inproceedings{sun-etal-2024-critique,
title = "The Critique of Critique",
author = "Sun, Shichao and
Li, Junlong and
Yuan, Weizhe and
Yuan, Ruifeng and
Li, Wenjie and
Liu, Pengfei",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.538",
pages = "9077--9096",
abstract = "Critique, as a natural language description for assessing the quality of model-generated content, has played a vital role in the training, evaluation, and refinement of LLMs. However, a systematic method to evaluate the quality of critique is lacking. In this paper, we pioneer the critique of critique, termed $\textbf{MetaCritique}$, which builds specific quantification criteria. To achieve a reliable evaluation outcome, we propose Atomic Information Units (AIUs), which describe the critique in a more fine-grained manner. MetaCritique aggregates each AIU{'}s judgment for the overall score. Moreover, MetaCritique delivers a natural language rationale for the intricate reasoning within each judgment. Lastly, we construct a meta-evaluation dataset covering 4 tasks across 16 public datasets involving human-written and LLM-generated critiques. Experiments demonstrate that MetaCritique can achieve near-human performance. Our study can facilitate future research in LLM critiques based on our following observations and released resources: (1) superior critiques judged by MetaCritique can lead to better refinements, indicating that it can potentially enhance the alignment of existing LLMs; (2) the leaderboard of critique models reveals that open-source critique models commonly suffer from factuality issues; (3) relevant code and data are publicly available at https://anonymous.4open.science/r/MetaCritique-ARR/ to support deeper exploration; (4) an $\textbf{API}$ at PyPI with the usage documentation in Appendix C allows users to assess the critique conveniently.",
}
| Critique, as a natural language description for assessing the quality of model-generated content, has played a vital role in the training, evaluation, and refinement of LLMs. However, a systematic method to evaluate the quality of critique is lacking. In this paper, we pioneer the critique of critique, termed $\textbf{MetaCritique}$, which builds specific quantification criteria. To achieve a reliable evaluation outcome, we propose Atomic Information Units (AIUs), which describe the critique in a more fine-grained manner. MetaCritique aggregates each AIU{'}s judgment for the overall score. Moreover, MetaCritique delivers a natural language rationale for the intricate reasoning within each judgment. Lastly, we construct a meta-evaluation dataset covering 4 tasks across 16 public datasets involving human-written and LLM-generated critiques. Experiments demonstrate that MetaCritique can achieve near-human performance. Our study can facilitate future research in LLM critiques based on our following observations and released resources: (1) superior critiques judged by MetaCritique can lead to better refinements, indicating that it can potentially enhance the alignment of existing LLMs; (2) the leaderboard of critique models reveals that open-source critique models commonly suffer from factuality issues; (3) relevant code and data are publicly available at https://anonymous.4open.science/r/MetaCritique-ARR/ to support deeper exploration; (4) an $\textbf{API}$ at PyPI with the usage documentation in Appendix C allows users to assess the critique conveniently. | [
"Sun, Shichao",
"Li, Junlong",
"Yuan, Weizhe",
"Yuan, Ruifeng",
"Li, Wenjie",
"Liu, Pengfei"
]
| The Critique of Critique | findings-acl.538 | Poster | 2204.02162v1 |
https://aclanthology.org/2024.findings-acl.539.bib | @inproceedings{ma-etal-2024-coco,
title = "{C}o{C}o-Agent: A Comprehensive Cognitive {MLLM} Agent for Smartphone {GUI} Automation",
author = "Ma, Xinbei and
Zhang, Zhuosheng and
Zhao, Hai",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.539",
pages = "9097--9110",
abstract = "Multimodal large language models (MLLMs) have shown remarkable potential as human-like autonomous language agents to interact with real-world environments, especially for graphical user interface (GUI) automation.However, those GUI agents require comprehensive cognition including exhaustive perception and reliable action response.We propose a Comprehensive Cognitive LLM Agent, CoCo-Agent, with two novel approaches, comprehensive environment perception (CEP) and conditional action prediction (CAP), to systematically improve the GUI automation performance. First, CEP facilitates the GUI perception through different aspects and granularity, including screenshots and complementary detailed layouts for the visual channel and historical actions for the textual channel.Second, CAP decomposes the action prediction into sub-problems: determining the action type and then identifying the action target conditioned on the action type.With our technical design, our agent achieves state-of-the-art performance on AITW and META-GUI benchmarks, showing promising abilities in realistic scenarios. Code is available at \url{https://github.com/xbmxb/CoCo-Agent}.",
}
| Multimodal large language models (MLLMs) have shown remarkable potential as human-like autonomous language agents to interact with real-world environments, especially for graphical user interface (GUI) automation.However, those GUI agents require comprehensive cognition including exhaustive perception and reliable action response.We propose a Comprehensive Cognitive LLM Agent, CoCo-Agent, with two novel approaches, comprehensive environment perception (CEP) and conditional action prediction (CAP), to systematically improve the GUI automation performance. First, CEP facilitates the GUI perception through different aspects and granularity, including screenshots and complementary detailed layouts for the visual channel and historical actions for the textual channel.Second, CAP decomposes the action prediction into sub-problems: determining the action type and then identifying the action target conditioned on the action type.With our technical design, our agent achieves state-of-the-art performance on AITW and META-GUI benchmarks, showing promising abilities in realistic scenarios. Code is available at \url{https://github.com/xbmxb/CoCo-Agent}. | [
"Ma, Xinbei",
"Zhang, Zhuosheng",
"Zhao, Hai"
]
| {C}o{C}o-Agent: A Comprehensive Cognitive {MLLM} Agent for Smartphone {GUI} Automation | findings-acl.539 | Poster | 2402.11941v3 |
https://aclanthology.org/2024.findings-acl.540.bib | @inproceedings{fan-etal-2024-frva,
title = "{FRVA}: Fact-Retrieval and Verification Augmented Entailment Tree Generation for Explainable Question Answering",
author = "Fan, Yue and
Zhang, Hu and
Li, Ru and
Wang, YuJie and
Tan, Hongye and
Liang, Jiye",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.540",
pages = "9111--9128",
abstract = "Structured entailment tree can exhibit the reasoning chains from knowledge facts to predicted answers, which is important for constructing an explainable question answering system. Existing works mainly include directly generating the entire tree and stepwise generating the proof steps. The stepwise methods can exploit combinatoriality and generalize to longer steps, but they have large fact search spaces and error accumulation problems resulting in the generation of invalid steps. In this paper, inspired by the Dual Process Theory in cognitive science, we propose FRVA, a Fact-Retrieval and Verification Augmented bidirectional entailment tree generation method that contains two systems. Specifically, System 1 makes intuitive judgments through the fact retrieval module and filters irrelevant facts to reduce the search space. System 2 designs a deductive-abductive bidirectional reasoning module, and we construct cross-verification and multi-view contrastive learning to make the generated proof steps closer to the target hypothesis. We enhance the reliability of the stepwise proofs to mitigate error propagation. Experiment results on EntailmentBank show that FRVA outperforms previous models and achieves state-of-the-art performance in fact selection and structural correctness.",
}
| Structured entailment tree can exhibit the reasoning chains from knowledge facts to predicted answers, which is important for constructing an explainable question answering system. Existing works mainly include directly generating the entire tree and stepwise generating the proof steps. The stepwise methods can exploit combinatoriality and generalize to longer steps, but they have large fact search spaces and error accumulation problems resulting in the generation of invalid steps. In this paper, inspired by the Dual Process Theory in cognitive science, we propose FRVA, a Fact-Retrieval and Verification Augmented bidirectional entailment tree generation method that contains two systems. Specifically, System 1 makes intuitive judgments through the fact retrieval module and filters irrelevant facts to reduce the search space. System 2 designs a deductive-abductive bidirectional reasoning module, and we construct cross-verification and multi-view contrastive learning to make the generated proof steps closer to the target hypothesis. We enhance the reliability of the stepwise proofs to mitigate error propagation. Experiment results on EntailmentBank show that FRVA outperforms previous models and achieves state-of-the-art performance in fact selection and structural correctness. | [
"Fan, Yue",
"Zhang, Hu",
"Li, Ru",
"Wang, YuJie",
"Tan, Hongye",
"Liang, Jiye"
]
| {FRVA}: Fact-Retrieval and Verification Augmented Entailment Tree Generation for Explainable Question Answering | findings-acl.540 | Poster | 2205.02593v1 |
https://aclanthology.org/2024.findings-acl.541.bib | @inproceedings{zhang-etal-2024-p4,
title = "P4: Plug-and-Play Discrete Prompting for Large Language Models Personalization",
author = "Zhang, Yuansen and
Wang, Xiao and
Chen, Tianze and
Fu, Jiayi and
Gui, Tao and
Zhang, Qi",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.541",
pages = "9129--9144",
abstract = "Empowering Large Language Models (LLMs) with distinct human-like personality traits has become an innovative task for developing advanced dialog systems.Although LLMs demonstrate impressive capabilities in following instructions, directly prompting them to exhibit certain personalities through manually crafted instructions may result in sub-optimal performance.In this paper, we propose a plug-and-play prompting method to manipulate the LLMs{'} personality traits.Specifically, we append discrete personalized suffixes, automatically generated through an aggregated gradient-based search method, to the user query or dialog histories and induce LLMs to respond with target personalities.In addition, due to the high redundancy of the search space, we adopt a reward-based strategy to prune the vocabulary and focus exclusively on influential tokens.Experiment results on four models ranging from 1.1B to 13B show that our method achieves 79.9{\%} accuracy in customizing LLMs{'} personalities, significantly outperforming other prompting methods (65.5{\%}) and model editing methods.Our method also excels in generation fluency and quality with the lowest generation perplexity and the highest GPT-4 evaluation scores.",
}
| Empowering Large Language Models (LLMs) with distinct human-like personality traits has become an innovative task for developing advanced dialog systems.Although LLMs demonstrate impressive capabilities in following instructions, directly prompting them to exhibit certain personalities through manually crafted instructions may result in sub-optimal performance.In this paper, we propose a plug-and-play prompting method to manipulate the LLMs{'} personality traits.Specifically, we append discrete personalized suffixes, automatically generated through an aggregated gradient-based search method, to the user query or dialog histories and induce LLMs to respond with target personalities.In addition, due to the high redundancy of the search space, we adopt a reward-based strategy to prune the vocabulary and focus exclusively on influential tokens.Experiment results on four models ranging from 1.1B to 13B show that our method achieves 79.9{\%} accuracy in customizing LLMs{'} personalities, significantly outperforming other prompting methods (65.5{\%}) and model editing methods.Our method also excels in generation fluency and quality with the lowest generation perplexity and the highest GPT-4 evaluation scores. | [
"Zhang, Yuansen",
"Wang, Xiao",
"Chen, Tianze",
"Fu, Jiayi",
"Gui, Tao",
"Zhang, Qi"
]
| P4: Plug-and-Play Discrete Prompting for Large Language Models Personalization | findings-acl.541 | Poster | 2407.17115v1 |
https://aclanthology.org/2024.findings-acl.542.bib | @inproceedings{guo-etal-2024-large,
title = "Large Language Models Can Learn Representation in Natural Language",
author = "Guo, Yiduo and
Liang, Yaobo and
Zhao, Dongyan and
Duan, Nan",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.542",
pages = "9145--9154",
abstract = "One major challenge for Large Language Models (LLMs) is completing complex tasks involving multiple entities, such as tool APIs. To tackle this, one approach is to retrieve relevant entities to enhance LLMs in task completion. A crucial issue here is obtaining accurate natural language representations for each entity to aid in retriever precision. In this paper, we propose the Natural Language Representation Optimization Problem, which aims to refine entity descriptions for improved retrieval and LLM utilization. We introduce the Learning to Represent with Natural Language method, which utilizes LLMs to optimize entity representations consisting of text patterns based on environmental feedback. We iteratively prompt LLMs to enhance or adjust patterns based on entity samples and evaluate their effectiveness through environmental feedback. Our method successfully learns human-understandable representations for classification tasks (e.g., instructions and documents) and API call tasks (e.g., APIbench and Virtual Home), significantly improving GPT-4{'}s task performance.",
}
| One major challenge for Large Language Models (LLMs) is completing complex tasks involving multiple entities, such as tool APIs. To tackle this, one approach is to retrieve relevant entities to enhance LLMs in task completion. A crucial issue here is obtaining accurate natural language representations for each entity to aid in retriever precision. In this paper, we propose the Natural Language Representation Optimization Problem, which aims to refine entity descriptions for improved retrieval and LLM utilization. We introduce the Learning to Represent with Natural Language method, which utilizes LLMs to optimize entity representations consisting of text patterns based on environmental feedback. We iteratively prompt LLMs to enhance or adjust patterns based on entity samples and evaluate their effectiveness through environmental feedback. Our method successfully learns human-understandable representations for classification tasks (e.g., instructions and documents) and API call tasks (e.g., APIbench and Virtual Home), significantly improving GPT-4{'}s task performance. | [
"Guo, Yiduo",
"Liang, Yaobo",
"Zhao, Dongyan",
"Duan, Nan"
]
| Large Language Models Can Learn Representation in Natural Language | findings-acl.542 | Poster | 2012.14388v3 |
https://aclanthology.org/2024.findings-acl.543.bib | @inproceedings{fang-etal-2024-ctc,
title = "{CTC}-based Non-autoregressive Textless Speech-to-Speech Translation",
author = "Fang, Qingkai and
Ma, Zhengrui and
Zhou, Yan and
Zhang, Min and
Feng, Yang",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.543",
pages = "9155--9161",
abstract = "Direct speech-to-speech translation (S2ST) has achieved impressive translation quality, but it often faces the challenge of slow decoding due to the considerable length of speech sequences. Recently, some research has turned to non-autoregressive (NAR) models to expedite decoding, yet the translation quality typically lags behind autoregressive (AR) models significantly. In this paper, we investigate the performance of CTC-based NAR models in S2ST, as these models have shown impressive results in machine translation. Experimental results demonstrate that by combining pretraining, knowledge distillation, and advanced NAR training techniques such as glancing training and non-monotonic latent alignments, CTC-based NAR models achieve translation quality comparable to the AR model, while preserving up to 26.81$\times$ decoding speedup.",
}
| Direct speech-to-speech translation (S2ST) has achieved impressive translation quality, but it often faces the challenge of slow decoding due to the considerable length of speech sequences. Recently, some research has turned to non-autoregressive (NAR) models to expedite decoding, yet the translation quality typically lags behind autoregressive (AR) models significantly. In this paper, we investigate the performance of CTC-based NAR models in S2ST, as these models have shown impressive results in machine translation. Experimental results demonstrate that by combining pretraining, knowledge distillation, and advanced NAR training techniques such as glancing training and non-monotonic latent alignments, CTC-based NAR models achieve translation quality comparable to the AR model, while preserving up to 26.81$\times$ decoding speedup. | [
"Fang, Qingkai",
"Ma, Zhengrui",
"Zhou, Yan",
"Zhang, Min",
"Feng, Yang"
]
| {CTC}-based Non-autoregressive Textless Speech-to-Speech Translation | findings-acl.543 | Poster | 2211.00115v1 |
https://aclanthology.org/2024.findings-acl.544.bib | @inproceedings{fan-etal-2024-rrnorm,
title = "{RRN}orm: A Novel Framework for {C}hinese Disease Diagnoses Normalization via {LLM}-Driven Terminology Component Recognition and Reconstruction",
author = "Fan, Yongqi and
Zhu, Yansha and
Xue, Kui and
Liu, Jingping and
Ruan, Tong",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.544",
pages = "9162--9175",
abstract = "The Clinical Terminology Normalization aims at finding standard terms from a given termbase for mentions extracted from clinical texts. However, we found that extracted mentions suffer from the multi-implication problem, especially disease diagnoses. The reason for this is that physicians often use abbreviations, conjunctions, and juxtapositions when writing diagnoses, and it is difficult to manually decompose. To address this problem, we propose a Terminology Component Recognition and Reconstruction strategy that leverages the reasoning capability of large language models (LLMs) to recognize the components of terms, enabling automated decomposition and transforming original mentions into multiple atomic mentions. Furthermore, we adopt the mainstream {``}Recall and Rank{''} framework to apply the benefits of the above strategy to the task flow. By leveraging the LLM incorporating the advanced sampling strategies, we design a sampling algorithm for atomic mentions and train the recall model using contrastive learning. Besides the information about the components is also used as knowledge to guide the final term ranking and selection. The experimental results show that our proposed strategy effectively improves the performance of the terminology normalization task and our proposed approach achieves state-of-the-art on the experimental dataset. We release our code and data on the repository https://github.com/yuugaochyan/RRNorm.",
}
| The Clinical Terminology Normalization aims at finding standard terms from a given termbase for mentions extracted from clinical texts. However, we found that extracted mentions suffer from the multi-implication problem, especially disease diagnoses. The reason for this is that physicians often use abbreviations, conjunctions, and juxtapositions when writing diagnoses, and it is difficult to manually decompose. To address this problem, we propose a Terminology Component Recognition and Reconstruction strategy that leverages the reasoning capability of large language models (LLMs) to recognize the components of terms, enabling automated decomposition and transforming original mentions into multiple atomic mentions. Furthermore, we adopt the mainstream {``}Recall and Rank{''} framework to apply the benefits of the above strategy to the task flow. By leveraging the LLM incorporating the advanced sampling strategies, we design a sampling algorithm for atomic mentions and train the recall model using contrastive learning. Besides the information about the components is also used as knowledge to guide the final term ranking and selection. The experimental results show that our proposed strategy effectively improves the performance of the terminology normalization task and our proposed approach achieves state-of-the-art on the experimental dataset. We release our code and data on the repository https://github.com/yuugaochyan/RRNorm. | [
"Fan, Yongqi",
"Zhu, Yansha",
"Xue, Kui",
"Liu, Jingping",
"Ruan, Tong"
]
| {RRN}orm: A Novel Framework for {C}hinese Disease Diagnoses Normalization via {LLM}-Driven Terminology Component Recognition and Reconstruction | findings-acl.544 | Poster | 2101.09101v1 |
https://aclanthology.org/2024.findings-acl.545.bib | @inproceedings{zhang-etal-2024-unexpected,
title = "Unexpected Phenomenon: {LLM}s{'} Spurious Associations in Information Extraction",
author = "Zhang, Weiyan and
Lu, Wanpeng and
Wang, Jiacheng and
Wang, Yating and
Chen, Lihan and
Jiang, Haiyun and
Liu, Jingping and
Ruan, Tong",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.545",
pages = "9176--9190",
abstract = "Information extraction plays a critical role in natural language processing. When applying large language models (LLMs) to this domain, we discover an unexpected phenomenon: LLMs{'} spurious associations. In tasks such as relation extraction, LLMs can accurately identify entity pairs, even if the given relation (label) is semantically unrelated to the pre-defined original one. To find these labels, we design two strategies in this study, including forward label extension and backward label validation. We also leverage the extended labels to improve model performance. Our comprehensive experiments show that spurious associations occur consistently in both Chinese and English datasets across various LLM sizes. Moreover, the use of extended labels significantly enhances LLM performance in information extraction tasks. Remarkably, there is a performance increase of 9.55{\%}, 11.42{\%}, and 21.27{\%} in F1 scores on the SciERC, ACE05, and DuEE datasets, respectively.",
}
| Information extraction plays a critical role in natural language processing. When applying large language models (LLMs) to this domain, we discover an unexpected phenomenon: LLMs{'} spurious associations. In tasks such as relation extraction, LLMs can accurately identify entity pairs, even if the given relation (label) is semantically unrelated to the pre-defined original one. To find these labels, we design two strategies in this study, including forward label extension and backward label validation. We also leverage the extended labels to improve model performance. Our comprehensive experiments show that spurious associations occur consistently in both Chinese and English datasets across various LLM sizes. Moreover, the use of extended labels significantly enhances LLM performance in information extraction tasks. Remarkably, there is a performance increase of 9.55{\%}, 11.42{\%}, and 21.27{\%} in F1 scores on the SciERC, ACE05, and DuEE datasets, respectively. | [
"Zhang, Weiyan",
"Lu, Wanpeng",
"Wang, Jiacheng",
"Wang, Yating",
"Chen, Lihan",
"Jiang, Haiyun",
"Liu, Jingping",
"Ruan, Tong"
]
| Unexpected Phenomenon: {LLM}s{'} Spurious Associations in Information Extraction | findings-acl.545 | Poster | 2204.01138v1 |
https://aclanthology.org/2024.findings-acl.546.bib | @inproceedings{zhang-etal-2024-autocap,
title = "{A}uto{CAP}: Towards Automatic Cross-lingual Alignment Planning for Zero-shot Chain-of-Thought",
author = "Zhang, Yongheng and
Chen, Qiguang and
Li, Min and
Che, Wanxiang and
Qin, Libo",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.546",
pages = "9191--9200",
abstract = "Cross-lingual chain-of-thought can effectively complete reasoning tasks across languages, which gains increasing attention.Recently, dominant approaches in the literature improve cross-lingual alignment capabilities by integrating reasoning knowledge from different languages. Despite achieving excellent performance, current methods still have two main challenges: (1) Manual language specification: They still highly rely on manually selecting the languages to integrate, severely affecting their generalizability; (2) Static weight allocation: Current methods simply integrate all languages equally. In fact, different language reasoning paths should have different weights to achieve better complementation and integration. Motivated by this, we introduce an Automatic Cross-lingual Alignment Planning (AutoCAP) for zero-shot chain-of-thought to address the above challenges. The core of AutoCAP consists of two components: (1) Automatic Language Selection Prompting to guide LLMs to select appropriate languages and (2) Automatic Weight Allocation Prompting to automatically allocate alignment weight scores to each reasoning path. Extensive experiments on several benchmarks reveal that AutoCAP achieves state-of-the-art performance, surpassing previous methods that required manual effort.",
}
| Cross-lingual chain-of-thought can effectively complete reasoning tasks across languages, which gains increasing attention.Recently, dominant approaches in the literature improve cross-lingual alignment capabilities by integrating reasoning knowledge from different languages. Despite achieving excellent performance, current methods still have two main challenges: (1) Manual language specification: They still highly rely on manually selecting the languages to integrate, severely affecting their generalizability; (2) Static weight allocation: Current methods simply integrate all languages equally. In fact, different language reasoning paths should have different weights to achieve better complementation and integration. Motivated by this, we introduce an Automatic Cross-lingual Alignment Planning (AutoCAP) for zero-shot chain-of-thought to address the above challenges. The core of AutoCAP consists of two components: (1) Automatic Language Selection Prompting to guide LLMs to select appropriate languages and (2) Automatic Weight Allocation Prompting to automatically allocate alignment weight scores to each reasoning path. Extensive experiments on several benchmarks reveal that AutoCAP achieves state-of-the-art performance, surpassing previous methods that required manual effort. | [
"Zhang, Yongheng",
"Chen, Qiguang",
"Li, Min",
"Che, Wanxiang",
"Qin, Libo"
]
| {A}uto{CAP}: Towards Automatic Cross-lingual Alignment Planning for Zero-shot Chain-of-Thought | findings-acl.546 | Poster | 2406.13940v1 |
https://aclanthology.org/2024.findings-acl.547.bib | @inproceedings{sun-etal-2024-lcs,
title = "{LCS}: A Language Converter Strategy for Zero-Shot Neural Machine Translation",
author = "Sun, Zengkui and
Liu, Yijin and
Meng, Fandong and
Xu, Jinan and
Chen, Yufeng and
Zhou, Jie",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.547",
pages = "9201--9214",
abstract = "Multilingual neural machine translation models generally distinguish translation directions by the language tag (LT) in front of the source or target sentences. However, current LT strategies cannot indicate the desired target language as expected on zero-shot translation, i.e., the off-target issue. Our analysis reveals that the indication of the target language is sensitive to the placement of the target LT. For example, when placing the target LT on the decoder side, the indication would rapidly degrade along with decoding steps, while placing the target LT on the encoder side would lead to copying or paraphrasing the source input. To address the above issues, we propose a simple yet effective strategy named Language Converter Strategy (LCS). By introducing the target language embedding into the top encoder layers, LCS mitigates confusion in the encoder and ensures stable language indication for the decoder. Experimental results on MultiUN, TED, and OPUS-100 datasets demonstrate that LCS could significantly mitigate the off-target issue, with language accuracy up to 95.28{\%}, 96.21{\%}, and 85.35{\%} meanwhile outperforming the vanilla LT strategy by 3.07, 3,3, and 7.93 BLEU scores on zero-shot translation, respectively.",
}
| Multilingual neural machine translation models generally distinguish translation directions by the language tag (LT) in front of the source or target sentences. However, current LT strategies cannot indicate the desired target language as expected on zero-shot translation, i.e., the off-target issue. Our analysis reveals that the indication of the target language is sensitive to the placement of the target LT. For example, when placing the target LT on the decoder side, the indication would rapidly degrade along with decoding steps, while placing the target LT on the encoder side would lead to copying or paraphrasing the source input. To address the above issues, we propose a simple yet effective strategy named Language Converter Strategy (LCS). By introducing the target language embedding into the top encoder layers, LCS mitigates confusion in the encoder and ensures stable language indication for the decoder. Experimental results on MultiUN, TED, and OPUS-100 datasets demonstrate that LCS could significantly mitigate the off-target issue, with language accuracy up to 95.28{\%}, 96.21{\%}, and 85.35{\%} meanwhile outperforming the vanilla LT strategy by 3.07, 3,3, and 7.93 BLEU scores on zero-shot translation, respectively. | [
"Sun, Zengkui",
"Liu, Yijin",
"Meng, F",
"ong",
"Xu, Jinan",
"Chen, Yufeng",
"Zhou, Jie"
]
| {LCS}: A Language Converter Strategy for Zero-Shot Neural Machine Translation | findings-acl.547 | Poster | 2406.02876v2 |
https://aclanthology.org/2024.findings-acl.548.bib | @inproceedings{liu-etal-2024-llms,
title = "Are {LLM}s Capable of Data-based Statistical and Causal Reasoning? Benchmarking Advanced Quantitative Reasoning with Data",
author = "Liu, Xiao and
Wu, Zirui and
Wu, Xueqing and
Lu, Pan and
Chang, Kai-Wei and
Feng, Yansong",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.548",
pages = "9215--9235",
abstract = "Quantitative reasoning is a critical skill to analyze data, yet the assessment of such ability remains limited. To address this gap, we introduce the Quantitative Reasoning with Data (QRData) benchmark, aiming to evaluate Large Language Models{'} capability in statistical and causal reasoning with real-world data. The benchmark comprises a carefully constructed dataset of 411 questions accompanied by data sheets from textbooks, online learning materials, and academic papers. To compare models{'} quantitative reasoning abilities on data and text, we enrich the benchmark with an auxiliary set of 290 text-only questions, namely QRText. We evaluate natural language reasoning, program-based reasoning, and agent reasoning methods including Chain-of-Thought, Program-of-Thoughts, ReAct, and code interpreter assistants on diverse models. The strongest model GPT-4 achieves an accuracy of 58{\%}, which has much room for improvement. Among open-source models, Deepseek-coder-instruct, a code LLM pretrained on 2T tokens, gets the highest accuracy of 37{\%}. Analysis reveals that models encounter difficulties in data analysis and causal reasoning, and struggle in using causal knowledge and provided data simultaneously. Code and data are in https://github.com/xxxiaol/QRData.",
}
| Quantitative reasoning is a critical skill to analyze data, yet the assessment of such ability remains limited. To address this gap, we introduce the Quantitative Reasoning with Data (QRData) benchmark, aiming to evaluate Large Language Models{'} capability in statistical and causal reasoning with real-world data. The benchmark comprises a carefully constructed dataset of 411 questions accompanied by data sheets from textbooks, online learning materials, and academic papers. To compare models{'} quantitative reasoning abilities on data and text, we enrich the benchmark with an auxiliary set of 290 text-only questions, namely QRText. We evaluate natural language reasoning, program-based reasoning, and agent reasoning methods including Chain-of-Thought, Program-of-Thoughts, ReAct, and code interpreter assistants on diverse models. The strongest model GPT-4 achieves an accuracy of 58{\%}, which has much room for improvement. Among open-source models, Deepseek-coder-instruct, a code LLM pretrained on 2T tokens, gets the highest accuracy of 37{\%}. Analysis reveals that models encounter difficulties in data analysis and causal reasoning, and struggle in using causal knowledge and provided data simultaneously. Code and data are in https://github.com/xxxiaol/QRData. | [
"Liu, Xiao",
"Wu, Zirui",
"Wu, Xueqing",
"Lu, Pan",
"Chang, Kai-Wei",
"Feng, Yansong"
]
| Are {LLM}s Capable of Data-based Statistical and Causal Reasoning? Benchmarking Advanced Quantitative Reasoning with Data | findings-acl.548 | Poster | 2402.17644v2 |
https://aclanthology.org/2024.findings-acl.549.bib | @inproceedings{yi-etal-2024-vulnerability,
title = "On the Vulnerability of Safety Alignment in Open-Access {LLM}s",
author = "Yi, Jingwei and
Ye, Rui and
Chen, Qisi and
Zhu, Bin and
Chen, Siheng and
Lian, Defu and
Sun, Guangzhong and
Xie, Xing and
Wu, Fangzhao",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.549",
pages = "9236--9260",
abstract = "Large language models (LLMs) possess immense capabilities but are susceptible to malicious exploitation. To mitigate the risk, safety alignment is employed to align LLMs with ethical standards. However, safety-aligned LLMs may remain vulnerable to carefully crafted jailbreak attacks, but these attacks often face high rejection rates and limited harmfulness. In this paper, we expose the vulnerabilities of safety alignment in open-access LLMs, which can significantly enhance the success rate and harmfulness of jailbreak attacks. Through reverse alignment, achieved by accessing model parameters, we show the feasibility of efficiently fine-tuning LLMs to undermine their inherent safeguards. We investigate two types of reverse alignment techniques: reverse supervised fine-tuning (RSFT) and reverse preference optimization (RPO). RSFT operates by supervising the fine-tuning of LLMs to reverse their inherent values. We also explore how to prepare data needed for RSFT. RPO optimizes LLMs to enhance their preference for harmful content, reversing the models{'} safety alignment. Our extensive experiments reveal that open-access high-performance LLMs can be adeptly reverse-aligned to output harmful content, even in the absence of manually curated malicious datasets. Our research acts as a whistleblower for the community, emphasizing the need to pay more attention to safety of open-accessing LLMs. It also underscores the limitations of current safety alignment approaches and calls for research on robust safety alignment methods to counteract malicious fine-tuning attacks.",
}
| Large language models (LLMs) possess immense capabilities but are susceptible to malicious exploitation. To mitigate the risk, safety alignment is employed to align LLMs with ethical standards. However, safety-aligned LLMs may remain vulnerable to carefully crafted jailbreak attacks, but these attacks often face high rejection rates and limited harmfulness. In this paper, we expose the vulnerabilities of safety alignment in open-access LLMs, which can significantly enhance the success rate and harmfulness of jailbreak attacks. Through reverse alignment, achieved by accessing model parameters, we show the feasibility of efficiently fine-tuning LLMs to undermine their inherent safeguards. We investigate two types of reverse alignment techniques: reverse supervised fine-tuning (RSFT) and reverse preference optimization (RPO). RSFT operates by supervising the fine-tuning of LLMs to reverse their inherent values. We also explore how to prepare data needed for RSFT. RPO optimizes LLMs to enhance their preference for harmful content, reversing the models{'} safety alignment. Our extensive experiments reveal that open-access high-performance LLMs can be adeptly reverse-aligned to output harmful content, even in the absence of manually curated malicious datasets. Our research acts as a whistleblower for the community, emphasizing the need to pay more attention to safety of open-accessing LLMs. It also underscores the limitations of current safety alignment approaches and calls for research on robust safety alignment methods to counteract malicious fine-tuning attacks. | [
"Yi, Jingwei",
"Ye, Rui",
"Chen, Qisi",
"Zhu, Bin",
"Chen, Siheng",
"Lian, Defu",
"Sun, Guangzhong",
"Xie, Xing",
"Wu, Fangzhao"
]
| On the Vulnerability of Safety Alignment in Open-Access {LLM}s | findings-acl.549 | Poster | 2312.06924v2 |
https://aclanthology.org/2024.findings-acl.550.bib | @inproceedings{yang-etal-2024-pek,
title = "{PEK}: A Parameter-Efficient Framework for Knowledge-Grounded Dialogue Generation",
author = "Yang, Pan and
Song, Dandan and
Wu, Zhijing and
Zhou, Yanru",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.550",
pages = "9261--9273",
abstract = "Pre-trained language models (PLMs) have shown great dialogue generation capability in different scenarios. However, the huge VRAM consumption when fine-tuning them is one of their drawbacks. PEFT approaches can significantly reduce the number of trainable parameters, which enables us to fine-tune larger dialogue generation models. However, the reduction in parameter quantity can diminish a PLM{'}s expressive capacity and affect the PLM{'}s learning from certain specific examples like knowledge-related conversations. Previous works have demonstrated that injecting external knowledge into dialogue generation models can improve the model{'}s performance in knowledge-related conversations. Nonetheless, these methods are designed for the scenario where most parameters of the entire framework are trainable. In this paper, we propose PEK, a parameter-efficient framework for knowledge-enhanced dialogue generation. It enables PLMs to leverage external knowledge documents and knowledge graphs to enhance its generation capabilities with an acceptable number of trainable parameters. Evaluation results on the Wizard of Wikipedia and CMU{\_}DoG datasets show that our approach outperforms baseline methods on multiple evaluation metrics, which validates the effectiveness of our approach.",
}
| Pre-trained language models (PLMs) have shown great dialogue generation capability in different scenarios. However, the huge VRAM consumption when fine-tuning them is one of their drawbacks. PEFT approaches can significantly reduce the number of trainable parameters, which enables us to fine-tune larger dialogue generation models. However, the reduction in parameter quantity can diminish a PLM{'}s expressive capacity and affect the PLM{'}s learning from certain specific examples like knowledge-related conversations. Previous works have demonstrated that injecting external knowledge into dialogue generation models can improve the model{'}s performance in knowledge-related conversations. Nonetheless, these methods are designed for the scenario where most parameters of the entire framework are trainable. In this paper, we propose PEK, a parameter-efficient framework for knowledge-enhanced dialogue generation. It enables PLMs to leverage external knowledge documents and knowledge graphs to enhance its generation capabilities with an acceptable number of trainable parameters. Evaluation results on the Wizard of Wikipedia and CMU{\_}DoG datasets show that our approach outperforms baseline methods on multiple evaluation metrics, which validates the effectiveness of our approach. | [
"Yang, Pan",
"Song, D",
"an",
"Wu, Zhijing",
"Zhou, Yanru"
]
| {PEK}: A Parameter-Efficient Framework for Knowledge-Grounded Dialogue Generation | findings-acl.550 | Poster | 2312.16864v1 |
https://aclanthology.org/2024.findings-acl.551.bib | @inproceedings{zheng-etal-2024-evidence,
title = "Evidence Retrieval is almost All You Need for Fact Verification",
author = "Zheng, Liwen and
Li, Chaozhuo and
Zhang, Xi and
Shang, Yu-Ming and
Huang, Feiran and
Jia, Haoran",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.551",
pages = "9274--9281",
abstract = "Current fact verification methods generally follow the two-stage training paradigm: evidence retrieval and claim verification. While existing works focus on developing sophisticated claim verification modules, the fundamental importance of evidence retrieval is largely ignored. Existing approaches usually adopt the heuristic semantic similarity-based retrieval strategy, resulting in the task-irrelevant evidence and undesirable performance. In this paper, we concentrate on evidence retrieval and propose a Retrieval-Augmented Verification framework RAV, consisting of two major modules: the hybrid evidence retrieval and the joint fact verification. Hybrid evidence retrieval module incorporates an efficient retriever for preliminary pruning of candidate evidence, succeeded by a ranker that generates more precise sorting results. Under this end-to-end training paradigm, gradients from the claim verification can be back-propagated to enhance evidence selection. Experimental results on FEVER dataset demonstrate the superiority of RAV.",
}
| Current fact verification methods generally follow the two-stage training paradigm: evidence retrieval and claim verification. While existing works focus on developing sophisticated claim verification modules, the fundamental importance of evidence retrieval is largely ignored. Existing approaches usually adopt the heuristic semantic similarity-based retrieval strategy, resulting in the task-irrelevant evidence and undesirable performance. In this paper, we concentrate on evidence retrieval and propose a Retrieval-Augmented Verification framework RAV, consisting of two major modules: the hybrid evidence retrieval and the joint fact verification. Hybrid evidence retrieval module incorporates an efficient retriever for preliminary pruning of candidate evidence, succeeded by a ranker that generates more precise sorting results. Under this end-to-end training paradigm, gradients from the claim verification can be back-propagated to enhance evidence selection. Experimental results on FEVER dataset demonstrate the superiority of RAV. | [
"Zheng, Liwen",
"Li, Chaozhuo",
"Zhang, Xi",
"Shang, Yu-Ming",
"Huang, Feiran",
"Jia, Haoran"
]
| Evidence Retrieval is almost All You Need for Fact Verification | findings-acl.551 | Poster | 2407.15291v1 |
https://aclanthology.org/2024.findings-acl.552.bib | @inproceedings{sun-etal-2024-outdated,
title = "Outdated Issue Aware Decoding for Factual Knowledge Editing",
author = "Sun, Zengkui and
Liu, Yijin and
Wang, Jiaan and
Meng, Fandong and
Xu, Jinan and
Chen, Yufeng and
Zhou, Jie",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.552",
pages = "9282--9293",
abstract = "Recently, Knowledge Editing has received increasing attention, since it could update the specific knowledge from outdated ones in pretrained models without re-training. However, as pointed out by recent studies, existing related methods tend to merely memorize the superficial word composition of the edited knowledge, rather than truly learning and absorbing it. Consequently, on the reasoning questions, we discover that existing methods struggle to utilize the edited knowledge to reason the new answer, and tend to retain outdated responses, which are generated by the original models utilizing original knowledge. Nevertheless, the outdated responses are unexpected for the correct answers to reasoning questions, which we named as the outdated issue. To alleviate this issue, in this paper, we propose a simple yet effective decoding strategy, i.e., outDated ISsue aware deCOding (DISCO), to enhance the performance of edited models on reasoning questions. Specifically, we capture the difference in the probability distribution between the original and edited models. Further, we amplify the difference of the token prediction in the edited model to alleviate the outdated issue, and thus enhance the model performance w.r.t the edited knowledge. Experimental results suggest that applying DISCO could enhance edited models to reason, e.g., on reasoning questions, DISCO outperforms the prior SOTA method by 12.99 F1 scores, and reduces the ratio of the outdated issue to 5.78{\%} on the zsRE dataset.",
}
| Recently, Knowledge Editing has received increasing attention, since it could update the specific knowledge from outdated ones in pretrained models without re-training. However, as pointed out by recent studies, existing related methods tend to merely memorize the superficial word composition of the edited knowledge, rather than truly learning and absorbing it. Consequently, on the reasoning questions, we discover that existing methods struggle to utilize the edited knowledge to reason the new answer, and tend to retain outdated responses, which are generated by the original models utilizing original knowledge. Nevertheless, the outdated responses are unexpected for the correct answers to reasoning questions, which we named as the outdated issue. To alleviate this issue, in this paper, we propose a simple yet effective decoding strategy, i.e., outDated ISsue aware deCOding (DISCO), to enhance the performance of edited models on reasoning questions. Specifically, we capture the difference in the probability distribution between the original and edited models. Further, we amplify the difference of the token prediction in the edited model to alleviate the outdated issue, and thus enhance the model performance w.r.t the edited knowledge. Experimental results suggest that applying DISCO could enhance edited models to reason, e.g., on reasoning questions, DISCO outperforms the prior SOTA method by 12.99 F1 scores, and reduces the ratio of the outdated issue to 5.78{\%} on the zsRE dataset. | [
"Sun, Zengkui",
"Liu, Yijin",
"Wang, Jiaan",
"Meng, F",
"ong",
"Xu, Jinan",
"Chen, Yufeng",
"Zhou, Jie"
]
| Outdated Issue Aware Decoding for Factual Knowledge Editing | findings-acl.552 | Poster | 2406.02882v3 |
https://aclanthology.org/2024.findings-acl.553.bib | @inproceedings{spliethover-etal-2024-disentangling,
title = "Disentangling Dialect from Social Bias via Multitask Learning to Improve Fairness",
author = {Splieth{\"o}ver, Maximilian and
Menon, Sai Nikhil and
Wachsmuth, Henning},
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.553",
pages = "9294--9313",
abstract = "Dialects introduce syntactic and lexical variations in language that occur in regional or social groups. Most NLP methods are not sensitive to such variations. This may lead to unfair behavior of the methods, conveying negative bias towards dialect speakers. While previous work has studied dialect-related fairness for aspects like hate speech, other aspects of biased language, such as lewdness, remain fully unexplored. To fill this gap, we investigate performance disparities between dialects in the detection of five aspects of biased language and how to mitigate them. To alleviate bias, we present a multitask learning approach that models dialect language as an auxiliary task to incorporate syntactic and lexical variations. In our experiments with African-American English dialect, we provide empirical evidence that complementing common learning approaches with dialect modeling improves their fairness. Furthermore, the results suggest that multitask learning achieves state-of-the-art performance and helps to detect properties of biased language more reliably.",
}
| Dialects introduce syntactic and lexical variations in language that occur in regional or social groups. Most NLP methods are not sensitive to such variations. This may lead to unfair behavior of the methods, conveying negative bias towards dialect speakers. While previous work has studied dialect-related fairness for aspects like hate speech, other aspects of biased language, such as lewdness, remain fully unexplored. To fill this gap, we investigate performance disparities between dialects in the detection of five aspects of biased language and how to mitigate them. To alleviate bias, we present a multitask learning approach that models dialect language as an auxiliary task to incorporate syntactic and lexical variations. In our experiments with African-American English dialect, we provide empirical evidence that complementing common learning approaches with dialect modeling improves their fairness. Furthermore, the results suggest that multitask learning achieves state-of-the-art performance and helps to detect properties of biased language more reliably. | [
"Splieth{\\\"o}ver, Maximilian",
"Menon, Sai Nikhil",
"Wachsmuth, Henning"
]
| Disentangling Dialect from Social Bias via Multitask Learning to Improve Fairness | findings-acl.553 | Poster | 2406.09977v1 |
https://aclanthology.org/2024.findings-acl.554.bib | @inproceedings{meisenbacher-etal-2024-dp,
title = "{DP}-{MLM}: Differentially Private Text Rewriting Using Masked Language Models",
author = "Meisenbacher, Stephen and
Chevli, Maulik and
Vladika, Juraj and
Matthes, Florian",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.554",
pages = "9314--9328",
abstract = "The task of text privatization using Differential Privacy has recently taken the form of $\textit{text rewriting}$, in which an input text is obfuscated via the use of generative (large) language models. While these methods have shown promising results in the ability to preserve privacy, these methods rely on autoregressive models which lack a mechanism to contextualize the private rewriting process. In response to this, we propose $\textbf{DP-MLM}$, a new method for differentially private text rewriting based on leveraging masked language models (MLMs) to rewrite text in a semantically similar $\textit{and}$ obfuscated manner. We accomplish this with a simple contextualization technique, whereby we rewrite a text one token at a time. We find that utilizing encoder-only MLMs provides better utility preservation at lower $\varepsilon$ levels, as compared to previous methods relying on larger models with a decoder. In addition, MLMs allow for greater customization of the rewriting mechanism, as opposed to generative approaches. We make the code for $\textbf{DP-MLM}$ public and reusable, found at https://github.com/sjmeis/DPMLM.",
}
| The task of text privatization using Differential Privacy has recently taken the form of $\textit{text rewriting}$, in which an input text is obfuscated via the use of generative (large) language models. While these methods have shown promising results in the ability to preserve privacy, these methods rely on autoregressive models which lack a mechanism to contextualize the private rewriting process. In response to this, we propose $\textbf{DP-MLM}$, a new method for differentially private text rewriting based on leveraging masked language models (MLMs) to rewrite text in a semantically similar $\textit{and}$ obfuscated manner. We accomplish this with a simple contextualization technique, whereby we rewrite a text one token at a time. We find that utilizing encoder-only MLMs provides better utility preservation at lower $\varepsilon$ levels, as compared to previous methods relying on larger models with a decoder. In addition, MLMs allow for greater customization of the rewriting mechanism, as opposed to generative approaches. We make the code for $\textbf{DP-MLM}$ public and reusable, found at https://github.com/sjmeis/DPMLM. | [
"Meisenbacher, Stephen",
"Chevli, Maulik",
"Vladika, Juraj",
"Matthes, Florian"
]
| {DP}-{MLM}: Differentially Private Text Rewriting Using Masked Language Models | findings-acl.554 | Poster | 2407.00637v1 |
https://aclanthology.org/2024.findings-acl.555.bib | @inproceedings{mogrovejo-solorio-2024-question,
title = "Question-Instructed Visual Descriptions for Zero-Shot Video Answering",
author = "Mogrovejo, David and
Solorio, Thamar",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.555",
pages = "9329--9339",
abstract = "We present Q-ViD, a simple approach for video question answering (video QA), that unlike prior methods, which are based on complex architectures, computationally expensive pipelines or use closed models like GPTs, Q-ViD relies on a single instruction-aware open vision-language model (InstructBLIP) to tackle videoQA using frame descriptions. Specifically, we create captioning instruction prompts that rely on the target questions about the videos and leverage InstructBLIP to obtain video frame captions that are useful to the task at hand. Subsequently, we form descriptions of the whole video using the question-dependent frame captions, and feed that information, along with a question-answering prompt, to a large language model (LLM). The LLM is our reasoning module, and performs the final step of multiple-choice QA. Our simple Q-ViD framework achieves competitive or even higher performances than current state of the art models on a diverse range of videoQA benchmarks, including NExT-QA, STAR, How2QA, TVQA and IntentQA.",
}
| We present Q-ViD, a simple approach for video question answering (video QA), that unlike prior methods, which are based on complex architectures, computationally expensive pipelines or use closed models like GPTs, Q-ViD relies on a single instruction-aware open vision-language model (InstructBLIP) to tackle videoQA using frame descriptions. Specifically, we create captioning instruction prompts that rely on the target questions about the videos and leverage InstructBLIP to obtain video frame captions that are useful to the task at hand. Subsequently, we form descriptions of the whole video using the question-dependent frame captions, and feed that information, along with a question-answering prompt, to a large language model (LLM). The LLM is our reasoning module, and performs the final step of multiple-choice QA. Our simple Q-ViD framework achieves competitive or even higher performances than current state of the art models on a diverse range of videoQA benchmarks, including NExT-QA, STAR, How2QA, TVQA and IntentQA. | [
"Mogrovejo, David",
"Solorio, Thamar"
]
| Question-Instructed Visual Descriptions for Zero-Shot Video Answering | findings-acl.555 | Poster | 2304.04227v3 |
https://aclanthology.org/2024.findings-acl.556.bib | @inproceedings{ma-etal-2024-ex,
title = "{EX}-{FEVER}: A Dataset for Multi-hop Explainable Fact Verification",
author = "Ma, Huanhuan and
Xu, Weizhi and
Wei, Yifan and
Chen, Liuji and
Wang, Liang and
Liu, Qiang and
Wu, Shu and
Wang, Liang",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.556",
pages = "9340--9353",
abstract = "Fact verification aims to automatically probe the veracity of a claim based on several pieces of evidence. Existing works are always engaging in accuracy improvement, let alone explainability, a critical capability of fact verification systems.Constructing an explainable fact verification system in a complex multi-hop scenario is consistently impeded by the absence of a relevant, high-quality dataset. Previous datasets either suffer from excessive simplification or fail to incorporate essential considerations for explainability. To address this, we present EX-FEVER, a pioneering dataset for multi-hop explainable fact verification. With over 60,000 claims involving 2-hop and 3-hop reasoning, each is created by summarizing and modifying information from hyperlinked Wikipedia documents. Each instance is accompanied by a veracity label and an explanation that outlines the reasoning path supporting the veracity classification. Additionally, we demonstrate a novel baseline system on our EX-FEVER dataset, showcasing document retrieval, explanation generation, and claim verification, and validate the significance of our dataset. Furthermore, we highlight the potential of utilizing Large Language Models in the fact verification task. We hope our dataset could make a significant contribution by providing ample opportunities to explore the integration of natural language explanations in the domain of fact verification.",
}
| Fact verification aims to automatically probe the veracity of a claim based on several pieces of evidence. Existing works are always engaging in accuracy improvement, let alone explainability, a critical capability of fact verification systems.Constructing an explainable fact verification system in a complex multi-hop scenario is consistently impeded by the absence of a relevant, high-quality dataset. Previous datasets either suffer from excessive simplification or fail to incorporate essential considerations for explainability. To address this, we present EX-FEVER, a pioneering dataset for multi-hop explainable fact verification. With over 60,000 claims involving 2-hop and 3-hop reasoning, each is created by summarizing and modifying information from hyperlinked Wikipedia documents. Each instance is accompanied by a veracity label and an explanation that outlines the reasoning path supporting the veracity classification. Additionally, we demonstrate a novel baseline system on our EX-FEVER dataset, showcasing document retrieval, explanation generation, and claim verification, and validate the significance of our dataset. Furthermore, we highlight the potential of utilizing Large Language Models in the fact verification task. We hope our dataset could make a significant contribution by providing ample opportunities to explore the integration of natural language explanations in the domain of fact verification. | [
"Ma, Huanhuan",
"Xu, Weizhi",
"Wei, Yifan",
"Chen, Liuji",
"Wang, Liang",
"Liu, Qiang",
"Wu, Shu",
"Wang, Liang"
]
| {EX}-{FEVER}: A Dataset for Multi-hop Explainable Fact Verification | findings-acl.556 | Poster | 2310.09754v3 |
https://aclanthology.org/2024.findings-acl.557.bib | @inproceedings{chen-etal-2024-agent,
title = "Agent-{FLAN}: Designing Data and Methods of Effective Agent Tuning for Large Language Models",
author = "Chen, Zehui and
Liu, Kuikun and
Wang, Qiuchen and
Zhang, Wenwei and
Liu, Jiangning and
Lin, Dahua and
Chen, Kai and
Zhao, Feng",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.557",
pages = "9354--9366",
abstract = "Open-sourced Large Language Models (LLMs) have achieved great success in various NLP tasks, however, they are still far inferior to API-based models when acting as agents. How to integrate agent ability into general LLMs becomes a crucial and urgent problem.This paper first delivers three key observations: (1) the current agent training corpus is entangled with both formats following and agent reasoning, which significantly shifts from the distribution of its pre-training data; (2) LLMs exhibit different learning speeds on the capabilities required by agent tasks; and (3) current approaches have side-effects when improving agent abilities by introducing hallucinations. Based on the above findings, we propose Agent-FLAN to effectively Fine-tune LANguage models for Agents.Through careful decomposition and redesign of the training corpus, Agent-FLAN enables Llama2-7B to outperform prior best works by 3.5{\%} across various agent evaluation datasets. With comprehensively constructed negative samples, Agent-FLAN greatly alleviates the hallucination issues based on our established evaluation benchmark. Besides, it consistently improves the agent capability of LLMs when scaling model sizes while slightly enhancing the general capability of LLMs. The code and models are available at https://github.com/InternLM/Agent-FLAN.",
}
| Open-sourced Large Language Models (LLMs) have achieved great success in various NLP tasks, however, they are still far inferior to API-based models when acting as agents. How to integrate agent ability into general LLMs becomes a crucial and urgent problem.This paper first delivers three key observations: (1) the current agent training corpus is entangled with both formats following and agent reasoning, which significantly shifts from the distribution of its pre-training data; (2) LLMs exhibit different learning speeds on the capabilities required by agent tasks; and (3) current approaches have side-effects when improving agent abilities by introducing hallucinations. Based on the above findings, we propose Agent-FLAN to effectively Fine-tune LANguage models for Agents.Through careful decomposition and redesign of the training corpus, Agent-FLAN enables Llama2-7B to outperform prior best works by 3.5{\%} across various agent evaluation datasets. With comprehensively constructed negative samples, Agent-FLAN greatly alleviates the hallucination issues based on our established evaluation benchmark. Besides, it consistently improves the agent capability of LLMs when scaling model sizes while slightly enhancing the general capability of LLMs. The code and models are available at https://github.com/InternLM/Agent-FLAN. | [
"Chen, Zehui",
"Liu, Kuikun",
"Wang, Qiuchen",
"Zhang, Wenwei",
"Liu, Jiangning",
"Lin, Dahua",
"Chen, Kai",
"Zhao, Feng"
]
| Agent-{FLAN}: Designing Data and Methods of Effective Agent Tuning for Large Language Models | findings-acl.557 | Poster | 2403.12881v1 |
https://aclanthology.org/2024.findings-acl.558.bib | @inproceedings{fadeeva-etal-2024-fact,
title = "Fact-Checking the Output of Large Language Models via Token-Level Uncertainty Quantification",
author = "Fadeeva, Ekaterina and
Rubashevskii, Aleksandr and
Shelmanov, Artem and
Petrakov, Sergey and
Li, Haonan and
Mubarak, Hamdy and
Tsymbalov, Evgenii and
Kuzmin, Gleb and
Panchenko, Alexander and
Baldwin, Timothy and
Nakov, Preslav and
Panov, Maxim",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.558",
pages = "9367--9385",
abstract = "Large language models (LLMs) are notorious for hallucinating, i.e., producing erroneous claims in their output. Such hallucinations can be dangerous, as occasional factual inaccuracies in the generated text might be obscured by the rest of the output being generally factually correct, making it extremely hard for the users to spot them. Current services that leverage LLMs usually do not provide any means for detecting unreliable generations. Here, we aim to bridge this gap. In particular, we propose a novel fact-checking and hallucination detection pipeline based on token-level uncertainty quantification. Uncertainty scores leverage information encapsulated in the output of a neural network or its layers to detect unreliable predictions, and we show that they can be used to fact-check the atomic claims in the LLM output. Moreover, we present a novel token-level uncertainty quantification method that removes the impact of uncertainty about what claim to generate on the current step and what surface form to use. Our method Claim Conditioned Probability (CCP) measures only the uncertainty of a particular claim value expressed by the model. Experiments on the task of biography generation demonstrate strong improvements for CCP compared to the baselines for seven different LLMs and four languages. Human evaluation reveals that the fact-checking pipeline based on uncertainty quantification is competitive with a fact-checking tool that leverages external knowledge.",
}
| Large language models (LLMs) are notorious for hallucinating, i.e., producing erroneous claims in their output. Such hallucinations can be dangerous, as occasional factual inaccuracies in the generated text might be obscured by the rest of the output being generally factually correct, making it extremely hard for the users to spot them. Current services that leverage LLMs usually do not provide any means for detecting unreliable generations. Here, we aim to bridge this gap. In particular, we propose a novel fact-checking and hallucination detection pipeline based on token-level uncertainty quantification. Uncertainty scores leverage information encapsulated in the output of a neural network or its layers to detect unreliable predictions, and we show that they can be used to fact-check the atomic claims in the LLM output. Moreover, we present a novel token-level uncertainty quantification method that removes the impact of uncertainty about what claim to generate on the current step and what surface form to use. Our method Claim Conditioned Probability (CCP) measures only the uncertainty of a particular claim value expressed by the model. Experiments on the task of biography generation demonstrate strong improvements for CCP compared to the baselines for seven different LLMs and four languages. Human evaluation reveals that the fact-checking pipeline based on uncertainty quantification is competitive with a fact-checking tool that leverages external knowledge. | [
"Fadeeva, Ekaterina",
"Rubashevskii, Aleks",
"r",
"Shelmanov, Artem",
"Petrakov, Sergey",
"Li, Haonan",
"Mubarak, Hamdy",
"Tsymbalov, Evgenii",
"Kuzmin, Gleb",
"Panchenko, Alex",
"er",
"Baldwin, Timothy",
"Nakov, Preslav",
"Panov, Maxim"
]
| Fact-Checking the Output of Large Language Models via Token-Level Uncertainty Quantification | findings-acl.558 | Poster | 2406.19712v1 |
https://aclanthology.org/2024.findings-acl.559.bib | @inproceedings{zhao-etal-2024-deciphering,
title = "Deciphering the Impact of Pretraining Data on Large Language Models through Machine Unlearning",
author = "Zhao, Yang and
Du, Li and
Ding, Xiao and
Xiong, Kai and
Sun, Zhouhao and
Jun, Shi and
Liu, Ting and
Qin, Bing",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.559",
pages = "9386--9406",
abstract = "Through pretraining on a corpus with various sources, Large Language Models (LLMs) have gained impressive performance. However, the impact of each component of the pretraining corpus remains opaque. As a result, the organization of the pretraining corpus is still empirical and may deviate from the optimal. To address this issue, we systematically analyze the impact of 48 datasets from 5 major categories of pretraining data of LLMs and measure their impacts on LLMs using benchmarks about nine major categories of model capabilities. Our analyses provide empirical results about the contribution of multiple corpora on the performances of LLMs, along with their joint impact patterns, including complementary, orthogonal, and correlational relationships. We also identify a set of {``}high-impact data{''} such as Books that is significantly related to a set of model capabilities. These findings provide insights into the organization of data to support more efficient pretraining of LLMs.",
}
| Through pretraining on a corpus with various sources, Large Language Models (LLMs) have gained impressive performance. However, the impact of each component of the pretraining corpus remains opaque. As a result, the organization of the pretraining corpus is still empirical and may deviate from the optimal. To address this issue, we systematically analyze the impact of 48 datasets from 5 major categories of pretraining data of LLMs and measure their impacts on LLMs using benchmarks about nine major categories of model capabilities. Our analyses provide empirical results about the contribution of multiple corpora on the performances of LLMs, along with their joint impact patterns, including complementary, orthogonal, and correlational relationships. We also identify a set of {``}high-impact data{''} such as Books that is significantly related to a set of model capabilities. These findings provide insights into the organization of data to support more efficient pretraining of LLMs. | [
"Zhao, Yang",
"Du, Li",
"Ding, Xiao",
"Xiong, Kai",
"Sun, Zhouhao",
"Jun, Shi",
"Liu, Ting",
"Qin, Bing"
]
| Deciphering the Impact of Pretraining Data on Large Language Models through Machine Unlearning | findings-acl.559 | Poster | 2402.11537v2 |
https://aclanthology.org/2024.findings-acl.560.bib | @inproceedings{chimoto-etal-2024-critical,
title = "Critical Learning Periods: Leveraging Early Training Dynamics for Efficient Data Pruning",
author = "Chimoto, Everlyn and
Gala, Jay and
Ahia, Orevaoghene and
Kreutzer, Julia and
Bassett, Bruce and
Hooker, Sara",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.560",
pages = "9407--9426",
abstract = "Neural Machine Translation models are extremely data and compute-hungry. However, not all datapoints contribute equally to model training and generalization. Data pruning to remove the low-value data points has the benefit of drastically reducing the compute budget without significantdrop in model performance. In this paper, we propose a new data pruning technique: CheckpointsAcross Time (CAT ), that leverages early model training dynamics to identify the most relevantdata points for model performance. We benchmark CAT against several data pruning techniquesincluding COMET-QE, LASER and LaBSE. We find that CAT outperforms the benchmarks onIndo-European languages on multiple test sets. When applied to English-German, English-Frenchand English-Swahili translation tasks, CAT achieves comparable performance to using the fulldataset, while pruning up to 50{\%} of training data. We inspect the data points that CAT selectsand find that it tends to favour longer sentences and sentences with unique or rare words.",
}
| Neural Machine Translation models are extremely data and compute-hungry. However, not all datapoints contribute equally to model training and generalization. Data pruning to remove the low-value data points has the benefit of drastically reducing the compute budget without significantdrop in model performance. In this paper, we propose a new data pruning technique: CheckpointsAcross Time (CAT ), that leverages early model training dynamics to identify the most relevantdata points for model performance. We benchmark CAT against several data pruning techniquesincluding COMET-QE, LASER and LaBSE. We find that CAT outperforms the benchmarks onIndo-European languages on multiple test sets. When applied to English-German, English-Frenchand English-Swahili translation tasks, CAT achieves comparable performance to using the fulldataset, while pruning up to 50{\%} of training data. We inspect the data points that CAT selectsand find that it tends to favour longer sentences and sentences with unique or rare words. | [
"Chimoto, Everlyn",
"Gala, Jay",
"Ahia, Orevaoghene",
"Kreutzer, Julia",
"Bassett, Bruce",
"Hooker, Sara"
]
| Critical Learning Periods: Leveraging Early Training Dynamics for Efficient Data Pruning | findings-acl.560 | Poster | 2405.19462v2 |
https://aclanthology.org/2024.findings-acl.561.bib | @inproceedings{ragazzi-etal-2024-token,
title = "What Are You Token About? Differentiable Perturbed Top-$k$ Token Selection for Scientific Document Summarization",
author = "Ragazzi, Luca and
Italiani, Paolo and
Moro, Gianluca and
Panni, Mattia",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.561",
pages = "9427--9440",
abstract = "Scientific document summarization aims to condense complex and long articles in both technical and plain-language terms to facilitate the accessibility and dissemination of scientific findings. Existing datasets suffer from a deficiency in source heterogeneity, as their data predominantly stem from a single common resource, hindering effective model training and generalizability. First, we introduce SciLay, a novel dataset that includes documents from multiple natural science journals with expert-authored technical and lay summaries. Second, we propose PrunePert, a new transformer-based model that incorporates a differentiable perturbed top-$k$ encoder layer to prune irrelevant tokens in end-to-end learning. Experimental results show that our model achieves a nearly 2x speed-up compared to a state-of-the-art linear transformer, remaining comparable in effectiveness. Additional examinations underscore the importance of employing a training dataset that includes different sources to enhance the generalizability of the models. Code is available at https://github.com/disi-unibo-nlp/sci-lay.",
}
| Scientific document summarization aims to condense complex and long articles in both technical and plain-language terms to facilitate the accessibility and dissemination of scientific findings. Existing datasets suffer from a deficiency in source heterogeneity, as their data predominantly stem from a single common resource, hindering effective model training and generalizability. First, we introduce SciLay, a novel dataset that includes documents from multiple natural science journals with expert-authored technical and lay summaries. Second, we propose PrunePert, a new transformer-based model that incorporates a differentiable perturbed top-$k$ encoder layer to prune irrelevant tokens in end-to-end learning. Experimental results show that our model achieves a nearly 2x speed-up compared to a state-of-the-art linear transformer, remaining comparable in effectiveness. Additional examinations underscore the importance of employing a training dataset that includes different sources to enhance the generalizability of the models. Code is available at https://github.com/disi-unibo-nlp/sci-lay. | [
"Ragazzi, Luca",
"Italiani, Paolo",
"Moro, Gianluca",
"Panni, Mattia"
]
| What Are You Token About? Differentiable Perturbed Top-$k$ Token Selection for Scientific Document Summarization | findings-acl.561 | Poster | 2111.11591v2 |
https://aclanthology.org/2024.findings-acl.562.bib | @inproceedings{picco-etal-2024-description,
title = "Description Boosting for Zero-Shot Entity and Relation Classification",
author = "Picco, Gabriele and
Fuchs, Leopold and
Mart{\'\i}nez Galindo, Marcos and
Purpura, Alberto and
L{\'o}pez, Vanessa and
Thanh Lam, Hoang",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.562",
pages = "9441--9457",
abstract = "Zero-shot entity and relation classification models leverage available external information of unseen classes {--} e.g., textual descriptions {--} to annotate input text data. Thanks to the minimum data requirement, Zero-Shot Learning (ZSL) methods have high value in practice, especially in applications where labeled data is scarce. Even though recent research in ZSL has demonstrated significant results, our analysis reveals that those methods are sensitive to provided textual descriptions of entities (or relations). Even a minor modification of descriptions can lead to a change in the decision boundary between entity (or relation) classes. In this paper, we formally define the problem of identifying effective descriptions for zero shot inference. We propose a strategy for generating variations of an initial description, a heuristic for ranking them and an ensemble method capable of boosting the predictions of zero-shot models through description enhancement. Empirical results on four different entity and relation classification datasets show that our proposed method outperform existing approaches and achieve new SOTA results on these datasets under the ZSL settings. The source code of the proposed solutions and the evaluation framework are open-sourced.",
}
| Zero-shot entity and relation classification models leverage available external information of unseen classes {--} e.g., textual descriptions {--} to annotate input text data. Thanks to the minimum data requirement, Zero-Shot Learning (ZSL) methods have high value in practice, especially in applications where labeled data is scarce. Even though recent research in ZSL has demonstrated significant results, our analysis reveals that those methods are sensitive to provided textual descriptions of entities (or relations). Even a minor modification of descriptions can lead to a change in the decision boundary between entity (or relation) classes. In this paper, we formally define the problem of identifying effective descriptions for zero shot inference. We propose a strategy for generating variations of an initial description, a heuristic for ranking them and an ensemble method capable of boosting the predictions of zero-shot models through description enhancement. Empirical results on four different entity and relation classification datasets show that our proposed method outperform existing approaches and achieve new SOTA results on these datasets under the ZSL settings. The source code of the proposed solutions and the evaluation framework are open-sourced. | [
"Picco, Gabriele",
"Fuchs, Leopold",
"Mart{\\'\\i}nez Galindo, Marcos",
"Purpura, Alberto",
"L{\\'o}pez, Vanessa",
"Thanh Lam, Hoang"
]
| Description Boosting for Zero-Shot Entity and Relation Classification | findings-acl.562 | Poster | 2406.02245v1 |
https://aclanthology.org/2024.findings-acl.563.bib | @inproceedings{wang-etal-2024-domain-aware,
title = "Domain-Aware $k$-Nearest-Neighbor Knowledge Distillation for Machine Translation",
author = "Wang, Zhexuan and
Liu, Shudong and
Liu, Xuebo and
Zhang, Miao and
Wong, Derek and
Zhang, Min",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.563",
pages = "9458--9469",
abstract = "$k$NN-MT has utilized neighborhood knowledge for auxiliary decoding, significantly improving translation performance. Subsequently, $k$NN-KD transitions the use of neighborhood knowledge from the decoding phase to the training phase, to address the temporal and spatial inefficiencies inherent in $k$NN-MT. However, $k$NN-KD transfers all the $k$NN knowledge arbitrarily, which has the potential to restrict the learning of student models. In this paper, we propose a novel domain-aware $k$NN-KD method, which filters out domain-relevant neighborhood knowledge for learning in the distillation process. Notably, this entire process exclusively utilizes the neighborhood knowledge of the original model, eliminating the need for establishing any additional datastores. Experiments on four domain translation tasks demonstrate that our method achieves state-of-the-art performance, realizing an average gain of 1.55 COMET and 1.42 BLEU scores, by further enhancing the translation of rare words. Source code can be accessed at https://github.com/wangzx1219/Dk-KD.",
}
| $k$NN-MT has utilized neighborhood knowledge for auxiliary decoding, significantly improving translation performance. Subsequently, $k$NN-KD transitions the use of neighborhood knowledge from the decoding phase to the training phase, to address the temporal and spatial inefficiencies inherent in $k$NN-MT. However, $k$NN-KD transfers all the $k$NN knowledge arbitrarily, which has the potential to restrict the learning of student models. In this paper, we propose a novel domain-aware $k$NN-KD method, which filters out domain-relevant neighborhood knowledge for learning in the distillation process. Notably, this entire process exclusively utilizes the neighborhood knowledge of the original model, eliminating the need for establishing any additional datastores. Experiments on four domain translation tasks demonstrate that our method achieves state-of-the-art performance, realizing an average gain of 1.55 COMET and 1.42 BLEU scores, by further enhancing the translation of rare words. Source code can be accessed at https://github.com/wangzx1219/Dk-KD. | [
"Wang, Zhexuan",
"Liu, Shudong",
"Liu, Xuebo",
"Zhang, Miao",
"Wong, Derek",
"Zhang, Min"
]
| Domain-Aware $k$-Nearest-Neighbor Knowledge Distillation for Machine Translation | findings-acl.563 | Poster | 2003.02877v3 |
https://aclanthology.org/2024.findings-acl.564.bib | @inproceedings{liu-etal-2024-beyond-single,
title = "Beyond Single-Event Extraction: Towards Efficient Document-Level Multi-Event Argument Extraction",
author = "Liu, Wanlong and
Zhou, Li and
Zeng, DingYi and
Xiao, Yichen and
Cheng, Shaohuan and
Zhang, Chen and
Lee, Grandee and
Zhang, Malu and
Chen, Wenyu",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.564",
pages = "9470--9487",
abstract = "Recent mainstream event argument extraction methods process each event in isolation, resulting in inefficient inference and ignoring the correlations among multiple events. To address these limitations, here we propose a multiple-event argument extraction model DEEIA (Dependency-guided Encoding and Event-specific Information Aggregation), capable of extracting arguments from all events within a document simultaneously. The proposed DEEIA model employs a multi-event prompt mechanism, comprising DE and EIA modules. The DE module is designed to improve the correlation between prompts and their corresponding event contexts, whereas the EIA module provides event-specific information to improve contextual understanding. Extensive experiments show that our method achieves new state-of-the-art performance on four public datasets (RAMS, WikiEvents, MLEE, and ACE05), while significantly saving the inference time compared to the baselines. Further analyses demonstrate the effectiveness of the proposed modules.",
}
| Recent mainstream event argument extraction methods process each event in isolation, resulting in inefficient inference and ignoring the correlations among multiple events. To address these limitations, here we propose a multiple-event argument extraction model DEEIA (Dependency-guided Encoding and Event-specific Information Aggregation), capable of extracting arguments from all events within a document simultaneously. The proposed DEEIA model employs a multi-event prompt mechanism, comprising DE and EIA modules. The DE module is designed to improve the correlation between prompts and their corresponding event contexts, whereas the EIA module provides event-specific information to improve contextual understanding. Extensive experiments show that our method achieves new state-of-the-art performance on four public datasets (RAMS, WikiEvents, MLEE, and ACE05), while significantly saving the inference time compared to the baselines. Further analyses demonstrate the effectiveness of the proposed modules. | [
"Liu, Wanlong",
"Zhou, Li",
"Zeng, DingYi",
"Xiao, Yichen",
"Cheng, Shaohuan",
"Zhang, Chen",
"Lee, Gr",
"ee",
"Zhang, Malu",
"Chen, Wenyu"
]
| Beyond Single-Event Extraction: Towards Efficient Document-Level Multi-Event Argument Extraction | findings-acl.564 | Poster | 2305.08495v1 |
https://aclanthology.org/2024.findings-acl.565.bib | @inproceedings{xu-etal-2024-revisiting,
title = "Revisiting Interpolation Augmentation for Speech-to-Text Generation",
author = "Xu, Chen and
Wang, Jie and
Liu, Xiaoqian and
Dong, Qian and
Zhang, Chunliang and
Xiao, Tong and
Zhu, JingBo and
Man, Dapeng and
Yang, Wu",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.565",
pages = "9488--9499",
abstract = "Speech-to-text (S2T) generation systems frequently face challenges in low-resource scenarios, primarily due to the lack of extensive labeled datasets. One emerging solution is constructing virtual training samples by interpolating inputs and labels, which has notably enhanced system generalization in other domains. Despite its potential, this technique{'}s application in S2T tasks has remained under-explored. In this paper, we delve into the utility of interpolation augmentation, guided by several pivotal questions. Our findings reveal that employing an appropriate strategy in interpolation augmentation significantly enhances performance across diverse tasks, architectures, and data scales, offering a promising avenue for more robust S2T systems in resource-constrained settings.",
}
| Speech-to-text (S2T) generation systems frequently face challenges in low-resource scenarios, primarily due to the lack of extensive labeled datasets. One emerging solution is constructing virtual training samples by interpolating inputs and labels, which has notably enhanced system generalization in other domains. Despite its potential, this technique{'}s application in S2T tasks has remained under-explored. In this paper, we delve into the utility of interpolation augmentation, guided by several pivotal questions. Our findings reveal that employing an appropriate strategy in interpolation augmentation significantly enhances performance across diverse tasks, architectures, and data scales, offering a promising avenue for more robust S2T systems in resource-constrained settings. | [
"Xu, Chen",
"Wang, Jie",
"Liu, Xiaoqian",
"Dong, Qian",
"Zhang, Chunliang",
"Xiao, Tong",
"Zhu, JingBo",
"Man, Dapeng",
"Yang, Wu"
]
| Revisiting Interpolation Augmentation for Speech-to-Text Generation | findings-acl.565 | Poster | 2406.15846v1 |
https://aclanthology.org/2024.findings-acl.566.bib | @inproceedings{ulmer-etal-2024-bootstrapping,
title = "Bootstrapping {LLM}-based Task-Oriented Dialogue Agents via Self-Talk",
author = "Ulmer, Dennis and
Mansimov, Elman and
Lin, Kaixiang and
Sun, Lijia and
Gao, Xibin and
Zhang, Yi",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.566",
pages = "9500--9522",
abstract = "Large language models (LLMs) are powerful dialogue agents, but specializing them towards fulfilling a specific function can be challenging. Instructing tuning, i.e. tuning models on instruction and sample responses generated by humans (Ouyang et al., 2022), has proven as an effective method to do so, yet requires a number of data samples that a) might not be available or b) costly to generate. Furthermore, this cost increases when the goal is to make the LLM follow a specific workflow within a dialogue instead of single instructions. Inspired by the self-play technique in reinforcement learning and the use of LLMs to simulate human agents, we propose a more effective method for data collection through LLMs engaging in a conversation in various roles. This approach generates a training data via {``}self-talk{''} of LLMs that can be refined and utilized for supervised fine-tuning. We introduce an automated way to measure the (partial) success of a dialogue. This metric is used to filter the generated conversational data that is fed back in LLM for training. Based on our automated and human evaluations of conversation quality, we demonstrate that such self-talk data improves results. In addition, we examine the various characteristics that showcase the quality of generated dialogues and how they can be connected to their potential utility as training data.",
}
| Large language models (LLMs) are powerful dialogue agents, but specializing them towards fulfilling a specific function can be challenging. Instructing tuning, i.e. tuning models on instruction and sample responses generated by humans (Ouyang et al., 2022), has proven as an effective method to do so, yet requires a number of data samples that a) might not be available or b) costly to generate. Furthermore, this cost increases when the goal is to make the LLM follow a specific workflow within a dialogue instead of single instructions. Inspired by the self-play technique in reinforcement learning and the use of LLMs to simulate human agents, we propose a more effective method for data collection through LLMs engaging in a conversation in various roles. This approach generates a training data via {``}self-talk{''} of LLMs that can be refined and utilized for supervised fine-tuning. We introduce an automated way to measure the (partial) success of a dialogue. This metric is used to filter the generated conversational data that is fed back in LLM for training. Based on our automated and human evaluations of conversation quality, we demonstrate that such self-talk data improves results. In addition, we examine the various characteristics that showcase the quality of generated dialogues and how they can be connected to their potential utility as training data. | [
"Ulmer, Dennis",
"Mansimov, Elman",
"Lin, Kaixiang",
"Sun, Lijia",
"Gao, Xibin",
"Zhang, Yi"
]
| Bootstrapping {LLM}-based Task-Oriented Dialogue Agents via Self-Talk | findings-acl.566 | Poster | 2107.11904v1 |
https://aclanthology.org/2024.findings-acl.567.bib | @inproceedings{wang-li-2024-semantic,
title = "Semantic are Beacons: A Semantic Perspective for Unveiling Parameter-Efficient Fine-Tuning in Knowledge Learning",
author = "Wang, Renzhi and
Li, Piji",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.567",
pages = "9523--9537",
abstract = "Parameter-Efficient Fine-Tuning (PEFT) methods enable efficient adaptation of Large Language Models (LLMs) to various downstream applications. However, the effectiveness of the PEFT diminishes notably when downstream tasks require accurate learning of specific knowledge. In this paper, we adopt a semantic perspective to investigate this phenomenon, uncovering the reasons behind PEFT{'}s limitations in knowledge learning task. Our findings reveals that: (1) PEFT presents a notable risk of pushing the model away from the intended knowledge target; (2) multiple knowledge interfere with each other, and such interference suppresses the learning and expression of knowledge features. Based on these insights, we introduce a data filtering strategy to exclude data that is detrimental to knowledge learning and a re-weighted learning strategy to make the model attentive to semantic distance during knowledge learning. Experimental results demonstrate the effectiveness of the proposed method on open-source large language model, further validate the semantic challenge in PEFT, thus paving the way for future research.",
}
| Parameter-Efficient Fine-Tuning (PEFT) methods enable efficient adaptation of Large Language Models (LLMs) to various downstream applications. However, the effectiveness of the PEFT diminishes notably when downstream tasks require accurate learning of specific knowledge. In this paper, we adopt a semantic perspective to investigate this phenomenon, uncovering the reasons behind PEFT{'}s limitations in knowledge learning task. Our findings reveals that: (1) PEFT presents a notable risk of pushing the model away from the intended knowledge target; (2) multiple knowledge interfere with each other, and such interference suppresses the learning and expression of knowledge features. Based on these insights, we introduce a data filtering strategy to exclude data that is detrimental to knowledge learning and a re-weighted learning strategy to make the model attentive to semantic distance during knowledge learning. Experimental results demonstrate the effectiveness of the proposed method on open-source large language model, further validate the semantic challenge in PEFT, thus paving the way for future research. | [
"Wang, Renzhi",
"Li, Piji"
]
| Semantic are Beacons: A Semantic Perspective for Unveiling Parameter-Efficient Fine-Tuning in Knowledge Learning | findings-acl.567 | Poster | 2405.18292v1 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.