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.368.bib | @inproceedings{cui-etal-2024-unveiling,
title = "Unveiling the Art of Heading Design: A Harmonious Blend of Summarization, Neology, and Algorithm",
author = "Cui, Shaobo and
Feng, Yiyang and
Mao, Yisong and
Hou, Yifan and
Faltings, Boi",
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.368",
pages = "6149--6174",
abstract = "Crafting an appealing heading is crucial for attracting readers and marketing work or products. A popular way is to summarize the main idea with a refined description and a memorable acronym. However, there lacks a systematic study and a formal benchmark including datasets and metrics. Motivated by this absence, we introduce LOgogram, a novel benchmark comprising 6,653 paper abstracts with corresponding descriptions and acronyms. To measure the quality of heading generation, we propose a set of evaluation metrics from three aspects: summarization, neology, and algorithm. Additionally, we explore three strategies for heading generation(generation ordering, tokenization of acronyms, and framework design) under various prevalent learning paradigms(supervised fine-tuning, in-context learning with Large Language Models(LLMs), and reinforcement learning) on our benchmark. Our experimental results indicate the difficulty in identifying a practice that excels across all summarization, neologistic, and algorithmic aspects.",
}
| Crafting an appealing heading is crucial for attracting readers and marketing work or products. A popular way is to summarize the main idea with a refined description and a memorable acronym. However, there lacks a systematic study and a formal benchmark including datasets and metrics. Motivated by this absence, we introduce LOgogram, a novel benchmark comprising 6,653 paper abstracts with corresponding descriptions and acronyms. To measure the quality of heading generation, we propose a set of evaluation metrics from three aspects: summarization, neology, and algorithm. Additionally, we explore three strategies for heading generation(generation ordering, tokenization of acronyms, and framework design) under various prevalent learning paradigms(supervised fine-tuning, in-context learning with Large Language Models(LLMs), and reinforcement learning) on our benchmark. Our experimental results indicate the difficulty in identifying a practice that excels across all summarization, neologistic, and algorithmic aspects. | [
"Cui, Shaobo",
"Feng, Yiyang",
"Mao, Yisong",
"Hou, Yifan",
"Faltings, Boi"
]
| Unveiling the Art of Heading Design: A Harmonious Blend of Summarization, Neology, and Algorithm | findings-acl.368 | Poster | 2006.03743v1 |
https://aclanthology.org/2024.findings-acl.369.bib | @inproceedings{wuehrl-etal-2024-understanding,
title = "Understanding Fine-grained Distortions in Reports of Scientific Findings",
author = "Wuehrl, Amelie and
Wright, Dustin and
Klinger, Roman and
Augenstein, Isabelle",
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.369",
pages = "6175--6191",
abstract = "Distorted science communication harms individuals and society as it can lead to unhealthy behavior change and decrease trust in scientific institutions. Given the rapidly increasing volume of science communication in recent years, a fine-grained understanding of how findings from scientific publications are reported to the general public, and methods to detect distortions from the original work automatically, are crucial. Prior work focused on individual aspects of distortions or worked with unpaired data. In this work, we make three foundational contributions towards addressing this problem: (1) annotating 1,600 instances of scientific findings from academic papers paired with corresponding findings as reported in news articles and tweets wrt. four characteristics: causality, certainty, generality and sensationalism; (2) establishing baselines for automatically detecting these characteristics; and (3) analyzing the prevalence of changes in these characteristics in both human-annotated and large-scale unlabeled data. Our results show that scientific findings frequently undergo subtle distortions when reported. Tweets distort findings more often than science news reports. Detecting fine-grained distortions automatically poses a challenging task. In our experiments, fine-tuned task-specific models consistently outperform few-shot LLM prompting.",
}
| Distorted science communication harms individuals and society as it can lead to unhealthy behavior change and decrease trust in scientific institutions. Given the rapidly increasing volume of science communication in recent years, a fine-grained understanding of how findings from scientific publications are reported to the general public, and methods to detect distortions from the original work automatically, are crucial. Prior work focused on individual aspects of distortions or worked with unpaired data. In this work, we make three foundational contributions towards addressing this problem: (1) annotating 1,600 instances of scientific findings from academic papers paired with corresponding findings as reported in news articles and tweets wrt. four characteristics: causality, certainty, generality and sensationalism; (2) establishing baselines for automatically detecting these characteristics; and (3) analyzing the prevalence of changes in these characteristics in both human-annotated and large-scale unlabeled data. Our results show that scientific findings frequently undergo subtle distortions when reported. Tweets distort findings more often than science news reports. Detecting fine-grained distortions automatically poses a challenging task. In our experiments, fine-tuned task-specific models consistently outperform few-shot LLM prompting. | [
"Wuehrl, Amelie",
"Wright, Dustin",
"Klinger, Roman",
"Augenstein, Isabelle"
]
| Understanding Fine-grained Distortions in Reports of Scientific Findings | findings-acl.369 | Poster | 2402.12431v1 |
https://aclanthology.org/2024.findings-acl.370.bib | @inproceedings{jin-etal-2024-mm,
title = "{MM}-{SOC}: Benchmarking Multimodal Large Language Models in Social Media Platforms",
author = "Jin, Yiqiao and
Choi, Minje and
Verma, Gaurav and
Wang, Jindong and
Kumar, Srijan",
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.370",
pages = "6192--6210",
abstract = "Social media platforms are hubs for multimodal information exchange, encompassing text, images, and videos, making it challenging for machines to comprehend the information or emotions associated with interactions in online spaces. Multimodal Large Language Models (MLLMs) have emerged as a promising solution to address these challenges, yet struggle with accurately interpreting human emotions and complex contents like misinformation. This paper introduces MM-Soc, a comprehensive benchmark designed to evaluate MLLMs{'} understanding of multimodal social media content. MM-Soc compiles prominent multimodal datasets and incorporates a novel large-scale YouTube tagging dataset, targeting a range of tasks from misinformation detection, hate speech detection, and social context generation. Through our exhaustive evaluation on ten size-variants of four open-source MLLMs, we have identified significant performance disparities, highlighting the need for advancements in models{'} social understanding capabilities. Our analysis reveals that, in a zero-shot setting, various types of MLLMs generally exhibit difficulties in handling social media tasks. However, MLLMs demonstrate performance improvements post fine-tuning, suggesting potential pathways for improvement.",
}
| Social media platforms are hubs for multimodal information exchange, encompassing text, images, and videos, making it challenging for machines to comprehend the information or emotions associated with interactions in online spaces. Multimodal Large Language Models (MLLMs) have emerged as a promising solution to address these challenges, yet struggle with accurately interpreting human emotions and complex contents like misinformation. This paper introduces MM-Soc, a comprehensive benchmark designed to evaluate MLLMs{'} understanding of multimodal social media content. MM-Soc compiles prominent multimodal datasets and incorporates a novel large-scale YouTube tagging dataset, targeting a range of tasks from misinformation detection, hate speech detection, and social context generation. Through our exhaustive evaluation on ten size-variants of four open-source MLLMs, we have identified significant performance disparities, highlighting the need for advancements in models{'} social understanding capabilities. Our analysis reveals that, in a zero-shot setting, various types of MLLMs generally exhibit difficulties in handling social media tasks. However, MLLMs demonstrate performance improvements post fine-tuning, suggesting potential pathways for improvement. | [
"Jin, Yiqiao",
"Choi, Minje",
"Verma, Gaurav",
"Wang, Jindong",
"Kumar, Srijan"
]
| {MM}-{SOC}: Benchmarking Multimodal Large Language Models in Social Media Platforms | findings-acl.370 | Poster | 2402.14154v2 |
https://aclanthology.org/2024.findings-acl.371.bib | @inproceedings{srivastava-etal-2024-instances,
title = "Instances Need More Care: Rewriting Prompts for Instances with {LLM}s in the Loop Yields Better Zero-Shot Performance",
author = "Srivastava, Saurabh and
Huang, Chengyue and
Fan, Weiguo and
Yao, Ziyu",
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.371",
pages = "6211--6232",
abstract = "Large language models (LLMs) have revolutionized zero-shot task performance, mitigating the need for task-specific annotations while enhancing task generalizability. Despite its advancements, current methods using trigger phrases such as {``}Let{'}s think step by step{''} remain limited. This study introduces PRomPTed, an approach that optimizes the zero-shot prompts for individual task instances following an innovative manner of {``}LLMs in the loop{''}.Our comprehensive evaluation across 13 datasets and 10 task types based on GPT-4 reveals that PRomPTed significantly outperforms both the naive zero-shot approaches and a strong baseline (i.e., {``}Output Refinement{''}) which refines the task output instead of the input prompt. Our experimental results also confirmed the generalization of this advantage to the relatively weaker GPT-3.5. Even more intriguingly, we found that leveraging GPT-3.5 to rewrite prompts for the stronger GPT-4 not only matches but occasionally exceeds the efficacy of using GPT-4 as the prompt rewriter. Our research thus presents a huge value in not only enhancing zero-shot LLM performance but also potentially enabling supervising LLMs with their weaker counterparts, a capability attracting much interest recently. Finally, our additional experiments confirm the generalization of the advantages to open-source LLMs such as Mistral 7B and Mixtral 8x7B.",
}
| Large language models (LLMs) have revolutionized zero-shot task performance, mitigating the need for task-specific annotations while enhancing task generalizability. Despite its advancements, current methods using trigger phrases such as {``}Let{'}s think step by step{''} remain limited. This study introduces PRomPTed, an approach that optimizes the zero-shot prompts for individual task instances following an innovative manner of {``}LLMs in the loop{''}.Our comprehensive evaluation across 13 datasets and 10 task types based on GPT-4 reveals that PRomPTed significantly outperforms both the naive zero-shot approaches and a strong baseline (i.e., {``}Output Refinement{''}) which refines the task output instead of the input prompt. Our experimental results also confirmed the generalization of this advantage to the relatively weaker GPT-3.5. Even more intriguingly, we found that leveraging GPT-3.5 to rewrite prompts for the stronger GPT-4 not only matches but occasionally exceeds the efficacy of using GPT-4 as the prompt rewriter. Our research thus presents a huge value in not only enhancing zero-shot LLM performance but also potentially enabling supervising LLMs with their weaker counterparts, a capability attracting much interest recently. Finally, our additional experiments confirm the generalization of the advantages to open-source LLMs such as Mistral 7B and Mixtral 8x7B. | [
"Srivastava, Saurabh",
"Huang, Chengyue",
"Fan, Weiguo",
"Yao, Ziyu"
]
| Instances Need More Care: Rewriting Prompts for Instances with {LLM}s in the Loop Yields Better Zero-Shot Performance | findings-acl.371 | Poster | 2310.02107v4 |
https://aclanthology.org/2024.findings-acl.372.bib | @inproceedings{xiong-etal-2024-benchmarking,
title = "Benchmarking Retrieval-Augmented Generation for Medicine",
author = "Xiong, Guangzhi and
Jin, Qiao and
Lu, Zhiyong and
Zhang, Aidong",
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.372",
pages = "6233--6251",
abstract = "While large language models (LLMs) have achieved state-of-the-art performance on a wide range of medical question answering (QA) tasks, they still face challenges with hallucinations and outdated knowledge. Retrieval-augmented generation (RAG) is a promising solution and has been widely adopted. However, a RAG system can involve multiple flexible components, and there is a lack of best practices regarding the optimal RAG setting for various medical purposes. To systematically evaluate such systems, we propose the Medical Information Retrieval-Augmented Generation Evaluation (MIRAGE), a first-of-its-kind benchmark including 7,663 questions from five medical QA datasets. Using MIRAGE, we conducted large-scale experiments with over 1.8 trillion prompt tokens on 41 combinations of different corpora, retrievers, and backbone LLMs through the MedRAG toolkit introduced in this work. Overall, MedRAG improves the accuracy of six different LLMs by up to 18{\%} over chain-of-thought prompting, elevating the performance of GPT-3.5 and Mixtral to GPT-4-level. Our results show that the combination of various medical corpora and retrievers achieves the best performance. In addition, we discovered a log-linear scaling property and the {``}lost-in-the-middle{''} effects in medical RAG. We believe our comprehensive evaluations can serve as practical guidelines for implementing RAG systems for medicine.",
}
| While large language models (LLMs) have achieved state-of-the-art performance on a wide range of medical question answering (QA) tasks, they still face challenges with hallucinations and outdated knowledge. Retrieval-augmented generation (RAG) is a promising solution and has been widely adopted. However, a RAG system can involve multiple flexible components, and there is a lack of best practices regarding the optimal RAG setting for various medical purposes. To systematically evaluate such systems, we propose the Medical Information Retrieval-Augmented Generation Evaluation (MIRAGE), a first-of-its-kind benchmark including 7,663 questions from five medical QA datasets. Using MIRAGE, we conducted large-scale experiments with over 1.8 trillion prompt tokens on 41 combinations of different corpora, retrievers, and backbone LLMs through the MedRAG toolkit introduced in this work. Overall, MedRAG improves the accuracy of six different LLMs by up to 18{\%} over chain-of-thought prompting, elevating the performance of GPT-3.5 and Mixtral to GPT-4-level. Our results show that the combination of various medical corpora and retrievers achieves the best performance. In addition, we discovered a log-linear scaling property and the {``}lost-in-the-middle{''} effects in medical RAG. We believe our comprehensive evaluations can serve as practical guidelines for implementing RAG systems for medicine. | [
"Xiong, Guangzhi",
"Jin, Qiao",
"Lu, Zhiyong",
"Zhang, Aidong"
]
| Benchmarking Retrieval-Augmented Generation for Medicine | findings-acl.372 | Poster | 2311.09774v1 |
https://aclanthology.org/2024.findings-acl.373.bib | @inproceedings{yuan-etal-2024-chatmusician,
title = "{C}hat{M}usician: Understanding and Generating Music Intrinsically with {LLM}",
author = "Yuan, Ruibin and
Lin, Hanfeng and
Wang, Yi and
Tian, Zeyue and
Wu, Shangda and
Shen, Tianhao and
Zhang, Ge and
Wu, Yuhang and
Liu, Cong and
Zhou, Ziya and
Xue, Liumeng and
Ma, Ziyang and
Liu, Qin and
Zheng, Tianyu and
Li, Yizhi and
Ma, Yinghao and
Liang, Yiming and
Chi, Xiaowei and
Liu, Ruibo and
Wang, Zili and
Lin, Chenghua and
Liu, Qifeng and
Jiang, Tao and
Huang, Wenhao and
Chen, Wenhu and
Fu, Jie and
Benetos, Emmanouil and
Xia, Gus and
Dannenberg, Roger and
Xue, Wei and
Kang, Shiyin and
Guo, Yike",
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.373",
pages = "6252--6271",
abstract = "While LLMs demonstrate impressive capabilities in musical knowledge, we find that music reasoning is still an unsolved task.We introduce ChatMusician, an open-source large language model (LLM) that integrates intrinsic musical abilities. It is based on continual pre-training and finetuning LLaMA2 on a text-compatible music representation, ABC notation, and the music is treated as a second language.ChatMusician can understand and generate music with a pure text tokenizer without external multi-modal neural structures or tokenizers. Interestingly, endowing musical abilities does not harm language abilities, even achieving a slightly higher MMLU score.ChatMusician is capable of composing well-structured, full-length music, condition on texts, chords, melodies, motifs, musical forms, etc.On our meticulously curated college-level music understanding benchmark, MusicTheoryBench, ChatMusician surpasses LLaMA2 and GPT-3.5 by a noticeable margin. We show that ChatMusician preserves or even surpasses the original LLaMA2 7B{'}s language abilities by evaluating on MMLU benchmark.Our work reveals that LLMs can be an excellent compressor for music, which can be seen as humanity{'}s creative language, but there remains significant territory to be conquered.We release our 5B token music-language corpora MusicPiles, the collected MusicTheoryBench, code, model and demo.",
}
| While LLMs demonstrate impressive capabilities in musical knowledge, we find that music reasoning is still an unsolved task.We introduce ChatMusician, an open-source large language model (LLM) that integrates intrinsic musical abilities. It is based on continual pre-training and finetuning LLaMA2 on a text-compatible music representation, ABC notation, and the music is treated as a second language.ChatMusician can understand and generate music with a pure text tokenizer without external multi-modal neural structures or tokenizers. Interestingly, endowing musical abilities does not harm language abilities, even achieving a slightly higher MMLU score.ChatMusician is capable of composing well-structured, full-length music, condition on texts, chords, melodies, motifs, musical forms, etc.On our meticulously curated college-level music understanding benchmark, MusicTheoryBench, ChatMusician surpasses LLaMA2 and GPT-3.5 by a noticeable margin. We show that ChatMusician preserves or even surpasses the original LLaMA2 7B{'}s language abilities by evaluating on MMLU benchmark.Our work reveals that LLMs can be an excellent compressor for music, which can be seen as humanity{'}s creative language, but there remains significant territory to be conquered.We release our 5B token music-language corpora MusicPiles, the collected MusicTheoryBench, code, model and demo. | [
"Yuan, Ruibin",
"Lin, Hanfeng",
"Wang, Yi",
"Tian, Zeyue",
"Wu, Shangda",
"Shen, Tianhao",
"Zhang, Ge",
"Wu, Yuhang",
"Liu, Cong",
"Zhou, Ziya",
"Xue, Liumeng",
"Ma, Ziyang",
"Liu, Qin",
"Zheng, Tianyu",
"Li, Yizhi",
"Ma, Yinghao",
"Liang, Yiming",
"Chi, Xiaowei",
"Liu, Ruibo",
"Wang, Zili",
"Lin, Chenghua",
"Liu, Qifeng",
"Jiang, Tao",
"Huang, Wenhao",
"Chen, Wenhu",
"Fu, Jie",
"Benetos, Emmanouil",
"Xia, Gus",
"Dannenberg, Roger",
"Xue, Wei",
"Kang, Shiyin",
"Guo, Yike"
]
| {C}hat{M}usician: Understanding and Generating Music Intrinsically with {LLM} | findings-acl.373 | Poster | 2407.21531v1 |
https://aclanthology.org/2024.findings-acl.374.bib | @inproceedings{tan-etal-2024-towards,
title = "Towards Robust Temporal Reasoning of Large Language Models via a Multi-Hop {QA} Dataset and Pseudo-Instruction Tuning",
author = "Tan, Qingyu and
Ng, Hwee Tou and
Bing, Lidong",
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.374",
pages = "6272--6286",
abstract = "Knowledge in the real world is being updated constantly. However, it is costly to frequently update large language models (LLMs). Therefore, it is crucial for LLMs to understand the concept of temporal knowledge. However, prior works on temporal question answering (TQA) did not emphasize multi-answer and multi-hop types of temporal reasoning. In this paper, we propose a complex temporal question-answering dataset Complex-TR that focuses on multi-answer and multi-hop temporal reasoning. Besides, we also propose a novel data augmentation strategy to improve the complex temporal reasoning capability and robustness of LLMs. We conducted experiments on multiple temporal QA datasets. Experimental results show that our method is able to improve LLMs{'} performance on temporal QA benchmarks by significant margins.",
}
| Knowledge in the real world is being updated constantly. However, it is costly to frequently update large language models (LLMs). Therefore, it is crucial for LLMs to understand the concept of temporal knowledge. However, prior works on temporal question answering (TQA) did not emphasize multi-answer and multi-hop types of temporal reasoning. In this paper, we propose a complex temporal question-answering dataset Complex-TR that focuses on multi-answer and multi-hop temporal reasoning. Besides, we also propose a novel data augmentation strategy to improve the complex temporal reasoning capability and robustness of LLMs. We conducted experiments on multiple temporal QA datasets. Experimental results show that our method is able to improve LLMs{'} performance on temporal QA benchmarks by significant margins. | [
"Tan, Qingyu",
"Ng, Hwee Tou",
"Bing, Lidong"
]
| Towards Robust Temporal Reasoning of Large Language Models via a Multi-Hop {QA} Dataset and Pseudo-Instruction Tuning | findings-acl.374 | Poster | 2311.09821v2 |
https://aclanthology.org/2024.findings-acl.375.bib | @inproceedings{voronov-etal-2024-mind,
title = "Mind Your Format: Towards Consistent Evaluation of In-Context Learning Improvements",
author = "Voronov, Anton and
Wolf, Lena and
Ryabinin, Max",
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.375",
pages = "6287--6310",
abstract = "Large language models demonstrate a remarkable capability for learning to solve new tasks from a few examples.The $\textit{prompt template}$, or the way the input examples are formatted to obtain the prompt, is an important yet often overlooked aspect of in-context learning.In this work, we conduct a comprehensive study of the template format{'}s influence on the in-context learning performance.We evaluate the impact of the prompt template across 21 models (from 770M to 70B parameters) and 4 standard classification datasets. We show that a poor choice of the template can reduce the performance of the strongest models and inference methods to a random guess level.More importantly, the best templates do not transfer between different setups and even between models of the same family.Our findings show that the currently prevalent approach to evaluation, which ignores template selection, may give misleading results due to different templates in different works.As a first step towards mitigating this issue, we propose $\textit{Template Ensembles}$ that aggregate model predictions across several templates.This simple test-time augmentation boosts average performance while being robust to the choice of random set of templates.",
}
| Large language models demonstrate a remarkable capability for learning to solve new tasks from a few examples.The $\textit{prompt template}$, or the way the input examples are formatted to obtain the prompt, is an important yet often overlooked aspect of in-context learning.In this work, we conduct a comprehensive study of the template format{'}s influence on the in-context learning performance.We evaluate the impact of the prompt template across 21 models (from 770M to 70B parameters) and 4 standard classification datasets. We show that a poor choice of the template can reduce the performance of the strongest models and inference methods to a random guess level.More importantly, the best templates do not transfer between different setups and even between models of the same family.Our findings show that the currently prevalent approach to evaluation, which ignores template selection, may give misleading results due to different templates in different works.As a first step towards mitigating this issue, we propose $\textit{Template Ensembles}$ that aggregate model predictions across several templates.This simple test-time augmentation boosts average performance while being robust to the choice of random set of templates. | [
"Voronov, Anton",
"Wolf, Lena",
"Ryabinin, Max"
]
| Mind Your Format: Towards Consistent Evaluation of In-Context Learning Improvements | findings-acl.375 | Poster | 2401.06766v3 |
https://aclanthology.org/2024.findings-acl.376.bib | @inproceedings{liu-etal-2024-knowledge-graph,
title = "Knowledge Graph-Enhanced Large Language Models via Path Selection",
author = "Liu, Haochen and
Wang, Song and
Zhu, Yaochen and
Dong, Yushun and
Li, Jundong",
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.376",
pages = "6311--6321",
abstract = "Large Language Models (LLMs) have shown unprecedented performance in various real-world applications. However, they are known to generate factually inaccurate outputs, a.k.a. the hallucination problem. In recent years, incorporating external knowledge extracted from Knowledge Graphs (KGs) has become a promising strategy to improve the factual accuracy of LLM-generated outputs. Nevertheless, most existing explorations rely on LLMs themselves to perform KG knowledge extraction, which is highly inflexible as LLMs can only provide binary judgment on whether a certain knowledge (e.g., a knowledge path in KG) should be used. In addition, LLMs tend to pick only knowledge with direct semantic relationship with the input text, while potentially useful knowledge with indirect semantics can be ignored. In this work, we propose a principled framework KELP with three stages to handle the above problems. Specifically, KELP is able to achieve finer granularity of flexible knowledge extraction by generating scores for knowledge paths with input texts via latent semantic matching. Meanwhile, knowledge paths with indirect semantic relationships with the input text can also be considered via trained encoding between the selected paths in KG and the input text. Experiments on real-world datasets validate the effectiveness of KELP.",
}
| Large Language Models (LLMs) have shown unprecedented performance in various real-world applications. However, they are known to generate factually inaccurate outputs, a.k.a. the hallucination problem. In recent years, incorporating external knowledge extracted from Knowledge Graphs (KGs) has become a promising strategy to improve the factual accuracy of LLM-generated outputs. Nevertheless, most existing explorations rely on LLMs themselves to perform KG knowledge extraction, which is highly inflexible as LLMs can only provide binary judgment on whether a certain knowledge (e.g., a knowledge path in KG) should be used. In addition, LLMs tend to pick only knowledge with direct semantic relationship with the input text, while potentially useful knowledge with indirect semantics can be ignored. In this work, we propose a principled framework KELP with three stages to handle the above problems. Specifically, KELP is able to achieve finer granularity of flexible knowledge extraction by generating scores for knowledge paths with input texts via latent semantic matching. Meanwhile, knowledge paths with indirect semantic relationships with the input text can also be considered via trained encoding between the selected paths in KG and the input text. Experiments on real-world datasets validate the effectiveness of KELP. | [
"Liu, Haochen",
"Wang, Song",
"Zhu, Yaochen",
"Dong, Yushun",
"Li, Jundong"
]
| Knowledge Graph-Enhanced Large Language Models via Path Selection | findings-acl.376 | Poster | 2406.13862v1 |
https://aclanthology.org/2024.findings-acl.377.bib | @inproceedings{huang-etal-2024-ottawa,
title = "{OTTAWA}: Optimal {T}ranspor{T} Adaptive Word Aligner for Hallucination and Omission Translation Errors Detection",
author = "Huang, Chenyang and
Ghaddar, Abbas and
Kobyzev, Ivan and
Rezagholizadeh, Mehdi and
Zaiane, Osmar and
Chen, Boxing",
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.377",
pages = "6322--6334",
abstract = "Recently, there has been considerable attention on detecting hallucinations and omissions in Machine Translation (MT) systems. The two dominant approaches to tackle this task involve analyzing the MT system{'}s internal states or relying on the output of external tools, such as sentence similarity or MT quality estimators. In this work, we introduce OTTAWA, a novel Optimal Transport (OT)-based word aligner specifically designed to enhance the detection of hallucinations and omissions in MT systems. Our approach explicitly models the missing alignments by introducing a {``}null{''} vector, for which we propose a novel one-side constrained OT setting to allow an adaptive null alignment. Our approach yields competitive results compared to state-of-the-art methods across 18 language pairs on the HalOmi benchmark. In addition, it shows promising features, such as the ability to distinguish between both error types and perform word-level detection without accessing the MT system{'}s internal states.",
}
| Recently, there has been considerable attention on detecting hallucinations and omissions in Machine Translation (MT) systems. The two dominant approaches to tackle this task involve analyzing the MT system{'}s internal states or relying on the output of external tools, such as sentence similarity or MT quality estimators. In this work, we introduce OTTAWA, a novel Optimal Transport (OT)-based word aligner specifically designed to enhance the detection of hallucinations and omissions in MT systems. Our approach explicitly models the missing alignments by introducing a {``}null{''} vector, for which we propose a novel one-side constrained OT setting to allow an adaptive null alignment. Our approach yields competitive results compared to state-of-the-art methods across 18 language pairs on the HalOmi benchmark. In addition, it shows promising features, such as the ability to distinguish between both error types and perform word-level detection without accessing the MT system{'}s internal states. | [
"Huang, Chenyang",
"Ghaddar, Abbas",
"Kobyzev, Ivan",
"Rezagholizadeh, Mehdi",
"Zaiane, Osmar",
"Chen, Boxing"
]
| {OTTAWA}: Optimal {T}ranspor{T} Adaptive Word Aligner for Hallucination and Omission Translation Errors Detection | findings-acl.377 | Poster | 2406.01919v1 |
https://aclanthology.org/2024.findings-acl.378.bib | @inproceedings{yu-etal-2024-onsep,
title = "{ONSEP}: A Novel Online Neural-Symbolic Framework for Event Prediction Based on Large Language Model",
author = "Yu, Xuanqing and
Sun, Wangtao and
Li, Jingwei and
Liu, Kang and
Liu, Chengbao and
Tan, 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.378",
pages = "6335--6350",
abstract = "In the realm of event prediction, temporal knowledge graph forecasting (TKGF) stands as a pivotal technique. Previous approaches face the challenges of not utilizing experience during testing and relying on a single short-term history, which limits adaptation to evolving data. In this paper, we introduce the Online Neural-Symbolic Event Prediction (ONSEP) framework, which innovates by integrating dynamic causal rule mining (DCRM) and dual history augmented generation (DHAG). DCRM dynamically constructs causal rules from real-time data, allowing for swift adaptation to new causal relationships. In parallel, DHAG merges short-term and long-term historical contexts, leveraging a bi-branch approach to enrich event prediction. Our framework demonstrates notable performance enhancements across diverse datasets, with significant Hit@k (k=1,3,10) improvements, showcasing its ability to augment large language models (LLMs) for event prediction without necessitating extensive retraining. The ONSEP framework not only advances the field of TKGF but also underscores the potential of neural-symbolic approaches in adapting to dynamic data environments.",
}
| In the realm of event prediction, temporal knowledge graph forecasting (TKGF) stands as a pivotal technique. Previous approaches face the challenges of not utilizing experience during testing and relying on a single short-term history, which limits adaptation to evolving data. In this paper, we introduce the Online Neural-Symbolic Event Prediction (ONSEP) framework, which innovates by integrating dynamic causal rule mining (DCRM) and dual history augmented generation (DHAG). DCRM dynamically constructs causal rules from real-time data, allowing for swift adaptation to new causal relationships. In parallel, DHAG merges short-term and long-term historical contexts, leveraging a bi-branch approach to enrich event prediction. Our framework demonstrates notable performance enhancements across diverse datasets, with significant Hit@k (k=1,3,10) improvements, showcasing its ability to augment large language models (LLMs) for event prediction without necessitating extensive retraining. The ONSEP framework not only advances the field of TKGF but also underscores the potential of neural-symbolic approaches in adapting to dynamic data environments. | [
"Yu, Xuanqing",
"Sun, Wangtao",
"Li, Jingwei",
"Liu, Kang",
"Liu, Chengbao",
"Tan, Jie"
]
| {ONSEP}: A Novel Online Neural-Symbolic Framework for Event Prediction Based on Large Language Model | findings-acl.378 | Poster | 2311.17351v1 |
https://aclanthology.org/2024.findings-acl.379.bib | @inproceedings{sun-etal-2024-speech,
title = "Speech-based Slot Filling using Large Language Models",
author = "Sun, Guangzhi and
Feng, Shutong and
Jiang, Dongcheng and
Zhang, Chao and
Gasic, Milica and
Woodland, Phil",
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.379",
pages = "6351--6362",
abstract = "Recently, advancements in large language models (LLMs) have shown an unprecedented ability across various language tasks. This paper investigates the potential application of LLMs to slot filling with noisy ASR transcriptions, via both in-context learning and task-specific fine-tuning. Dedicated prompt designs and noise-robust LoRA fine-tuning are proposed to improve the robustness of LLMs for slot filling with noisy ASR transcriptions. Moreover, a linearised knowledge injection (LKI) scheme is also proposed to integrate dynamic external knowledge into LLMs. Experiments were performed on SLURP to quantify the performance of LLMs, including GPT-3.5-turbo, GPT-4, LLaMA-13B, LLaMA-2-13B and Vicuna-13B (v1.1 and v1.5) with different ASR error rates. The use of the noise-robust fine-tuning together with LKI for Vicuna-13B-v1.5 achieved 6.7{\%} and 17.6{\%} absolute SLU-F1 improvements compared to a fully fine-tuned Flan-T5-XL model on the limited data setup and the zero-shot setup respectively.",
}
| Recently, advancements in large language models (LLMs) have shown an unprecedented ability across various language tasks. This paper investigates the potential application of LLMs to slot filling with noisy ASR transcriptions, via both in-context learning and task-specific fine-tuning. Dedicated prompt designs and noise-robust LoRA fine-tuning are proposed to improve the robustness of LLMs for slot filling with noisy ASR transcriptions. Moreover, a linearised knowledge injection (LKI) scheme is also proposed to integrate dynamic external knowledge into LLMs. Experiments were performed on SLURP to quantify the performance of LLMs, including GPT-3.5-turbo, GPT-4, LLaMA-13B, LLaMA-2-13B and Vicuna-13B (v1.1 and v1.5) with different ASR error rates. The use of the noise-robust fine-tuning together with LKI for Vicuna-13B-v1.5 achieved 6.7{\%} and 17.6{\%} absolute SLU-F1 improvements compared to a fully fine-tuned Flan-T5-XL model on the limited data setup and the zero-shot setup respectively. | [
"Sun, Guangzhi",
"Feng, Shutong",
"Jiang, Dongcheng",
"Zhang, Chao",
"Gasic, Milica",
"Woodl",
", Phil"
]
| Speech-based Slot Filling using Large Language Models | findings-acl.379 | Poster | 1811.01331v2 |
https://aclanthology.org/2024.findings-acl.380.bib | @inproceedings{li-etal-2024-big,
title = "Too Big to Fail: Larger Language Models are Disproportionately Resilient to Induction of Dementia-Related Linguistic Anomalies",
author = "Li, Changye and
Sheng, Zhecheng and
Cohen, Trevor and
Pakhomov, Serguei",
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.380",
pages = "6363--6377",
abstract = "As artificial neural networks grow in complexity, understanding their inner workings becomes increasingly challenging, which is particularly important in healthcare applications. The intrinsic evaluation metrics of autoregressive neural language models (NLMs), perplexity (PPL), can reflect how {``}surprised{''} an NLM model is at novel input. PPL has been widely used to understand the behavior of NLMs. Previous findings show that changes in PPL when masking attention layers in pre-trained transformer-based NLMs reflect linguistic anomalies associated with Alzheimer{'}s disease dementia. Building upon this, we explore a novel bidirectional attention head ablation method that exhibits properties attributed to the concepts of cognitive and brain reserve in human brain studies, which postulate that people with more neurons in the brain and more efficient processing are more resilient to neurodegeneration. Our results show that larger GPT-2 models require a disproportionately larger share of attention heads to be masked/ablated to display degradation of similar magnitude to masking in smaller models. These results suggest that the attention mechanism in transformer models may present an analogue to the notions of cognitive and brain reserve and could potentially be used to model certain aspects of the progression of neurodegenerative disorders and aging.",
}
| As artificial neural networks grow in complexity, understanding their inner workings becomes increasingly challenging, which is particularly important in healthcare applications. The intrinsic evaluation metrics of autoregressive neural language models (NLMs), perplexity (PPL), can reflect how {``}surprised{''} an NLM model is at novel input. PPL has been widely used to understand the behavior of NLMs. Previous findings show that changes in PPL when masking attention layers in pre-trained transformer-based NLMs reflect linguistic anomalies associated with Alzheimer{'}s disease dementia. Building upon this, we explore a novel bidirectional attention head ablation method that exhibits properties attributed to the concepts of cognitive and brain reserve in human brain studies, which postulate that people with more neurons in the brain and more efficient processing are more resilient to neurodegeneration. Our results show that larger GPT-2 models require a disproportionately larger share of attention heads to be masked/ablated to display degradation of similar magnitude to masking in smaller models. These results suggest that the attention mechanism in transformer models may present an analogue to the notions of cognitive and brain reserve and could potentially be used to model certain aspects of the progression of neurodegenerative disorders and aging. | [
"Li, Changye",
"Sheng, Zhecheng",
"Cohen, Trevor",
"Pakhomov, Serguei"
]
| Too Big to Fail: Larger Language Models are Disproportionately Resilient to Induction of Dementia-Related Linguistic Anomalies | findings-acl.380 | Poster | 2406.02830v1 |
https://aclanthology.org/2024.findings-acl.381.bib | @inproceedings{paz-argaman-etal-2024-hesum,
title = "{H}e{S}um: a Novel Dataset for Abstractive Text Summarization in {H}ebrew",
author = "Paz-Argaman, Tzuf and
Mondshine, Itai and
Achi Mordechai, Asaf and
Tsarfaty, Reut",
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.381",
pages = "6378--6388",
abstract = "While large language models (LLMs) excel in various natural language tasks in English, their performance in low-resource languages like Hebrew, especially for generative tasks such as abstractive summarization, remains unclear. The high morphological richness in Hebrew adds further challenges due to the ambiguity in sentence comprehension and the complexities in meaning construction.In this paper, we address this evaluation and resource gap by introducing HeSum, a novel benchmark dataset specifically designed for Hebrew abstractive text summarization. HeSum consists of 10,000 article-summary pairs sourced from Hebrew news websites written by professionals. Linguistic analysis confirms HeSum{'}s high abstractness and unique morphological challenges. We show that HeSum presents distinct difficulties even for state-of-the-art LLMs, establishing it as a valuable testbed for advancing generative language technology in Hebrew, and MRLs generative challenges in general.",
}
| While large language models (LLMs) excel in various natural language tasks in English, their performance in low-resource languages like Hebrew, especially for generative tasks such as abstractive summarization, remains unclear. The high morphological richness in Hebrew adds further challenges due to the ambiguity in sentence comprehension and the complexities in meaning construction.In this paper, we address this evaluation and resource gap by introducing HeSum, a novel benchmark dataset specifically designed for Hebrew abstractive text summarization. HeSum consists of 10,000 article-summary pairs sourced from Hebrew news websites written by professionals. Linguistic analysis confirms HeSum{'}s high abstractness and unique morphological challenges. We show that HeSum presents distinct difficulties even for state-of-the-art LLMs, establishing it as a valuable testbed for advancing generative language technology in Hebrew, and MRLs generative challenges in general. | [
"Paz-Argaman, Tzuf",
"Mondshine, Itai",
"Achi Mordechai, Asaf",
"Tsarfaty, Reut"
]
| {H}e{S}um: a Novel Dataset for Abstractive Text Summarization in {H}ebrew | findings-acl.381 | Poster | 2406.03897v2 |
https://aclanthology.org/2024.findings-acl.382.bib | @inproceedings{wang-zhao-2024-tram,
title = "{TRAM}: Benchmarking Temporal Reasoning for Large Language Models",
author = "Wang, Yuqing and
Zhao, Yun",
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.382",
pages = "6389--6415",
abstract = "Reasoning about time is essential for understanding the nuances of events described in natural language. Previous research on this topic has been limited in scope, characterized by a lack of standardized benchmarks that would allow for consistent evaluations across different studies. In this paper, we introduce TRAM, a temporal reasoning benchmark composed of ten datasets, encompassing various temporal aspects of events such as order, arithmetic, frequency, and duration, designed to facilitate a comprehensive evaluation of the TeR capabilities of large language models (LLMs). We evaluate popular LLMs like GPT-4 and Llama2 in zero-shot and few-shot scenarios, and establish baselines with BERT-based and domain-specific models. Our findings indicate that the best-performing model lags significantly behind human performance. It is our aspiration that TRAM will spur further progress in enhancing the TeR capabilities of LLMs.",
}
| Reasoning about time is essential for understanding the nuances of events described in natural language. Previous research on this topic has been limited in scope, characterized by a lack of standardized benchmarks that would allow for consistent evaluations across different studies. In this paper, we introduce TRAM, a temporal reasoning benchmark composed of ten datasets, encompassing various temporal aspects of events such as order, arithmetic, frequency, and duration, designed to facilitate a comprehensive evaluation of the TeR capabilities of large language models (LLMs). We evaluate popular LLMs like GPT-4 and Llama2 in zero-shot and few-shot scenarios, and establish baselines with BERT-based and domain-specific models. Our findings indicate that the best-performing model lags significantly behind human performance. It is our aspiration that TRAM will spur further progress in enhancing the TeR capabilities of LLMs. | [
"Wang, Yuqing",
"Zhao, Yun"
]
| {TRAM}: Benchmarking Temporal Reasoning for Large Language Models | findings-acl.382 | Poster | 2406.09072v1 |
https://aclanthology.org/2024.findings-acl.383.bib | @inproceedings{amayuelas-etal-2024-knowledge,
title = "Knowledge of Knowledge: Exploring Known-Unknowns Uncertainty with Large Language Models",
author = "Amayuelas, Alfonso and
Wong, Kyle and
Pan, Liangming and
Chen, Wenhu and
Wang, William 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.383",
pages = "6416--6432",
abstract = "This paper investigates the capabilities of Large Language Models (LLMs) in understanding their knowledge and uncertainty over questions. Specifically, we focus on addressing known-unknown questions, characterized by high uncertainty due to the absence of definitive answers. To facilitate our study, we collect a new dataset with Known-Unknown Questions (KUQ) and establish a categorization framework to clarify the origins of uncertainty in such queries. Subsequently, we examine the performance of open-source LLMs, fine-tuned using this dataset, in distinguishing between known and unknown queries within open-ended question-answering scenarios. The fine-tuned models demonstrated a significant improvement, achieving a considerable increase in F1-score relative to their pre-fine-tuning state. Through a comprehensive analysis, we reveal insights into the models{'} improved uncertainty articulation and their consequent efficacy in multi-agent debates. These findings help us understand how LLMs can be trained to identify and express uncertainty, improving our knowledge of how they understand and express complex or unclear information.",
}
| This paper investigates the capabilities of Large Language Models (LLMs) in understanding their knowledge and uncertainty over questions. Specifically, we focus on addressing known-unknown questions, characterized by high uncertainty due to the absence of definitive answers. To facilitate our study, we collect a new dataset with Known-Unknown Questions (KUQ) and establish a categorization framework to clarify the origins of uncertainty in such queries. Subsequently, we examine the performance of open-source LLMs, fine-tuned using this dataset, in distinguishing between known and unknown queries within open-ended question-answering scenarios. The fine-tuned models demonstrated a significant improvement, achieving a considerable increase in F1-score relative to their pre-fine-tuning state. Through a comprehensive analysis, we reveal insights into the models{'} improved uncertainty articulation and their consequent efficacy in multi-agent debates. These findings help us understand how LLMs can be trained to identify and express uncertainty, improving our knowledge of how they understand and express complex or unclear information. | [
"Amayuelas, Alfonso",
"Wong, Kyle",
"Pan, Liangming",
"Chen, Wenhu",
"Wang, William Yang"
]
| Knowledge of Knowledge: Exploring Known-Unknowns Uncertainty with Large Language Models | findings-acl.383 | Poster | 2003.10775v2 |
https://aclanthology.org/2024.findings-acl.384.bib | @inproceedings{cui-etal-2024-exploring,
title = "Exploring Defeasibility in Causal Reasoning",
author = "Cui, Shaobo and
Milikic, Lazar and
Feng, Yiyang and
Ismayilzada, Mete and
Paul, Debjit and
Bosselut, Antoine and
Faltings, Boi",
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.384",
pages = "6433--6452",
abstract = "Defeasibility in causal reasoning implies that the causal relationship between cause and effect can be strengthened or weakened. Namely, the causal strength between cause and effect should increase or decrease with the incorporation of strengthening arguments (supporters) or weakening arguments (defeaters), respectively. However, existing works ignore defeasibility in causal reasoning and fail to evaluate existing causal strength metrics in defeasible settings. In this work, we present $\delta$-CAUSAL, the first benchmark dataset for studying defeasibility in causal reasoning. $\delta$-CAUSAL includes around 11K events spanning ten domains, featuring defeasible causality pairs, namely, cause-effect pairs accompanied by supporters and defeaters. We further show that current causal strength metrics fail to reflect the change of causal strength with the incorporation of supporters or defeaters in $\delta$-CAUSAL. To this end, we propose CESAR (Causal Embedding aSsociation with Attention Rating), a metric that measures causal strength based on token-level causal relationships. CESAR achieves a significant 69.7{\%} relative improvement over existing metrics, increasing from 47.2{\%} to 80.1{\%} in capturing the causal strength change brought by supporters and defeaters. We further demonstrate even Large Language Models (LLMs) like GPT-3.5 still lag 4.5 and 10.7 points behind humans in generating supporters and defeaters, emphasizing the challenge posed by $\delta$-CAUSAL.",
}
| Defeasibility in causal reasoning implies that the causal relationship between cause and effect can be strengthened or weakened. Namely, the causal strength between cause and effect should increase or decrease with the incorporation of strengthening arguments (supporters) or weakening arguments (defeaters), respectively. However, existing works ignore defeasibility in causal reasoning and fail to evaluate existing causal strength metrics in defeasible settings. In this work, we present $\delta$-CAUSAL, the first benchmark dataset for studying defeasibility in causal reasoning. $\delta$-CAUSAL includes around 11K events spanning ten domains, featuring defeasible causality pairs, namely, cause-effect pairs accompanied by supporters and defeaters. We further show that current causal strength metrics fail to reflect the change of causal strength with the incorporation of supporters or defeaters in $\delta$-CAUSAL. To this end, we propose CESAR (Causal Embedding aSsociation with Attention Rating), a metric that measures causal strength based on token-level causal relationships. CESAR achieves a significant 69.7{\%} relative improvement over existing metrics, increasing from 47.2{\%} to 80.1{\%} in capturing the causal strength change brought by supporters and defeaters. We further demonstrate even Large Language Models (LLMs) like GPT-3.5 still lag 4.5 and 10.7 points behind humans in generating supporters and defeaters, emphasizing the challenge posed by $\delta$-CAUSAL. | [
"Cui, Shaobo",
"Milikic, Lazar",
"Feng, Yiyang",
"Ismayilzada, Mete",
"Paul, Debjit",
"Bosselut, Antoine",
"Faltings, Boi"
]
| Exploring Defeasibility in Causal Reasoning | findings-acl.384 | Poster | 2401.03183v2 |
https://aclanthology.org/2024.findings-acl.385.bib | @inproceedings{gandhi-etal-2024-better,
title = "Better Synthetic Data by Retrieving and Transforming Existing Datasets",
author = "Gandhi, Saumya and
Gala, Ritu and
Viswanathan, Vijay and
Wu, Tongshuang and
Neubig, Graham",
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.385",
pages = "6453--6466",
abstract = "Despite recent advances in large language models, building dependable and deployable NLP models typically requires abundant, high-quality training data. However, task-specific data is not available for many use cases, and manually curating task-specific data is labor-intensive. Recent work has studied prompt-driven synthetic data generation using large language models, but these generated datasets tend to lack complexity and diversity. To address these limitations, we introduce a method, {\_}DataTune{\_}, to make better use of existing, publicly available datasets to improve automatic dataset generation. DataTune performs {\_}dataset transformation{\_}, enabling the repurposing of publicly available datasets into a format that is directly aligned with the specific requirements of target tasks. On a diverse set of language-based tasks from the BIG-Bench benchmark, we find that finetuning language models via DataTune improves over a few-shot prompting baseline by 49{\%} and improves over existing methods that use synthetic or retrieved training data by 34{\%}. We find that dataset transformation significantly increases the diversity and difficulty of generated data on many tasks. We release a Python package and open-source repository to make this method accessible to the community (URL will be added upon acceptance).",
}
| Despite recent advances in large language models, building dependable and deployable NLP models typically requires abundant, high-quality training data. However, task-specific data is not available for many use cases, and manually curating task-specific data is labor-intensive. Recent work has studied prompt-driven synthetic data generation using large language models, but these generated datasets tend to lack complexity and diversity. To address these limitations, we introduce a method, {\_}DataTune{\_}, to make better use of existing, publicly available datasets to improve automatic dataset generation. DataTune performs {\_}dataset transformation{\_}, enabling the repurposing of publicly available datasets into a format that is directly aligned with the specific requirements of target tasks. On a diverse set of language-based tasks from the BIG-Bench benchmark, we find that finetuning language models via DataTune improves over a few-shot prompting baseline by 49{\%} and improves over existing methods that use synthetic or retrieved training data by 34{\%}. We find that dataset transformation significantly increases the diversity and difficulty of generated data on many tasks. We release a Python package and open-source repository to make this method accessible to the community (URL will be added upon acceptance). | [
"G",
"hi, Saumya",
"Gala, Ritu",
"Viswanathan, Vijay",
"Wu, Tongshuang",
"Neubig, Graham"
]
| Better Synthetic Data by Retrieving and Transforming Existing Datasets | findings-acl.385 | Poster | 2404.14361v3 |
https://aclanthology.org/2024.findings-acl.386.bib | @inproceedings{xiang-etal-2024-addressing,
title = "Addressing Order Sensitivity of In-Context Demonstration Examples in Causal Language Models",
author = "Xiang, Yanzheng and
Yan, Hanqi 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.386",
pages = "6467--6481",
abstract = "In-context learning has become a popular paradigm in natural language processing. However, its performance can be significantly influenced by the order of in-context demonstration examples. In this paper, we found that causal language models (CausalLMs) are more sensitive to this order compared to prefix language models (PrefixLMs). We attribute this phenomenon to the auto-regressive attention masks within CausalLMs, which restrict each token from accessing information from subsequent tokens. This results in different receptive fields for samples at different positions, thereby leading to representation disparities across positions. To tackle this challenge, we introduce an unsupervised fine-tuning method, termed the Information-Augmented and Consistency-Enhanced approach. This approach utilizes contrastive learning to align representations of in-context examples across different positions and introduces a consistency loss to ensure similar representations for inputs with different permutations. This enhances the model{'}s predictive consistency across permutations. Experimental results on five benchmarks suggest that our proposed method can reduce the sensitivity of CausalLMs to the order of in-context examples and exhibit robust generalizability, particularly when demonstrations are sourced from a candidate pool different from that used in the training phase, or when the number of in-context examples differs from what is used during training.",
}
| In-context learning has become a popular paradigm in natural language processing. However, its performance can be significantly influenced by the order of in-context demonstration examples. In this paper, we found that causal language models (CausalLMs) are more sensitive to this order compared to prefix language models (PrefixLMs). We attribute this phenomenon to the auto-regressive attention masks within CausalLMs, which restrict each token from accessing information from subsequent tokens. This results in different receptive fields for samples at different positions, thereby leading to representation disparities across positions. To tackle this challenge, we introduce an unsupervised fine-tuning method, termed the Information-Augmented and Consistency-Enhanced approach. This approach utilizes contrastive learning to align representations of in-context examples across different positions and introduces a consistency loss to ensure similar representations for inputs with different permutations. This enhances the model{'}s predictive consistency across permutations. Experimental results on five benchmarks suggest that our proposed method can reduce the sensitivity of CausalLMs to the order of in-context examples and exhibit robust generalizability, particularly when demonstrations are sourced from a candidate pool different from that used in the training phase, or when the number of in-context examples differs from what is used during training. | [
"Xiang, Yanzheng",
"Yan, Hanqi",
"Gui, Lin",
"He, Yulan"
]
| Addressing Order Sensitivity of In-Context Demonstration Examples in Causal Language Models | findings-acl.386 | Poster | 2402.15637v2 |
https://aclanthology.org/2024.findings-acl.387.bib | @inproceedings{plepi-etal-2024-perspective,
title = "Perspective Taking through Generating Responses to Conflict Situations",
author = "Plepi, Joan and
Welch, Charles and
Flek, Lucie",
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.387",
pages = "6482--6497",
abstract = "Although language model performance across diverse tasks continues to improve, these models still struggle to understand and explain the beliefs of other people. This skill requires perspective-taking, the process of conceptualizing the point of view of another person. Perspective taking becomes challenging when the text reflects more personal and potentially more controversial beliefs.We explore this task through natural language generation of responses to conflict situations. We evaluate novel modifications to recent architectures for conditioning generation on an individual{'}s comments and self-disclosure statements. Our work extends the Social-Chem-101 corpus, using 95k judgements written by 6k authors from English Reddit data, for each of whom we obtained 20-500 self-disclosure statements. Our evaluation methodology borrows ideas from both personalized generation and theory of mind literature. Our proposed perspective-taking models outperform recent work, especially the twin encoder model conditioned on self-disclosures with high similarity to the conflict situation.",
}
| Although language model performance across diverse tasks continues to improve, these models still struggle to understand and explain the beliefs of other people. This skill requires perspective-taking, the process of conceptualizing the point of view of another person. Perspective taking becomes challenging when the text reflects more personal and potentially more controversial beliefs.We explore this task through natural language generation of responses to conflict situations. We evaluate novel modifications to recent architectures for conditioning generation on an individual{'}s comments and self-disclosure statements. Our work extends the Social-Chem-101 corpus, using 95k judgements written by 6k authors from English Reddit data, for each of whom we obtained 20-500 self-disclosure statements. Our evaluation methodology borrows ideas from both personalized generation and theory of mind literature. Our proposed perspective-taking models outperform recent work, especially the twin encoder model conditioned on self-disclosures with high similarity to the conflict situation. | [
"Plepi, Joan",
"Welch, Charles",
"Flek, Lucie"
]
| Perspective Taking through Generating Responses to Conflict Situations | findings-acl.387 | Poster | 2310.00935v1 |
https://aclanthology.org/2024.findings-acl.388.bib | @inproceedings{lee-etal-2024-llm2llm,
title = "{LLM}2{LLM}: Boosting {LLM}s with Novel Iterative Data Enhancement",
author = "Lee, Nicholas and
Wattanawong, Thanakul and
Kim, Sehoon and
Mangalam, Karttikeya and
Shen, Sheng and
Anumanchipalli, Gopala and
Mahoney, Michael and
Keutzer, Kurt and
Gholami, Amir",
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.388",
pages = "6498--6526",
abstract = "Pretrained large language models (LLMs) are currently state-of-the-art for solving the vast majority of natural language processing tasks. While many real-world applications still require fine-tuning to reach satisfactory levels of performance, many of them are in the low-data regime, making fine-tuning challenging. To address this, we propose LLM2LLM, a targeted and iterative data augmentation strategy that uses a teacher LLM to enhance a small seed dataset by augmenting additional data that can be used for fine-tuning on a specific task. LLM2LLM (1) fine-tunes a baseline student LLM on the initial seed data, (2) evaluates and extracts data points that the model gets wrong, and (3) uses a teacher LLM to generate synthetic data based on these incorrect data points, which are then added back into the training data. This approach amplifies the signal from incorrectly predicted data points by the LLM during training and reintegrates them into the dataset to focus on more challenging examples for the LLM. Our results show that LLM2LLM significantly enhances the performance of LLMs in the low-data regime, outperforming both traditional fine-tuning and other data augmentation baselines. LLM2LLM reduces the dependence on labor-intensive data curation and paves the way for more scalable and performant LLM solutions, allowing us to tackle data-constrained domains and tasks. We achieve improvements up to 24.2{\%} on the GSM8K dataset, 32.6{\%} on CaseHOLD, 32.0{\%} on SNIPS, 52.6{\%} on TREC and 39.8{\%} on SST-2 over regular fine-tuning in the low-data regime using a Llama-2-7B student model. Our code is available at https://github.com/SqueezeAILab/LLM2LLM.",
}
| Pretrained large language models (LLMs) are currently state-of-the-art for solving the vast majority of natural language processing tasks. While many real-world applications still require fine-tuning to reach satisfactory levels of performance, many of them are in the low-data regime, making fine-tuning challenging. To address this, we propose LLM2LLM, a targeted and iterative data augmentation strategy that uses a teacher LLM to enhance a small seed dataset by augmenting additional data that can be used for fine-tuning on a specific task. LLM2LLM (1) fine-tunes a baseline student LLM on the initial seed data, (2) evaluates and extracts data points that the model gets wrong, and (3) uses a teacher LLM to generate synthetic data based on these incorrect data points, which are then added back into the training data. This approach amplifies the signal from incorrectly predicted data points by the LLM during training and reintegrates them into the dataset to focus on more challenging examples for the LLM. Our results show that LLM2LLM significantly enhances the performance of LLMs in the low-data regime, outperforming both traditional fine-tuning and other data augmentation baselines. LLM2LLM reduces the dependence on labor-intensive data curation and paves the way for more scalable and performant LLM solutions, allowing us to tackle data-constrained domains and tasks. We achieve improvements up to 24.2{\%} on the GSM8K dataset, 32.6{\%} on CaseHOLD, 32.0{\%} on SNIPS, 52.6{\%} on TREC and 39.8{\%} on SST-2 over regular fine-tuning in the low-data regime using a Llama-2-7B student model. Our code is available at https://github.com/SqueezeAILab/LLM2LLM. | [
"Lee, Nicholas",
"Wattanawong, Thanakul",
"Kim, Sehoon",
"Mangalam, Karttikeya",
"Shen, Sheng",
"Anumanchipalli, Gopala",
"Mahoney, Michael",
"Keutzer, Kurt",
"Gholami, Amir"
]
| {LLM}2{LLM}: Boosting {LLM}s with Novel Iterative Data Enhancement | findings-acl.388 | Poster | 2402.12146v3 |
https://aclanthology.org/2024.findings-acl.389.bib | @inproceedings{ernst-etal-2024-power,
title = "The Power of Summary-Source Alignments",
author = "Ernst, Ori and
Shapira, Ori and
Slobodkin, Aviv and
Adar, Sharon and
Bansal, Mohit and
Goldberger, Jacob and
Levy, Ran and
Dagan, Ido",
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.389",
pages = "6527--6548",
abstract = "Multi-document summarization (MDS) is a challenging task, often decomposed to subtasks of salience and redundancy detection, followed by text generation.In this context, alignment of corresponding sentences between a reference summary and its source documents has been leveraged to generate training data for some of the component tasks. Yet, this enabling alignment step has usually been applied heuristically on the sentence level on a limited number of subtasks.In this paper, we propose extending the summary-source alignment framework by (1) applying it at the more fine-grained proposition span level, (2) annotating alignment manually in a multi-document setup, and (3) revealing the great potential of summary-source alignments to yield several datasets for at least six different tasks. Specifically, for each of the tasks, we release a manually annotated test set that was derived automatically from the alignment annotation. We also release development and train sets in the same way, but from automatically derived alignments.Using the datasets, each task is demonstrated with baseline models and corresponding evaluation metrics to spur future research on this broad challenge.",
}
| Multi-document summarization (MDS) is a challenging task, often decomposed to subtasks of salience and redundancy detection, followed by text generation.In this context, alignment of corresponding sentences between a reference summary and its source documents has been leveraged to generate training data for some of the component tasks. Yet, this enabling alignment step has usually been applied heuristically on the sentence level on a limited number of subtasks.In this paper, we propose extending the summary-source alignment framework by (1) applying it at the more fine-grained proposition span level, (2) annotating alignment manually in a multi-document setup, and (3) revealing the great potential of summary-source alignments to yield several datasets for at least six different tasks. Specifically, for each of the tasks, we release a manually annotated test set that was derived automatically from the alignment annotation. We also release development and train sets in the same way, but from automatically derived alignments.Using the datasets, each task is demonstrated with baseline models and corresponding evaluation metrics to spur future research on this broad challenge. | [
"Ernst, Ori",
"Shapira, Ori",
"Slobodkin, Aviv",
"Adar, Sharon",
"Bansal, Mohit",
"Goldberger, Jacob",
"Levy, Ran",
"Dagan, Ido"
]
| The Power of Summary-Source Alignments | findings-acl.389 | Poster | 2303.08494v1 |
https://aclanthology.org/2024.findings-acl.390.bib | @inproceedings{bhatt-etal-2024-experimental,
title = "An Experimental Design Framework for Label-Efficient Supervised Finetuning of Large Language Models",
author = "Bhatt, Gantavya and
Chen, Yifang and
Das, Arnav and
Zhang, Jifan and
Truong, Sang and
Mussmann, Stephen and
Zhu, Yinglun and
Bilmes, Jeff and
Du, Simon and
Jamieson, Kevin and
Ash, Jordan and
Nowak, Robert",
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.390",
pages = "6549--6560",
abstract = "Supervised finetuning (SFT) on instruction datasets has played a crucial role in achieving the remarkable zero-shot generalization capabilities observed in modern large language models (LLMs). However, the annotation efforts required to produce high quality responses for instructions are becoming prohibitively expensive, especially as the number of tasks spanned by instruction datasets continues to increase. Active learning is effective in identifying useful subsets of samples to annotate from an unlabeled pool, but its high computational cost remains a barrier to its widespread applicability in the context of LLMs. To mitigate the annotation cost of SFT and circumvent the computational bottlenecks of active learning, we propose using experimental design. Experimental design techniques select the most informative samples to label, and typically maximize some notion of uncertainty and/or diversity. In our work, we implement a framework that evaluates several existing and novel experimental design techniques and find that these methods consistently yield significant gains in label efficiency with little computational overhead. On generative tasks, to reach the same generalization performance, our methods save 50{\%} of the annotation cost compared to random sampling.",
}
| Supervised finetuning (SFT) on instruction datasets has played a crucial role in achieving the remarkable zero-shot generalization capabilities observed in modern large language models (LLMs). However, the annotation efforts required to produce high quality responses for instructions are becoming prohibitively expensive, especially as the number of tasks spanned by instruction datasets continues to increase. Active learning is effective in identifying useful subsets of samples to annotate from an unlabeled pool, but its high computational cost remains a barrier to its widespread applicability in the context of LLMs. To mitigate the annotation cost of SFT and circumvent the computational bottlenecks of active learning, we propose using experimental design. Experimental design techniques select the most informative samples to label, and typically maximize some notion of uncertainty and/or diversity. In our work, we implement a framework that evaluates several existing and novel experimental design techniques and find that these methods consistently yield significant gains in label efficiency with little computational overhead. On generative tasks, to reach the same generalization performance, our methods save 50{\%} of the annotation cost compared to random sampling. | [
"Bhatt, Gantavya",
"Chen, Yifang",
"Das, Arnav",
"Zhang, Jifan",
"Truong, Sang",
"Mussmann, Stephen",
"Zhu, Yinglun",
"Bilmes, Jeff",
"Du, Simon",
"Jamieson, Kevin",
"Ash, Jordan",
"Nowak, Robert"
]
| An Experimental Design Framework for Label-Efficient Supervised Finetuning of Large Language Models | findings-acl.390 | Poster | 2407.02770v1 |
https://aclanthology.org/2024.findings-acl.391.bib | @inproceedings{tian-etal-2024-learning,
title = "Learning Multimodal Contrast with Cross-modal Memory and Reinforced Contrast Recognition",
author = "Tian, Yuanhe and
Xia, Fei and
Song, Yan",
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.391",
pages = "6561--6573",
abstract = "In many practical scenarios, contents from different modalities are not semantically aligned; for instance, visual and textual information may conflict with each other, resulting in non-compositional expression effects such as irony or humor. Effective modeling and smooth integration of multimodal information are crucial for achieving good understanding of the contrast across modalities. Being focusing on image-text matching, most current studies face challenges in identifying such contrast, leading to limitations in exploring the extended semantics when images and texts do not match. In this paper, we propose an LLM-based approach for learning multimodal contrast following the encoding-decoding paradigm, enhanced by a memory module with reinforced contrast recognition, and use a series of tasks that have the nature of multimodal contrast to verify our approach. The memory module learns the integration between visual and textual features with trainable memory vectors and the reinforced contrast recognition uses self-rejection sampling to optimize the memory to further enhance learning multimodal contrast. The resulted information, accompanied with visual and text features, is finally fed into the LLM to predict corresponding labels. We experiment our approach on four English and Chinese benchmark datasets, where it outperforms strong baselines and state-of-the-art studies.",
}
| In many practical scenarios, contents from different modalities are not semantically aligned; for instance, visual and textual information may conflict with each other, resulting in non-compositional expression effects such as irony or humor. Effective modeling and smooth integration of multimodal information are crucial for achieving good understanding of the contrast across modalities. Being focusing on image-text matching, most current studies face challenges in identifying such contrast, leading to limitations in exploring the extended semantics when images and texts do not match. In this paper, we propose an LLM-based approach for learning multimodal contrast following the encoding-decoding paradigm, enhanced by a memory module with reinforced contrast recognition, and use a series of tasks that have the nature of multimodal contrast to verify our approach. The memory module learns the integration between visual and textual features with trainable memory vectors and the reinforced contrast recognition uses self-rejection sampling to optimize the memory to further enhance learning multimodal contrast. The resulted information, accompanied with visual and text features, is finally fed into the LLM to predict corresponding labels. We experiment our approach on four English and Chinese benchmark datasets, where it outperforms strong baselines and state-of-the-art studies. | [
"Tian, Yuanhe",
"Xia, Fei",
"Song, Yan"
]
| Learning Multimodal Contrast with Cross-modal Memory and Reinforced Contrast Recognition | findings-acl.391 | Poster | 2401.17032v2 |
https://aclanthology.org/2024.findings-acl.392.bib | @inproceedings{bahrainian-etal-2024-text,
title = "Text Simplification via Adaptive Teaching",
author = "Bahrainian, Seyed Ali and
Dou, Jonathan and
Eickhoff, Carsten",
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.392",
pages = "6574--6584",
abstract = "Text simplification is the process of rewriting a piece of text using simpler vocabulary and grammatical structure in order to make the text more accessible and understandable for a larger audience. In this paper, we introduce a new text simplification model based on the notion of adaptive teaching using a teacher network and a text generation network. We name this new model Simplification via Adaptive Teaching (SAT). Our proposed model sets a new state-of-the-art performance in terms of standard simplification metrics such as SARI and D-SARI with a significant improvement over the previous state of the art on the D-Wikipedia dataset and the Wiki-Doc benchmark dataset. Moreover, we conduct a human evaluation in terms of text simplicity, correctness, and fluency to substantiate SAT{'}s performance.",
}
| Text simplification is the process of rewriting a piece of text using simpler vocabulary and grammatical structure in order to make the text more accessible and understandable for a larger audience. In this paper, we introduce a new text simplification model based on the notion of adaptive teaching using a teacher network and a text generation network. We name this new model Simplification via Adaptive Teaching (SAT). Our proposed model sets a new state-of-the-art performance in terms of standard simplification metrics such as SARI and D-SARI with a significant improvement over the previous state of the art on the D-Wikipedia dataset and the Wiki-Doc benchmark dataset. Moreover, we conduct a human evaluation in terms of text simplicity, correctness, and fluency to substantiate SAT{'}s performance. | [
"Bahrainian, Seyed Ali",
"Dou, Jonathan",
"Eickhoff, Carsten"
]
| Text Simplification via Adaptive Teaching | findings-acl.392 | Poster | 2305.12463v1 |
https://aclanthology.org/2024.findings-acl.393.bib | @inproceedings{gokceoglu-etal-2024-multi,
title = "A multi-level multi-label text classification dataset of 19th century Ottoman and {R}ussian literary and critical texts",
author = {Gokceoglu, Gokcen and
{\c{C}}avu{\c{s}}o{\u{g}}lu, Devrim and
Akbas, Emre and
Dolcerocca, {\"O}zen},
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.393",
pages = "6585--6596",
abstract = "This paper introduces a multi-level, multi-label text classification dataset comprising over 3000 documents. The dataset features literary and critical texts from 19th-century Ottoman Turkish and Russian. It is the first study to apply large language models (LLMs) to this dataset, sourced from prominent literary periodicals of the era. The texts have been meticulously organized and labeled. This was done according to a taxonomic framework that takes into account both their structural and semantic attributes. Articles are categorized and tagged with bibliometric metadata by human experts. We present baseline classification results using a classical bag-of-words (BoW) naive Bayes model and three modern LLMs: multilingual BERT, Falcon, and Llama-v2. We found that in certain cases, Bag of Words (BoW) outperforms Large Language Models (LLMs), emphasizing the need for additional research, especially in low-resource language settings. This dataset is expected to be a valuable resource for researchers in natural language processing and machine learning, especially for historical and low-resource languages. The dataset is publicly available.",
}
| This paper introduces a multi-level, multi-label text classification dataset comprising over 3000 documents. The dataset features literary and critical texts from 19th-century Ottoman Turkish and Russian. It is the first study to apply large language models (LLMs) to this dataset, sourced from prominent literary periodicals of the era. The texts have been meticulously organized and labeled. This was done according to a taxonomic framework that takes into account both their structural and semantic attributes. Articles are categorized and tagged with bibliometric metadata by human experts. We present baseline classification results using a classical bag-of-words (BoW) naive Bayes model and three modern LLMs: multilingual BERT, Falcon, and Llama-v2. We found that in certain cases, Bag of Words (BoW) outperforms Large Language Models (LLMs), emphasizing the need for additional research, especially in low-resource language settings. This dataset is expected to be a valuable resource for researchers in natural language processing and machine learning, especially for historical and low-resource languages. The dataset is publicly available. | [
"Gokceoglu, Gokcen",
"{\\c{C}}avu{\\c{s}}o{\\u{g}}lu, Devrim",
"Akbas, Emre",
"Dolcerocca, {\\\"O}zen"
]
| A multi-level multi-label text classification dataset of 19th century Ottoman and {R}ussian literary and critical texts | findings-acl.393 | Poster | 2407.15136v1 |
https://aclanthology.org/2024.findings-acl.394.bib | @inproceedings{cabello-akujuobi-2024-simple,
title = "It is Simple Sometimes: A Study On Improving Aspect-Based Sentiment Analysis Performance",
author = "Cabello, Laura and
Akujuobi, Uchenna",
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.394",
pages = "6597--6610",
abstract = "Aspect-Based Sentiment Analysis (ABSA) involves extracting opinions from textual data about specific entities and their corresponding aspects through various complementary subtasks. Several prior research has focused on developing ad hoc designs of varying complexities for these subtasks. In this paper, we build upon the instruction tuned model proposed by Scaria et al. (2023), who present an instruction-based model with task descriptions followed by in-context examples on ABSA subtasks. We propose PFInstruct, an extension to this instruction learning paradigm by appending an NLP-related task prefix to the task description. This simple approach leads to improved performance across all tested SemEval subtasks, surpassing previous state-of-the-art (SOTA) on the ATE subtask (Rest14) by +3.28 F1-score, and on the AOOE subtask by an average of +5.43 F1-score across SemEval datasets. Furthermore, we explore the impact of the prefix-enhanced prompt quality on the ABSA subtasks and find that even a noisy prefix enhances model performance compared to the baseline. Our method also achieves competitive results on a biomedical domain dataset (ERSA).",
}
| Aspect-Based Sentiment Analysis (ABSA) involves extracting opinions from textual data about specific entities and their corresponding aspects through various complementary subtasks. Several prior research has focused on developing ad hoc designs of varying complexities for these subtasks. In this paper, we build upon the instruction tuned model proposed by Scaria et al. (2023), who present an instruction-based model with task descriptions followed by in-context examples on ABSA subtasks. We propose PFInstruct, an extension to this instruction learning paradigm by appending an NLP-related task prefix to the task description. This simple approach leads to improved performance across all tested SemEval subtasks, surpassing previous state-of-the-art (SOTA) on the ATE subtask (Rest14) by +3.28 F1-score, and on the AOOE subtask by an average of +5.43 F1-score across SemEval datasets. Furthermore, we explore the impact of the prefix-enhanced prompt quality on the ABSA subtasks and find that even a noisy prefix enhances model performance compared to the baseline. Our method also achieves competitive results on a biomedical domain dataset (ERSA). | [
"Cabello, Laura",
"Akujuobi, Uchenna"
]
| It is Simple Sometimes: A Study On Improving Aspect-Based Sentiment Analysis Performance | findings-acl.394 | Poster | 2010.11731v2 |
https://aclanthology.org/2024.findings-acl.395.bib | @inproceedings{he-etal-2024-whose,
title = "Whose Emotions and Moral Sentiments do Language Models Reflect?",
author = "He, Zihao and
Guo, Siyi and
Rao, Ashwin and
Lerman, Kristina",
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.395",
pages = "6611--6631",
abstract = "Language models (LMs) are known to represent the perspectives of some social groups better than others, which may impact their performance, especially on subjective tasks such as content moderation and hate speech detection. To explore how LMs represent different perspectives, existing research focused on positional alignment, i.e., how closely the models mimic the opinions and stances of different groups, e.g., liberals or conservatives. However, human communication also encompasses emotional and moral dimensions. We define the problem of affective alignment, which measures how LMs{'} emotional and moral tone represents those of different groups. By comparing the affect of responses generated by 36 LMs to the affect of Twitter messages written by two ideological groups, we observe significant misalignment of LMs with both ideological groups. This misalignment is larger than the partisan divide in the U.S. Even after steering the LMs towards specific ideological perspectives, the misalignment and liberal tendencies of the model persist, suggesting a systemic bias within LMs.",
}
| Language models (LMs) are known to represent the perspectives of some social groups better than others, which may impact their performance, especially on subjective tasks such as content moderation and hate speech detection. To explore how LMs represent different perspectives, existing research focused on positional alignment, i.e., how closely the models mimic the opinions and stances of different groups, e.g., liberals or conservatives. However, human communication also encompasses emotional and moral dimensions. We define the problem of affective alignment, which measures how LMs{'} emotional and moral tone represents those of different groups. By comparing the affect of responses generated by 36 LMs to the affect of Twitter messages written by two ideological groups, we observe significant misalignment of LMs with both ideological groups. This misalignment is larger than the partisan divide in the U.S. Even after steering the LMs towards specific ideological perspectives, the misalignment and liberal tendencies of the model persist, suggesting a systemic bias within LMs. | [
"He, Zihao",
"Guo, Siyi",
"Rao, Ashwin",
"Lerman, Kristina"
]
| Whose Emotions and Moral Sentiments do Language Models Reflect? | findings-acl.395 | Poster | 2402.11114v2 |
https://aclanthology.org/2024.findings-acl.396.bib | @inproceedings{wang-etal-2024-llm-achieve,
title = "{LLM} can Achieve Self-Regulation via Hyperparameter Aware Generation",
author = "Wang, Siyin and
Li, Shimin and
Sun, Tianxiang and
Fu, Jinlan and
Cheng, Qinyuan and
Ye, Jiasheng and
Ye, Junjie and
Qiu, Xipeng and
Huang, Xuanjing",
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.396",
pages = "6632--6646",
abstract = "In the realm of Large Language Models (LLMs), users commonly employ diverse decoding strategies and adjust hyperparameters to control the generated text. However, a critical question emerges: Are LLMs conscious of the existence of these decoding strategies and capable of regulating themselves? The current decoding generation process often relies on empirical and heuristic manual adjustments to hyperparameters based on types of tasks and demands. However, this process is typically cumbersome, and the decoding hyperparameters may not always be optimal for each sample. To address the aforementioned challenges, we propose a novel text generation paradigm termed Hyperparameter Aware Generation (HAG). By leveraging hyperparameter-aware instruction tuning, the LLM autonomously determines the optimal decoding strategy and configs based on the input samples, enabling self-regulation. Our approach eliminates the need for extensive manual tuning, offering a more autonomous, self-regulate model behavior. Experimental results spanning six datasets across reasoning, creativity, translation, and mathematics tasks demonstrate that hyperparameter-aware instruction tuning empowers the LLMs to self-regulate the decoding strategy and hyperparameter. HAG extends the current paradigm in the text generation process, highlighting the feasibility of endowing the LLMs with self-regulate decoding strategies.",
}
| In the realm of Large Language Models (LLMs), users commonly employ diverse decoding strategies and adjust hyperparameters to control the generated text. However, a critical question emerges: Are LLMs conscious of the existence of these decoding strategies and capable of regulating themselves? The current decoding generation process often relies on empirical and heuristic manual adjustments to hyperparameters based on types of tasks and demands. However, this process is typically cumbersome, and the decoding hyperparameters may not always be optimal for each sample. To address the aforementioned challenges, we propose a novel text generation paradigm termed Hyperparameter Aware Generation (HAG). By leveraging hyperparameter-aware instruction tuning, the LLM autonomously determines the optimal decoding strategy and configs based on the input samples, enabling self-regulation. Our approach eliminates the need for extensive manual tuning, offering a more autonomous, self-regulate model behavior. Experimental results spanning six datasets across reasoning, creativity, translation, and mathematics tasks demonstrate that hyperparameter-aware instruction tuning empowers the LLMs to self-regulate the decoding strategy and hyperparameter. HAG extends the current paradigm in the text generation process, highlighting the feasibility of endowing the LLMs with self-regulate decoding strategies. | [
"Wang, Siyin",
"Li, Shimin",
"Sun, Tianxiang",
"Fu, Jinlan",
"Cheng, Qinyuan",
"Ye, Jiasheng",
"Ye, Junjie",
"Qiu, Xipeng",
"Huang, Xuanjing"
]
| {LLM} can Achieve Self-Regulation via Hyperparameter Aware Generation | findings-acl.396 | Poster | 2402.11251v1 |
https://aclanthology.org/2024.findings-acl.397.bib | @inproceedings{jiang-etal-2024-forward,
title = "Forward-Backward Reasoning in Large Language Models for Mathematical Verification",
author = "Jiang, Weisen and
Shi, Han and
Yu, Longhui and
Liu, Zhengying and
Zhang, Yu and
Li, Zhenguo and
Kwok, James",
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.397",
pages = "6647--6661",
abstract = "Self-Consistency samples diverse reasoning chains with answers and chooses the final answer by majority voting. It is based on forward reasoning and cannot further improve performance by sampling more reasoning chains when saturated. To further boost performance, we introduce backward reasoning to verify candidate answers. Specifically, for mathematical tasks, we mask a number in the question and ask the LLM to answer a backward question created by a simple template, i.e., to predict the masked number when a candidate answer is provided. Instead of using forward or backward reasoning alone, we propose **FOBAR** to combine **FO**rward and **BA**ckward **R**easoning for verification. Extensive experiments on six standard mathematical data sets and three LLMs show that FOBAR achieves state-of-the-art performance. In particular, FOBAR outperforms Self-Consistency, which uses forward reasoning alone, demonstrating that combining forward and backward reasoning is more accurate in verification. In addition, FOBAR achieves higher accuracy than existing verification methods, showing the effectiveness of the simple template used in backward reasoning and the proposed combination.",
}
| Self-Consistency samples diverse reasoning chains with answers and chooses the final answer by majority voting. It is based on forward reasoning and cannot further improve performance by sampling more reasoning chains when saturated. To further boost performance, we introduce backward reasoning to verify candidate answers. Specifically, for mathematical tasks, we mask a number in the question and ask the LLM to answer a backward question created by a simple template, i.e., to predict the masked number when a candidate answer is provided. Instead of using forward or backward reasoning alone, we propose **FOBAR** to combine **FO**rward and **BA**ckward **R**easoning for verification. Extensive experiments on six standard mathematical data sets and three LLMs show that FOBAR achieves state-of-the-art performance. In particular, FOBAR outperforms Self-Consistency, which uses forward reasoning alone, demonstrating that combining forward and backward reasoning is more accurate in verification. In addition, FOBAR achieves higher accuracy than existing verification methods, showing the effectiveness of the simple template used in backward reasoning and the proposed combination. | [
"Jiang, Weisen",
"Shi, Han",
"Yu, Longhui",
"Liu, Zhengying",
"Zhang, Yu",
"Li, Zhenguo",
"Kwok, James"
]
| Forward-Backward Reasoning in Large Language Models for Mathematical Verification | findings-acl.397 | Poster | 2405.16802v3 |
https://aclanthology.org/2024.findings-acl.398.bib | @inproceedings{han-etal-2024-towards,
title = "Towards Uncertainty-Aware Language Agent",
author = "Han, Jiuzhou and
Buntine, Wray and
Shareghi, Ehsan",
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.398",
pages = "6662--6685",
abstract = "While Language Agents have achieved promising success by placing Large Language Models at the core of a more versatile design that dynamically interacts with the external world, the existing approaches neglect the notion of uncertainty during these interactions. We present the Uncertainty-Aware Language Agent (UALA), a framework that orchestrates the interaction between the agent and the external world using uncertainty quantification. Compared with other well-known counterparts like ReAct, our extensive experiments across 3 representative tasks (HotpotQA, StrategyQA, MMLU) and various LLM sizes demonstrate that UALA brings a significant improvement of performance, while having a substantially lower reliance on the external world (i.e., reduced number of tool calls and tokens). Our analyses provide various insights including the great potential of UALA compared with agent fine-tuning, and underscore the unreliability of verbalised confidence of LLMs as a proxy for uncertainty.",
}
| While Language Agents have achieved promising success by placing Large Language Models at the core of a more versatile design that dynamically interacts with the external world, the existing approaches neglect the notion of uncertainty during these interactions. We present the Uncertainty-Aware Language Agent (UALA), a framework that orchestrates the interaction between the agent and the external world using uncertainty quantification. Compared with other well-known counterparts like ReAct, our extensive experiments across 3 representative tasks (HotpotQA, StrategyQA, MMLU) and various LLM sizes demonstrate that UALA brings a significant improvement of performance, while having a substantially lower reliance on the external world (i.e., reduced number of tool calls and tokens). Our analyses provide various insights including the great potential of UALA compared with agent fine-tuning, and underscore the unreliability of verbalised confidence of LLMs as a proxy for uncertainty. | [
"Han, Jiuzhou",
"Buntine, Wray",
"Shareghi, Ehsan"
]
| Towards Uncertainty-Aware Language Agent | findings-acl.398 | Poster | 2404.05337v1 |
https://aclanthology.org/2024.findings-acl.399.bib | @inproceedings{lin-etal-2024-detection,
title = "Detection and Positive Reconstruction of Cognitive Distortion Sentences: {M}andarin Dataset and Evaluation",
author = "Lin, Shuya and
Wang, Yuxiong and
Dong, Jonathan and
Ni, Shiguang",
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.399",
pages = "6686--6701",
abstract = "This research introduces a Positive Reconstruction Framework based on positive psychology theory. Overcoming negative thoughts can be challenging, our objective is to address and reframe them through a positive reinterpretation. To tackle this challenge, a two-fold approach is necessary: identifying cognitive distortions and suggesting a positively reframed alternative while preserving the original thought{'}s meaning. Recent studies have investigated the application of Natural Language Processing (NLP) models in English for each stage of this process. In this study, we emphasize the theoretical foundation for the Positive Reconstruction Framework, grounded in broaden-and-build theory. We provide a shared corpus containing 4001 instances for detecting cognitive distortions and 1900 instances for positive reconstruction in Mandarin. Leveraging recent NLP techniques, including transfer learning, fine-tuning pretrained networks, and prompt engineering, we demonstrate the effectiveness of automated tools for both tasks. In summary, our study contributes to multilingual positive reconstruction, highlighting the effectiveness of NLP in cognitive distortion detection and positive reconstruction.",
}
| This research introduces a Positive Reconstruction Framework based on positive psychology theory. Overcoming negative thoughts can be challenging, our objective is to address and reframe them through a positive reinterpretation. To tackle this challenge, a two-fold approach is necessary: identifying cognitive distortions and suggesting a positively reframed alternative while preserving the original thought{'}s meaning. Recent studies have investigated the application of Natural Language Processing (NLP) models in English for each stage of this process. In this study, we emphasize the theoretical foundation for the Positive Reconstruction Framework, grounded in broaden-and-build theory. We provide a shared corpus containing 4001 instances for detecting cognitive distortions and 1900 instances for positive reconstruction in Mandarin. Leveraging recent NLP techniques, including transfer learning, fine-tuning pretrained networks, and prompt engineering, we demonstrate the effectiveness of automated tools for both tasks. In summary, our study contributes to multilingual positive reconstruction, highlighting the effectiveness of NLP in cognitive distortion detection and positive reconstruction. | [
"Lin, Shuya",
"Wang, Yuxiong",
"Dong, Jonathan",
"Ni, Shiguang"
]
| Detection and Positive Reconstruction of Cognitive Distortion Sentences: {M}andarin Dataset and Evaluation | findings-acl.399 | Poster | 2405.15334v1 |
https://aclanthology.org/2024.findings-acl.400.bib | @inproceedings{han-etal-2024-pive,
title = "{P}i{V}e: Prompting with Iterative Verification Improving Graph-based Generative Capability of {LLM}s",
author = "Han, Jiuzhou and
Collier, Nigel and
Buntine, Wray and
Shareghi, Ehsan",
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.400",
pages = "6702--6718",
abstract = "Large language models (LLMs) have shown great abilities of solving various natural language tasks in different domains. Due to the training objective of LLMs and their pre-training data, LLMs are not very well equipped for tasks involving structured data generation. We propose a framework, Prompting with Iterative Verification (PiVe), to improve graph-based generative capability of LLMs. We show how a small language model could be trained to act as a verifier module for the output of an LLM(i.e., ChatGPT, GPT-4), and to iteratively improve its performance via fine-grained corrective instructions. We also show how the verifier module could apply iterative corrections offline for a more cost-effective solution to the text-to-graph generation task. Experiments on three graph-based datasets show consistent improvement gained via PiVe. Additionally, we create GenWiki-HIQ and highlight that the verifier module can be used as a data augmentation tool to help improve the quality of automatically generated parallel text-graph datasets.",
}
| Large language models (LLMs) have shown great abilities of solving various natural language tasks in different domains. Due to the training objective of LLMs and their pre-training data, LLMs are not very well equipped for tasks involving structured data generation. We propose a framework, Prompting with Iterative Verification (PiVe), to improve graph-based generative capability of LLMs. We show how a small language model could be trained to act as a verifier module for the output of an LLM(i.e., ChatGPT, GPT-4), and to iteratively improve its performance via fine-grained corrective instructions. We also show how the verifier module could apply iterative corrections offline for a more cost-effective solution to the text-to-graph generation task. Experiments on three graph-based datasets show consistent improvement gained via PiVe. Additionally, we create GenWiki-HIQ and highlight that the verifier module can be used as a data augmentation tool to help improve the quality of automatically generated parallel text-graph datasets. | [
"Han, Jiuzhou",
"Collier, Nigel",
"Buntine, Wray",
"Shareghi, Ehsan"
]
| {P}i{V}e: Prompting with Iterative Verification Improving Graph-based Generative Capability of {LLM}s | findings-acl.400 | Poster | 2305.12392v3 |
https://aclanthology.org/2024.findings-acl.401.bib | @inproceedings{gao-etal-2024-two,
title = "Two-stage Generative Question Answering on Temporal Knowledge Graph Using Large Language Models",
author = "Gao, Yifu and
Qiao, Linbo and
Kan, Zhigang and
Wen, Zhihua and
He, Yongquan and
Li, Dongsheng",
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.401",
pages = "6719--6734",
abstract = "Temporal knowledge graph question answering (TKGQA) poses a significant challenge task, due to the temporal constraints hidden in questions and the answers sought from dynamic structured knowledge. Although large language models (LLMs) have made considerable progress in their reasoning ability over structured data, their application to the TKGQA task is a relatively unexplored area. This paper first proposes a novel generative temporal knowledge graph question answering framework, GenTKGQA, which guides LLMs to answer temporal questions through two phases: Subgraph Retrieval and Answer Generation. First, we exploit LLM{'}s intrinsic knowledge to mine temporal constraints and structural links in the questions without extra training, thus narrowing down the subgraph search space in both temporal and structural dimensions. Next, we design virtual knowledge indicators to fuse the graph neural network signals of the subgraph and the text representations of the LLM in a non-shallow way, which helps the open-source LLM deeply understand the temporal order and structural dependencies among the retrieved facts through instruction tuning. Experimental results on two widely used datasets demonstrate the superiority of our model.",
}
| Temporal knowledge graph question answering (TKGQA) poses a significant challenge task, due to the temporal constraints hidden in questions and the answers sought from dynamic structured knowledge. Although large language models (LLMs) have made considerable progress in their reasoning ability over structured data, their application to the TKGQA task is a relatively unexplored area. This paper first proposes a novel generative temporal knowledge graph question answering framework, GenTKGQA, which guides LLMs to answer temporal questions through two phases: Subgraph Retrieval and Answer Generation. First, we exploit LLM{'}s intrinsic knowledge to mine temporal constraints and structural links in the questions without extra training, thus narrowing down the subgraph search space in both temporal and structural dimensions. Next, we design virtual knowledge indicators to fuse the graph neural network signals of the subgraph and the text representations of the LLM in a non-shallow way, which helps the open-source LLM deeply understand the temporal order and structural dependencies among the retrieved facts through instruction tuning. Experimental results on two widely used datasets demonstrate the superiority of our model. | [
"Gao, Yifu",
"Qiao, Linbo",
"Kan, Zhigang",
"Wen, Zhihua",
"He, Yongquan",
"Li, Dongsheng"
]
| Two-stage Generative Question Answering on Temporal Knowledge Graph Using Large Language Models | findings-acl.401 | Poster | 2402.16568v2 |
https://aclanthology.org/2024.findings-acl.402.bib | @inproceedings{akter-etal-2024-visreas,
title = "{VISREAS}: Complex Visual Reasoning with Unanswerable Questions",
author = "Akter, Syeda Nahida and
Lee, Sangwu and
Chang, Yingshan and
Bisk, Yonatan and
Nyberg, Eric",
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.402",
pages = "6735--6752",
abstract = "Verifying a question{'}s validity before answering is crucial in real-world applications, where users may provide imperfect instructions. In this scenario, an ideal model should address the discrepancies in the query and convey them to the users rather than generating the best possible answer. Addressing this requirement, we introduce a new compositional visual question-answering dataset, VisReas, that consists of answerable and unanswerable visual queries formulated by traversing and perturbing commonalities and differences among objects, attributes, and relations. VisReas contains 2.07M semantically diverse queries generated automatically using Visual Genome scene graphs. The unique feature of this task, validating question answerability with respect to an image before answering, and the poor performance of state-of-the-art models inspired the design of a new modular baseline, Logic2Vision that reasons by producing and executing pseudocode without any external modules to generate the answer. Logic2Vision outperforms generative models in VisReas (+4.82{\%} over LLaVA-1.5; +12.23{\%} over InstructBLIP) and achieves a significant gain in performance against the classification models.",
}
| Verifying a question{'}s validity before answering is crucial in real-world applications, where users may provide imperfect instructions. In this scenario, an ideal model should address the discrepancies in the query and convey them to the users rather than generating the best possible answer. Addressing this requirement, we introduce a new compositional visual question-answering dataset, VisReas, that consists of answerable and unanswerable visual queries formulated by traversing and perturbing commonalities and differences among objects, attributes, and relations. VisReas contains 2.07M semantically diverse queries generated automatically using Visual Genome scene graphs. The unique feature of this task, validating question answerability with respect to an image before answering, and the poor performance of state-of-the-art models inspired the design of a new modular baseline, Logic2Vision that reasons by producing and executing pseudocode without any external modules to generate the answer. Logic2Vision outperforms generative models in VisReas (+4.82{\%} over LLaVA-1.5; +12.23{\%} over InstructBLIP) and achieves a significant gain in performance against the classification models. | [
"Akter, Syeda Nahida",
"Lee, Sangwu",
"Chang, Yingshan",
"Bisk, Yonatan",
"Nyberg, Eric"
]
| {VISREAS}: Complex Visual Reasoning with Unanswerable Questions | findings-acl.402 | Poster | 2212.10189v2 |
https://aclanthology.org/2024.findings-acl.403.bib | @inproceedings{hu-etal-2024-unified,
title = "A Unified Generative Framework for Bilingual Euphemism Detection and Identification",
author = "Hu, Yuxue and
Li, Junsong and
Wang, Tongguan and
Su, Dongyu and
Su, Guixin and
Sha, Ying",
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.403",
pages = "6753--6766",
abstract = "Various euphemisms are emerging in social networks, attracting widespread attention from the natural language processing community. However, existing euphemism datasets are only domain-specific or language-specific. In addition, existing approaches to the study of euphemisms are one-sided. Either only the euphemism detection task or only the euphemism identification task is accomplished, lacking a unified framework. To this end, we construct a large-scale Bilingual Multi-category dataset of Euphemisms named BME, which covers a total of 12 categories for two languages, English and Chinese. Then, we first propose a unified generative model to Jointly conduct the tasks of bilingual Euphemism Detection and Identification named JointEDI. By comparing with LLMs and human evaluation, we demonstrate the effectiveness of the proposed JointEDI and the feasibility of unifying euphemism detection and euphemism identification tasks. Moreover, the BME dataset also provides a new reference standard for euphemism detection and euphemism identification.",
}
| Various euphemisms are emerging in social networks, attracting widespread attention from the natural language processing community. However, existing euphemism datasets are only domain-specific or language-specific. In addition, existing approaches to the study of euphemisms are one-sided. Either only the euphemism detection task or only the euphemism identification task is accomplished, lacking a unified framework. To this end, we construct a large-scale Bilingual Multi-category dataset of Euphemisms named BME, which covers a total of 12 categories for two languages, English and Chinese. Then, we first propose a unified generative model to Jointly conduct the tasks of bilingual Euphemism Detection and Identification named JointEDI. By comparing with LLMs and human evaluation, we demonstrate the effectiveness of the proposed JointEDI and the feasibility of unifying euphemism detection and euphemism identification tasks. Moreover, the BME dataset also provides a new reference standard for euphemism detection and euphemism identification. | [
"Hu, Yuxue",
"Li, Junsong",
"Wang, Tongguan",
"Su, Dongyu",
"Su, Guixin",
"Sha, Ying"
]
| A Unified Generative Framework for Bilingual Euphemism Detection and Identification | findings-acl.403 | Poster | 2103.16808v1 |
https://aclanthology.org/2024.findings-acl.404.bib | @inproceedings{cong-etal-2024-styledubber,
title = "{S}tyle{D}ubber: Towards Multi-Scale Style Learning for Movie Dubbing",
author = "Cong, Gaoxiang and
Qi, Yuankai and
Li, Liang and
Beheshti, Amin and
Zhang, Zhedong and
Hengel, Anton and
Yang, Ming-Hsuan and
Yan, Chenggang and
Huang, Qingming",
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.404",
pages = "6767--6779",
abstract = "Given a script, the challenge in Movie Dubbing (Visual Voice Cloning, V2C) is to generate speech that aligns well with the video in both time and emotion, based on the tone of a reference audio track. Existing state-of-the-art V2C models break the phonemes in the script according to the divisions between video frames, which solves the temporal alignment problem but leads to incomplete phoneme pronunciation and poor identity stability. To address this problem, we propose StyleDubber, which switches dubbing learning from the frame level to phoneme level. It contains three main components: (1) A multimodal style adaptor operating at the phoneme level to learn pronunciation style from the reference audio, and generate intermediate representations informed by the facial emotion presented in the video; (2) An utterance-level style learning module, which guides both the mel-spectrogram decoding and the refining processes from the intermediate embeddings to improve the overall style expression; And (3) a phoneme-guided lip aligner to maintain lip sync. Extensive experiments on two of the primary benchmarks, V2C and Grid, demonstrate the favorable performance of the proposed method as compared to the current state-of-the-art. The code will be made available at https://github.com/GalaxyCong/StyleDubber.",
}
| Given a script, the challenge in Movie Dubbing (Visual Voice Cloning, V2C) is to generate speech that aligns well with the video in both time and emotion, based on the tone of a reference audio track. Existing state-of-the-art V2C models break the phonemes in the script according to the divisions between video frames, which solves the temporal alignment problem but leads to incomplete phoneme pronunciation and poor identity stability. To address this problem, we propose StyleDubber, which switches dubbing learning from the frame level to phoneme level. It contains three main components: (1) A multimodal style adaptor operating at the phoneme level to learn pronunciation style from the reference audio, and generate intermediate representations informed by the facial emotion presented in the video; (2) An utterance-level style learning module, which guides both the mel-spectrogram decoding and the refining processes from the intermediate embeddings to improve the overall style expression; And (3) a phoneme-guided lip aligner to maintain lip sync. Extensive experiments on two of the primary benchmarks, V2C and Grid, demonstrate the favorable performance of the proposed method as compared to the current state-of-the-art. The code will be made available at https://github.com/GalaxyCong/StyleDubber. | [
"Cong, Gaoxiang",
"Qi, Yuankai",
"Li, Liang",
"Beheshti, Amin",
"Zhang, Zhedong",
"Hengel, Anton",
"Yang, Ming-Hsuan",
"Yan, Chenggang",
"Huang, Qingming"
]
| {S}tyle{D}ubber: Towards Multi-Scale Style Learning for Movie Dubbing | findings-acl.404 | Poster | 2402.12636v3 |
https://aclanthology.org/2024.findings-acl.405.bib | @inproceedings{yang-liu-2024-etas,
title = "{ETAS}: Zero-Shot Transformer Architecture Search via Network Trainability and Expressivity",
author = "Yang, Jiechao and
Liu, Yong",
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.405",
pages = "6780--6795",
abstract = "Transformer Architecture Search (TAS) methods aim to automate searching for the optimal Transformer architecture configurations for a given task. However, they are impeded by the prohibitive cost of evaluating Transformer architectures. Recently, several Zero-Shot TAS methods have been proposed to mitigate this problem by utilizing zero-cost proxies to evaluate Transformer architectures without training. Unfortunately, they are limited to specific computer vision or natural language processing tasks. Nonetheless, most of them are developed based on empirical observations and lack theoretical guarantees. To solve this problem, we develop a new zero-cost proxy called NTSR that combines two theoretically-inspired indicators to measure the trainability and expressivity of Transformer networks separately. We then integrate it into an effective regularized evolution framework called ETAS to demonstrate its efficacy on various tasks. The results show that our proposed NTSR proxy can consistently achieve a higher correlation with the true performance of Transformer networks on both computer vision and natural language processing tasks. Further, it can significantly accelerate the search process for finding the best-performing Transformer architecture configurations.",
}
| Transformer Architecture Search (TAS) methods aim to automate searching for the optimal Transformer architecture configurations for a given task. However, they are impeded by the prohibitive cost of evaluating Transformer architectures. Recently, several Zero-Shot TAS methods have been proposed to mitigate this problem by utilizing zero-cost proxies to evaluate Transformer architectures without training. Unfortunately, they are limited to specific computer vision or natural language processing tasks. Nonetheless, most of them are developed based on empirical observations and lack theoretical guarantees. To solve this problem, we develop a new zero-cost proxy called NTSR that combines two theoretically-inspired indicators to measure the trainability and expressivity of Transformer networks separately. We then integrate it into an effective regularized evolution framework called ETAS to demonstrate its efficacy on various tasks. The results show that our proposed NTSR proxy can consistently achieve a higher correlation with the true performance of Transformer networks on both computer vision and natural language processing tasks. Further, it can significantly accelerate the search process for finding the best-performing Transformer architecture configurations. | [
"Yang, Jiechao",
"Liu, Yong"
]
| {ETAS}: Zero-Shot Transformer Architecture Search via Network Trainability and Expressivity | findings-acl.405 | Poster | 1809.02209v2 |
https://aclanthology.org/2024.findings-acl.406.bib | @inproceedings{xu-etal-2024-reasoning,
title = "Reasoning Like a Doctor: Improving Medical Dialogue Systems via Diagnostic Reasoning Process Alignment",
author = "Xu, Kaishuai and
Cheng, Yi and
Hou, Wenjun and
Tan, Qiaoyu and
Li, Wenjie",
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.406",
pages = "6796--6814",
abstract = "Medical dialogue systems have attracted significant attention for their potential to act as medical assistants. Enabling these medical systems to emulate clinicians{'} diagnostic reasoning process has been the long-standing research focus. Previous studies rudimentarily realized the simulation of clinicians{'} diagnostic process by fine-tuning language models on high-quality dialogue datasets. Nonetheless, they overly focus on the outcomes of the clinician{'}s reasoning process while ignoring their internal thought processes and alignment with clinician preferences. Our work aims to build a medical dialogue system that aligns with clinicians{'} diagnostic reasoning processes. We propose a novel framework, Emulation, designed to generate an appropriate response that relies on abductive and deductive diagnostic reasoning analyses and aligns with clinician preferences through thought process modeling. Experimental results on two datasets confirm the efficacy of Emulation. Crucially, our framework furnishes clear explanations for the generated responses, enhancing its transparency in medical consultations.",
}
| Medical dialogue systems have attracted significant attention for their potential to act as medical assistants. Enabling these medical systems to emulate clinicians{'} diagnostic reasoning process has been the long-standing research focus. Previous studies rudimentarily realized the simulation of clinicians{'} diagnostic process by fine-tuning language models on high-quality dialogue datasets. Nonetheless, they overly focus on the outcomes of the clinician{'}s reasoning process while ignoring their internal thought processes and alignment with clinician preferences. Our work aims to build a medical dialogue system that aligns with clinicians{'} diagnostic reasoning processes. We propose a novel framework, Emulation, designed to generate an appropriate response that relies on abductive and deductive diagnostic reasoning analyses and aligns with clinician preferences through thought process modeling. Experimental results on two datasets confirm the efficacy of Emulation. Crucially, our framework furnishes clear explanations for the generated responses, enhancing its transparency in medical consultations. | [
"Xu, Kaishuai",
"Cheng, Yi",
"Hou, Wenjun",
"Tan, Qiaoyu",
"Li, Wenjie"
]
| Reasoning Like a Doctor: Improving Medical Dialogue Systems via Diagnostic Reasoning Process Alignment | findings-acl.406 | Poster | 2406.13934v1 |
https://aclanthology.org/2024.findings-acl.407.bib | @inproceedings{wu-etal-2024-conceptmath,
title = "{C}oncept{M}ath: A Bilingual Concept-wise Benchmark for Measuring Mathematical Reasoning of Large Language Models",
author = "Wu, Yanan and
Liu, Jie and
Bu, Xingyuan and
Liu, Jiaheng and
Zhou, Zhanhui and
Zhang, Yuanxing and
Zhang, Chenchen and
ZhiqiBai, ZhiqiBai and
Chen, Haibin and
Ge, Tiezheng and
Ouyang, Wanli and
Su, Wenbo and
Zheng, Bo",
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.407",
pages = "6815--6839",
abstract = "This paper introduces ConceptMath, a bilingual (English and Chinese), fine-grained benchmark that evaluates concept-wise mathematical reasoning of Large Language Models (LLMs). Unlike traditional benchmarks that evaluate general mathematical reasoning with an average accuracy, ConceptMath systemically organizes math problems under a hierarchy of math concepts, so that mathematical reasoning can be evaluated at different granularity with concept-wise accuracies. Based on our ConcepthMath, we then evaluate a broad range of LLMs, and we observe existing LLMs, though achieving high average accuracies on traditional benchmarks, exhibit significant performance variations across different math concepts and may even fail catastrophically on the most basic ones. Besides, we also introduce an efficient fine-tuning strategy to enhance the weaknesses of existing LLMs. Finally, we hope ConceptMath could guide the developers to understand the fine-grained mathematical abilities of their models and facilitate the growth of foundation models. Code is available at https://github.com/conceptmath/conceptmath.",
}
| This paper introduces ConceptMath, a bilingual (English and Chinese), fine-grained benchmark that evaluates concept-wise mathematical reasoning of Large Language Models (LLMs). Unlike traditional benchmarks that evaluate general mathematical reasoning with an average accuracy, ConceptMath systemically organizes math problems under a hierarchy of math concepts, so that mathematical reasoning can be evaluated at different granularity with concept-wise accuracies. Based on our ConcepthMath, we then evaluate a broad range of LLMs, and we observe existing LLMs, though achieving high average accuracies on traditional benchmarks, exhibit significant performance variations across different math concepts and may even fail catastrophically on the most basic ones. Besides, we also introduce an efficient fine-tuning strategy to enhance the weaknesses of existing LLMs. Finally, we hope ConceptMath could guide the developers to understand the fine-grained mathematical abilities of their models and facilitate the growth of foundation models. Code is available at https://github.com/conceptmath/conceptmath. | [
"Wu, Yanan",
"Liu, Jie",
"Bu, Xingyuan",
"Liu, Jiaheng",
"Zhou, Zhanhui",
"Zhang, Yuanxing",
"Zhang, Chenchen",
"ZhiqiBai, ZhiqiBai",
"Chen, Haibin",
"Ge, Tiezheng",
"Ouyang, Wanli",
"Su, Wenbo",
"Zheng, Bo"
]
| {C}oncept{M}ath: A Bilingual Concept-wise Benchmark for Measuring Mathematical Reasoning of Large Language Models | findings-acl.407 | Poster | 2402.14660v2 |
https://aclanthology.org/2024.findings-acl.408.bib | @inproceedings{chen-etal-2024-reinstruct,
title = "{REI}nstruct: Building Instruction Data from Unlabeled Corpus",
author = "Chen, Shu and
Guan, Xinyan and
Lu, Yaojie and
Lin, Hongyu and
Han, Xianpei and
Sun, Le",
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.408",
pages = "6840--6856",
abstract = "Manually annotating instruction data for large language models is difficult, costly, and hard to scale. Meanwhile, current automatic annotation methods typically rely on distilling synthetic data from proprietary LLMs, which not only limits the upper bound of the quality of the instruction data but also raises potential copyright issues. In this paper, we propose REInstruct, a simple and scalable method to automatically build instruction data from an unlabeled corpus without heavy reliance on proprietary LLMs and human annotation.Specifically, REInstruct first selects a subset of unlabeled texts that potentially contain well-structured helpful and insightful content and then generates instructions for these texts. To generate accurate and relevant responses for effective and robust training, REInstruct further proposes a rewriting-based approach to improve the quality of the generated instruction data. By training Llama-7b on a combination of 3k seed data and 32k synthetic data from REInstruct, fine-tuned model achieves a 65.41{\%} win rate on AlpacaEval leaderboard against text-davinci-003, outperforming other open-source, non-distilled instruction data construction methods. The code is publicly available at \url{https://github.com/cs32963/REInstruct}.",
}
| Manually annotating instruction data for large language models is difficult, costly, and hard to scale. Meanwhile, current automatic annotation methods typically rely on distilling synthetic data from proprietary LLMs, which not only limits the upper bound of the quality of the instruction data but also raises potential copyright issues. In this paper, we propose REInstruct, a simple and scalable method to automatically build instruction data from an unlabeled corpus without heavy reliance on proprietary LLMs and human annotation.Specifically, REInstruct first selects a subset of unlabeled texts that potentially contain well-structured helpful and insightful content and then generates instructions for these texts. To generate accurate and relevant responses for effective and robust training, REInstruct further proposes a rewriting-based approach to improve the quality of the generated instruction data. By training Llama-7b on a combination of 3k seed data and 32k synthetic data from REInstruct, fine-tuned model achieves a 65.41{\%} win rate on AlpacaEval leaderboard against text-davinci-003, outperforming other open-source, non-distilled instruction data construction methods. The code is publicly available at \url{https://github.com/cs32963/REInstruct}. | [
"Chen, Shu",
"Guan, Xinyan",
"Lu, Yaojie",
"Lin, Hongyu",
"Han, Xianpei",
"Sun, Le"
]
| {REI}nstruct: Building Instruction Data from Unlabeled Corpus | findings-acl.408 | Poster | 2210.09175v1 |
https://aclanthology.org/2024.findings-acl.409.bib | @inproceedings{chen-etal-2024-learning-maximize,
title = "Learning to Maximize Mutual Information for Chain-of-Thought Distillation",
author = "Chen, Xin and
Huang, Hanxian and
Gao, Yanjun and
Wang, Yi and
Zhao, Jishen and
Ding, Ke",
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.409",
pages = "6857--6868",
abstract = "Knowledge distillation, the technique of transferring knowledge from large, complex models to smaller ones, marks a pivotal step towards efficient AI deployment. Distilling Step-by-Step (DSS), a novel method utilizing chain-of-thought (CoT) distillation, has demonstrated promise by imbuing smaller models with the superior reasoning capabilities of their larger counterparts. In DSS, the distilled model acquires the ability to generate rationales and predict labels concurrently through a multi-task learning framework. However, DSS overlooks the intrinsic relationship between the two training tasks, leading to ineffective integration of CoT knowledge with the task of label prediction. To this end, we investigate the mutual relationship of the two tasks from Information Bottleneck perspective and formulate it as maximizing the mutual information of the representation features of the two tasks. We propose a variational approach to solve this optimization problem using a learning-based method. Our experimental results across four datasets demonstrate that our method outperforms the state-of-the-art DSS. Our findings offer insightful guidance for future research on language model distillation as well as applications involving CoT. Codes are available at https://github.com/xinchen9/cot{\_}distillation{\_}ACL2024.",
}
| Knowledge distillation, the technique of transferring knowledge from large, complex models to smaller ones, marks a pivotal step towards efficient AI deployment. Distilling Step-by-Step (DSS), a novel method utilizing chain-of-thought (CoT) distillation, has demonstrated promise by imbuing smaller models with the superior reasoning capabilities of their larger counterparts. In DSS, the distilled model acquires the ability to generate rationales and predict labels concurrently through a multi-task learning framework. However, DSS overlooks the intrinsic relationship between the two training tasks, leading to ineffective integration of CoT knowledge with the task of label prediction. To this end, we investigate the mutual relationship of the two tasks from Information Bottleneck perspective and formulate it as maximizing the mutual information of the representation features of the two tasks. We propose a variational approach to solve this optimization problem using a learning-based method. Our experimental results across four datasets demonstrate that our method outperforms the state-of-the-art DSS. Our findings offer insightful guidance for future research on language model distillation as well as applications involving CoT. Codes are available at https://github.com/xinchen9/cot{\_}distillation{\_}ACL2024. | [
"Chen, Xin",
"Huang, Hanxian",
"Gao, Yanjun",
"Wang, Yi",
"Zhao, Jishen",
"Ding, Ke"
]
| Learning to Maximize Mutual Information for Chain-of-Thought Distillation | findings-acl.409 | Poster | 2403.03348v3 |
https://aclanthology.org/2024.findings-acl.410.bib | @inproceedings{lin-etal-2024-pemt,
title = "{PEMT}: Multi-Task Correlation Guided Mixture-of-Experts Enables Parameter-Efficient Transfer Learning",
author = "Lin, Zhisheng and
Fu, Han and
Liu, Chenghao and
Li, Zhuo and
Sun, Jianling",
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.410",
pages = "6869--6883",
abstract = "Parameter-efficient fine-tuning (PEFT) has emerged as an effective method for adapting pre-trained language models to various tasks efficiently. Recently, there has been a growing interest in transferring knowledge from one or multiple tasks to the downstream target task to achieve performance improvements. However, current approaches typically either train adapters on individual tasks or distill shared knowledge from source tasks, failing to fully exploit task-specific knowledge and the correlation between source and target tasks. To overcome these limitations, we propose PEMT, a novel parameter-efficient fine-tuning framework based on multi-task transfer learning. PEMT extends the mixture-of-experts (MoE) framework to capture the transferable knowledge as a weighted combination of adapters trained on source tasks. These weights are determined by a gated unit, measuring the correlation between the target and each source task using task description prompt vectors. To fully exploit the task-specific knowledge, we also propose the Task Sparsity Loss to improve the sparsity of the gated unit. We conduct experiments on a broad range of tasks over 17 datasets. The experimental results demonstrate our PEMT yields stable improvements over full fine-tuning, and state-of-the-art PEFT and knowledge transferring methods on various tasks. The results highlight the effectiveness of our method which is capable of sufficiently exploiting the knowledge and correlation features across multiple tasks.",
}
| Parameter-efficient fine-tuning (PEFT) has emerged as an effective method for adapting pre-trained language models to various tasks efficiently. Recently, there has been a growing interest in transferring knowledge from one or multiple tasks to the downstream target task to achieve performance improvements. However, current approaches typically either train adapters on individual tasks or distill shared knowledge from source tasks, failing to fully exploit task-specific knowledge and the correlation between source and target tasks. To overcome these limitations, we propose PEMT, a novel parameter-efficient fine-tuning framework based on multi-task transfer learning. PEMT extends the mixture-of-experts (MoE) framework to capture the transferable knowledge as a weighted combination of adapters trained on source tasks. These weights are determined by a gated unit, measuring the correlation between the target and each source task using task description prompt vectors. To fully exploit the task-specific knowledge, we also propose the Task Sparsity Loss to improve the sparsity of the gated unit. We conduct experiments on a broad range of tasks over 17 datasets. The experimental results demonstrate our PEMT yields stable improvements over full fine-tuning, and state-of-the-art PEFT and knowledge transferring methods on various tasks. The results highlight the effectiveness of our method which is capable of sufficiently exploiting the knowledge and correlation features across multiple tasks. | [
"Lin, Zhisheng",
"Fu, Han",
"Liu, Chenghao",
"Li, Zhuo",
"Sun, Jianling"
]
| {PEMT}: Multi-Task Correlation Guided Mixture-of-Experts Enables Parameter-Efficient Transfer Learning | findings-acl.410 | Poster | 2303.16154v1 |
https://aclanthology.org/2024.findings-acl.411.bib | @inproceedings{liu-etal-2024-mathbench,
title = "{M}ath{B}ench: Evaluating the Theory and Application Proficiency of {LLM}s with a Hierarchical Mathematics Benchmark",
author = "Liu, Hongwei and
Zheng, Zilong and
Qiao, Yuxuan and
Duan, Haodong and
Fei, Zhiwei and
Zhou, Fengzhe and
Zhang, Wenwei and
Zhang, Songyang and
Lin, Dahua and
Chen, Kai",
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.411",
pages = "6884--6915",
abstract = "Recent advancements in large language models (LLMs) have showcased significant improvements in mathematics. However, traditional math benchmarks like GSM8k offer a unidimensional perspective, which fall short in providing a holistic assessment of the LLMs{'} math capabilities. To address this gap, we introduce MathBench, a new benchmark that rigorously assesses the mathematical capabilities of large language models. MathBench spans a wide range of mathematical disciplines, offering a detailed evaluation of both theoretical understanding and practical problem-solving skills. The benchmark progresses through five distinct stages, from basic arithmetic to college mathematics, and is structured to evaluate models at various depths of knowledge. Each stage includes theoretical questions and application problems, allowing us to measure a model{'}s mathematical proficiency and its ability to apply concepts in practical scenarios. MathBench aims to enhance the evaluation of LLMs{'} mathematical abilities, providing a nuanced view of their knowledge understanding levels and problem solving skills in a bilingual context.",
}
| Recent advancements in large language models (LLMs) have showcased significant improvements in mathematics. However, traditional math benchmarks like GSM8k offer a unidimensional perspective, which fall short in providing a holistic assessment of the LLMs{'} math capabilities. To address this gap, we introduce MathBench, a new benchmark that rigorously assesses the mathematical capabilities of large language models. MathBench spans a wide range of mathematical disciplines, offering a detailed evaluation of both theoretical understanding and practical problem-solving skills. The benchmark progresses through five distinct stages, from basic arithmetic to college mathematics, and is structured to evaluate models at various depths of knowledge. Each stage includes theoretical questions and application problems, allowing us to measure a model{'}s mathematical proficiency and its ability to apply concepts in practical scenarios. MathBench aims to enhance the evaluation of LLMs{'} mathematical abilities, providing a nuanced view of their knowledge understanding levels and problem solving skills in a bilingual context. | [
"Liu, Hongwei",
"Zheng, Zilong",
"Qiao, Yuxuan",
"Duan, Haodong",
"Fei, Zhiwei",
"Zhou, Fengzhe",
"Zhang, Wenwei",
"Zhang, Songyang",
"Lin, Dahua",
"Chen, Kai"
]
| {M}ath{B}ench: Evaluating the Theory and Application Proficiency of {LLM}s with a Hierarchical Mathematics Benchmark | findings-acl.411 | Poster | 2405.12209v1 |
https://aclanthology.org/2024.findings-acl.412.bib | @inproceedings{ren-etal-2024-identifying,
title = "Identifying Semantic Induction Heads to Understand In-Context Learning",
author = "Ren, Jie and
Guo, Qipeng and
Yan, Hang and
Liu, Dongrui and
Zhang, Quanshi and
Qiu, Xipeng and
Lin, Dahua",
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.412",
pages = "6916--6932",
abstract = "Although large language models (LLMs) have demonstrated remarkable performance, the lack of transparency in their inference logic raises concerns about their trustworthiness. To gain a better understanding of LLMs, we conduct a detailed analysis of the operations of attention heads and aim to better understand the in-context learning of LLMs. Specifically, we investigate whether attention heads encode two types of relationships between tokens present in natural languages: the syntactic dependency parsed from sentences and the relation within knowledge graphs. We find that certain attention heads exhibit a pattern where, when attending to subject tokens, they recall object tokens and increase the output logits of those object tokens. More crucially, the formulation of such semantic induction heads has a close correlation with the emergence of the in-context learning ability of language models. The study of semantic attention heads advances our understanding of the intricate operations of attention heads in transformers, and further provides new insights into the in-context learning of LLMs.",
}
| Although large language models (LLMs) have demonstrated remarkable performance, the lack of transparency in their inference logic raises concerns about their trustworthiness. To gain a better understanding of LLMs, we conduct a detailed analysis of the operations of attention heads and aim to better understand the in-context learning of LLMs. Specifically, we investigate whether attention heads encode two types of relationships between tokens present in natural languages: the syntactic dependency parsed from sentences and the relation within knowledge graphs. We find that certain attention heads exhibit a pattern where, when attending to subject tokens, they recall object tokens and increase the output logits of those object tokens. More crucially, the formulation of such semantic induction heads has a close correlation with the emergence of the in-context learning ability of language models. The study of semantic attention heads advances our understanding of the intricate operations of attention heads in transformers, and further provides new insights into the in-context learning of LLMs. | [
"Ren, Jie",
"Guo, Qipeng",
"Yan, Hang",
"Liu, Dongrui",
"Zhang, Quanshi",
"Qiu, Xipeng",
"Lin, Dahua"
]
| Identifying Semantic Induction Heads to Understand In-Context Learning | findings-acl.412 | Poster | 2402.13055v2 |
https://aclanthology.org/2024.findings-acl.413.bib | @inproceedings{jiang-etal-2024-chinese,
title = "{C}hinese Spelling Corrector Is Just a Language Learner",
author = "Jiang, Lai and
Wu, Hongqiu and
Zhao, Hai 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.413",
pages = "6933--6943",
abstract = "This paper emphasizes Chinese spelling correction by means of self-supervised learning, which means there are no annotated errors within the training data. Our intuition is that humans are naturally good correctors with exposure to error-free sentences, which contrasts with current unsupervised methods that strongly rely on the usage of confusion sets to produce parallel sentences. In this paper, we demonstrate that learning a spelling correction model is identical to learning a language model from error-free data alone, with decoding it in a greater search space. We propose \textit{Denoising Decoding Correction (D2C)}, which selectively imposes noise upon the source sentence to determine the underlying correct characters. Our method is largely inspired by the ability of language models to perform correction, including both BERT-based models and large language models (LLMs). We show that the self-supervised learning manner generally outperforms the confusion set in specific domains because it bypasses the need to introduce error characters to the training data which can impair the error patterns not included in the introduced error characters.",
}
| This paper emphasizes Chinese spelling correction by means of self-supervised learning, which means there are no annotated errors within the training data. Our intuition is that humans are naturally good correctors with exposure to error-free sentences, which contrasts with current unsupervised methods that strongly rely on the usage of confusion sets to produce parallel sentences. In this paper, we demonstrate that learning a spelling correction model is identical to learning a language model from error-free data alone, with decoding it in a greater search space. We propose \textit{Denoising Decoding Correction (D2C)}, which selectively imposes noise upon the source sentence to determine the underlying correct characters. Our method is largely inspired by the ability of language models to perform correction, including both BERT-based models and large language models (LLMs). We show that the self-supervised learning manner generally outperforms the confusion set in specific domains because it bypasses the need to introduce error characters to the training data which can impair the error patterns not included in the introduced error characters. | [
"Jiang, Lai",
"Wu, Hongqiu",
"Zhao, Hai",
"Zhang, Min"
]
| {C}hinese Spelling Corrector Is Just a Language Learner | findings-acl.413 | Poster | 2004.14166v2 |
https://aclanthology.org/2024.findings-acl.414.bib | @inproceedings{wu-etal-2024-logical,
title = "Logical Closed Loop: Uncovering Object Hallucinations in Large Vision-Language Models",
author = "Wu, Junfei and
Liu, Qiang and
Wang, Ding and
Zhang, Jinghao and
Wu, Shu and
Wang, Liang and
Tan, Tieniu",
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.414",
pages = "6944--6962",
abstract = "Object hallucination has been an Achilles{'} heel which hinders the broader applications of large vision-language models (LVLMs). Object hallucination refers to the phenomenon that the LVLMs claim non-existent objects in the image. To mitigate the object hallucinations, instruction tuning and external model-based detection methods have been proposed, which either require large-scare computational resources or depend on the detection result of external models. However, there remains an under-explored field to utilize the LVLM itself to alleviate object hallucinations. In this work, we adopt the intuition that the LVLM tends to respond logically consistently for existent objects but inconsistently for hallucinated objects. Therefore, we propose a Logical Closed Loop-based framework for Object Hallucination Detection and Mitigation, namely $\textbf{LogicCheckGPT}$. In specific, we devise logical consistency probing to raise questions with logical correlations, inquiring about attributes from objects and vice versa. Whether their responses can form a logical closed loop serves as an indicator of object hallucination. As a plug-and-play method, it can be seamlessly applied to all existing LVLMs. Comprehensive experiments conducted on three benchmarks across four LVLMs have demonstrated significant improvements brought by our method, indicating its effectiveness and generality.",
}
| Object hallucination has been an Achilles{'} heel which hinders the broader applications of large vision-language models (LVLMs). Object hallucination refers to the phenomenon that the LVLMs claim non-existent objects in the image. To mitigate the object hallucinations, instruction tuning and external model-based detection methods have been proposed, which either require large-scare computational resources or depend on the detection result of external models. However, there remains an under-explored field to utilize the LVLM itself to alleviate object hallucinations. In this work, we adopt the intuition that the LVLM tends to respond logically consistently for existent objects but inconsistently for hallucinated objects. Therefore, we propose a Logical Closed Loop-based framework for Object Hallucination Detection and Mitigation, namely $\textbf{LogicCheckGPT}$. In specific, we devise logical consistency probing to raise questions with logical correlations, inquiring about attributes from objects and vice versa. Whether their responses can form a logical closed loop serves as an indicator of object hallucination. As a plug-and-play method, it can be seamlessly applied to all existing LVLMs. Comprehensive experiments conducted on three benchmarks across four LVLMs have demonstrated significant improvements brought by our method, indicating its effectiveness and generality. | [
"Wu, Junfei",
"Liu, Qiang",
"Wang, Ding",
"Zhang, Jinghao",
"Wu, Shu",
"Wang, Liang",
"Tan, Tieniu"
]
| Logical Closed Loop: Uncovering Object Hallucinations in Large Vision-Language Models | findings-acl.414 | Poster | 2402.11622v2 |
https://aclanthology.org/2024.findings-acl.415.bib | @inproceedings{zhang-etal-2024-retrievalqa,
title = "{R}etrieval{QA}: Assessing Adaptive Retrieval-Augmented Generation for Short-form Open-Domain Question Answering",
author = "Zhang, Zihan and
Fang, Meng and
Chen, Ling",
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.415",
pages = "6963--6975",
abstract = "Adaptive retrieval-augmented generation (ARAG) aims to dynamically determine the necessity of retrieval for queries instead of retrieving indiscriminately to enhance the efficiency and relevance of the sourced information. However, previous works largely overlook the evaluation of ARAG approaches, leading to their effectiveness being understudied. This work presents a benchmark, RetrievalQA, comprising 1,271 short-form questions covering new world and long-tail knowledge. The knowledge necessary to answer the questions is absent from LLMs; therefore, external information must be retrieved to answer correctly. This makes RetrievalQA a suitable testbed to evaluate existing ARAG methods. We observe that calibration-based methods heavily rely on threshold tuning, while vanilla prompting is inadequate for guiding LLMs to make reliable retrieval decisions. Based on our findings, we propose Time-Aware Adaptive Retrieval (TA-ARE), a simple yet effective method that helps LLMs assess the necessity of retrieval without calibration or additional training.",
}
| Adaptive retrieval-augmented generation (ARAG) aims to dynamically determine the necessity of retrieval for queries instead of retrieving indiscriminately to enhance the efficiency and relevance of the sourced information. However, previous works largely overlook the evaluation of ARAG approaches, leading to their effectiveness being understudied. This work presents a benchmark, RetrievalQA, comprising 1,271 short-form questions covering new world and long-tail knowledge. The knowledge necessary to answer the questions is absent from LLMs; therefore, external information must be retrieved to answer correctly. This makes RetrievalQA a suitable testbed to evaluate existing ARAG methods. We observe that calibration-based methods heavily rely on threshold tuning, while vanilla prompting is inadequate for guiding LLMs to make reliable retrieval decisions. Based on our findings, we propose Time-Aware Adaptive Retrieval (TA-ARE), a simple yet effective method that helps LLMs assess the necessity of retrieval without calibration or additional training. | [
"Zhang, Zihan",
"Fang, Meng",
"Chen, Ling"
]
| {R}etrieval{QA}: Assessing Adaptive Retrieval-Augmented Generation for Short-form Open-Domain Question Answering | findings-acl.415 | Poster | 2209.11396v1 |
https://aclanthology.org/2024.findings-acl.416.bib | @inproceedings{chen-etal-2024-llast,
title = "{LL}a{ST}: Improved End-to-end Speech Translation System Leveraged by Large Language Models",
author = "Chen, Xi and
Zhang, Songyang and
Bai, Qibing and
Chen, Kai and
Nakamura, Satoshi",
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.416",
pages = "6976--6987",
abstract = "We introduces ***LLaST***, a framework for building high-performance Large Language model based Speech-to-text Translation systems. We address the limitations of end-to-end speech translation (E2E ST) models by exploring model architecture design and optimization techniques tailored for LLMs. Our approach includes LLM-based speech translation architecture design, ASR-augmented training, multilingual data augmentation, and dual-LoRA optimization. Our approach demonstrates superior performance on the CoVoST-2 benchmark and showcases exceptional scaling capabilities powered by LLMs.We believe this effective method will serve as a strong baseline for speech translation and provide insights for futureimprovements of the LLM-based speech translation framework.",
}
| We introduces ***LLaST***, a framework for building high-performance Large Language model based Speech-to-text Translation systems. We address the limitations of end-to-end speech translation (E2E ST) models by exploring model architecture design and optimization techniques tailored for LLMs. Our approach includes LLM-based speech translation architecture design, ASR-augmented training, multilingual data augmentation, and dual-LoRA optimization. Our approach demonstrates superior performance on the CoVoST-2 benchmark and showcases exceptional scaling capabilities powered by LLMs.We believe this effective method will serve as a strong baseline for speech translation and provide insights for futureimprovements of the LLM-based speech translation framework. | [
"Chen, Xi",
"Zhang, Songyang",
"Bai, Qibing",
"Chen, Kai",
"Nakamura, Satoshi"
]
| {LL}a{ST}: Improved End-to-end Speech Translation System Leveraged by Large Language Models | findings-acl.416 | Poster | 2107.06959v2 |
https://aclanthology.org/2024.findings-acl.417.bib | @inproceedings{gu-yang-2024-plan,
title = "Plan, Generate and Complicate: Improving Low-resource Dialogue State Tracking via Easy-to-Difficult Zero-shot Data Augmentation",
author = "Gu, Ming and
Yang, Yan",
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.417",
pages = "6988--7005",
abstract = "Data augmentation methods have been a promising direction to improve the performance of small models for low-resource dialogue state tracking. However, traditional methods rely on pre-defined user goals and neglect the importance of data complexity in this task. In this paper, we propose EDZ-DA, an Easy-to-Difficult Zero-shot Data Augmentation framework for low-resource dialogue state tracking that utilizes large language models to automatically catch the relationships of different domains and then generate the dialogue data. We also complicate the dialogues based on the domain relation to enhance the model{'}s capability for co-reference slot tracking. Furthermore, we permute slot values to mitigate the influence of output orders and the problem of incomplete value generation. Experimental results illustrate the superiority of our proposed method compared to previous strong data augmentation baselines on MultiWOZ.",
}
| Data augmentation methods have been a promising direction to improve the performance of small models for low-resource dialogue state tracking. However, traditional methods rely on pre-defined user goals and neglect the importance of data complexity in this task. In this paper, we propose EDZ-DA, an Easy-to-Difficult Zero-shot Data Augmentation framework for low-resource dialogue state tracking that utilizes large language models to automatically catch the relationships of different domains and then generate the dialogue data. We also complicate the dialogues based on the domain relation to enhance the model{'}s capability for co-reference slot tracking. Furthermore, we permute slot values to mitigate the influence of output orders and the problem of incomplete value generation. Experimental results illustrate the superiority of our proposed method compared to previous strong data augmentation baselines on MultiWOZ. | [
"Gu, Ming",
"Yang, Yan"
]
| Plan, Generate and Complicate: Improving Low-resource Dialogue State Tracking via Easy-to-Difficult Zero-shot Data Augmentation | findings-acl.417 | Poster | 2406.08860v1 |
https://aclanthology.org/2024.findings-acl.418.bib | @inproceedings{quan-2024-dmoerm,
title = "{DM}o{ERM}: Recipes of Mixture-of-Experts for Effective Reward Modeling",
author = "Quan, Shanghaoran",
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.418",
pages = "7006--7028",
abstract = "The performance of the reward model (RM) is a critical factor in improving the effectiveness of the large language model (LLM) during alignment fine-tuning. There remain two challenges in RM training: 1) training the same RM using various categories of data may cause its generalization performance to suffer from multi-task disturbance, and 2) the human annotation consistency rate is generally only 60{\%} to 75{\%}, causing training data to contain a lot of noise. To tackle these two challenges, we introduced the idea of Mixture-of-Experts (MoE) into the field of RM for the first time. We propose the Double-Layer MoE RM (DMoERM). The outer layer MoE is a sparse model. After classifying an input into task categories, we route it to the corresponding inner layer task-specific model. The inner layer MoE is a dense model. We decompose the specific task into multiple capability dimensions and individually fine-tune a LoRA expert on each one. Their outputs are then synthesized by an MLP to compute the final rewards. To minimize costs, we call a public LLM API to obtain the capability preference labels. The validation on manually labeled datasets confirms that our model attains superior consistency with human preference and outstrips advanced generative approaches. Meanwhile, through BoN sampling and RL experiments, we demonstrate that our model outperforms state-of-the-art ensemble methods of RM and mitigates the overoptimization problem. Our code is available at: https://github.com/quanshr/DMoERM.",
}
| The performance of the reward model (RM) is a critical factor in improving the effectiveness of the large language model (LLM) during alignment fine-tuning. There remain two challenges in RM training: 1) training the same RM using various categories of data may cause its generalization performance to suffer from multi-task disturbance, and 2) the human annotation consistency rate is generally only 60{\%} to 75{\%}, causing training data to contain a lot of noise. To tackle these two challenges, we introduced the idea of Mixture-of-Experts (MoE) into the field of RM for the first time. We propose the Double-Layer MoE RM (DMoERM). The outer layer MoE is a sparse model. After classifying an input into task categories, we route it to the corresponding inner layer task-specific model. The inner layer MoE is a dense model. We decompose the specific task into multiple capability dimensions and individually fine-tune a LoRA expert on each one. Their outputs are then synthesized by an MLP to compute the final rewards. To minimize costs, we call a public LLM API to obtain the capability preference labels. The validation on manually labeled datasets confirms that our model attains superior consistency with human preference and outstrips advanced generative approaches. Meanwhile, through BoN sampling and RL experiments, we demonstrate that our model outperforms state-of-the-art ensemble methods of RM and mitigates the overoptimization problem. Our code is available at: https://github.com/quanshr/DMoERM. | [
"Quan, Shanghaoran"
]
| {DM}o{ERM}: Recipes of Mixture-of-Experts for Effective Reward Modeling | findings-acl.418 | Poster | 2407.04185v2 |
https://aclanthology.org/2024.findings-acl.419.bib | @inproceedings{yamada-ri-2024-leia,
title = "{LEIA}: Facilitating Cross-lingual Knowledge Transfer in Language Models with Entity-based Data Augmentation",
author = "Yamada, Ikuya and
Ri, Ryokan",
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.419",
pages = "7029--7039",
abstract = "Adapting English-based large language models (LLMs) to other languages has become increasingly popular due to the efficiency and potential of cross-lingual transfer. However, existing language adaptation methods often overlook the benefits of cross-lingual supervision. In this study, we introduce LEIA, a language adaptation tuning method that utilizes Wikipedia entity names aligned across languages. This method involves augmenting the target language corpus with English entity names and training the model using left-to-right language modeling. We assess LEIA on diverse question answering datasets using 7B-parameter LLMs, demonstrating significant performance gains across various non-English languages.",
}
| Adapting English-based large language models (LLMs) to other languages has become increasingly popular due to the efficiency and potential of cross-lingual transfer. However, existing language adaptation methods often overlook the benefits of cross-lingual supervision. In this study, we introduce LEIA, a language adaptation tuning method that utilizes Wikipedia entity names aligned across languages. This method involves augmenting the target language corpus with English entity names and training the model using left-to-right language modeling. We assess LEIA on diverse question answering datasets using 7B-parameter LLMs, demonstrating significant performance gains across various non-English languages. | [
"Yamada, Ikuya",
"Ri, Ryokan"
]
| {LEIA}: Facilitating Cross-lingual Knowledge Transfer in Language Models with Entity-based Data Augmentation | findings-acl.419 | Poster | 2309.12763v2 |
https://aclanthology.org/2024.findings-acl.420.bib | @inproceedings{chen-etal-2024-comments,
title = "Comments as Natural Logic Pivots: Improve Code Generation via Comment Perspective",
author = "Chen, Yijie and
Liu, Yijin and
Meng, Fandong and
Chen, Yufeng and
Xu, Jinan 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.420",
pages = "7040--7051",
abstract = "Code generation aims to understand the problem description and generate corresponding code snippets, where existing works generally decompose such complex tasks into intermediate steps by prompting strategies, such as Chain-of-Thought and its variants. While these studies have achieved some success, their effectiveness is highly dependent on the capabilities of advanced Large Language Models (LLMs) such as GPT-4, particularly in terms of API calls, which significantly limits their practical applicability. Consequently, how to enhance the code generation capabilities of small and medium-scale code LLMs without significantly increasing training costs is an appealing challenge. In this paper, we suggest that code comments are the natural logic pivot between natural language and code language and propose using comments to boost the code generation ability of code LLMs. Concretely, we propose MANGO (comMents As Natural loGic pivOts), including a comment contrastive training strategy and a corresponding logical comment decoding strategy. Experiments are performed on HumanEval and MBPP, utilizing StarCoder and WizardCoder as backbone models, and encompassing model parameter sizes between 3B and 7B. The results indicate that MANGO significantly improves the code pass rate based on the strong baselines. Meanwhile, the robustness of the logical comment decoding strategy is notably higher than the Chain-of-thoughts prompting.",
}
| Code generation aims to understand the problem description and generate corresponding code snippets, where existing works generally decompose such complex tasks into intermediate steps by prompting strategies, such as Chain-of-Thought and its variants. While these studies have achieved some success, their effectiveness is highly dependent on the capabilities of advanced Large Language Models (LLMs) such as GPT-4, particularly in terms of API calls, which significantly limits their practical applicability. Consequently, how to enhance the code generation capabilities of small and medium-scale code LLMs without significantly increasing training costs is an appealing challenge. In this paper, we suggest that code comments are the natural logic pivot between natural language and code language and propose using comments to boost the code generation ability of code LLMs. Concretely, we propose MANGO (comMents As Natural loGic pivOts), including a comment contrastive training strategy and a corresponding logical comment decoding strategy. Experiments are performed on HumanEval and MBPP, utilizing StarCoder and WizardCoder as backbone models, and encompassing model parameter sizes between 3B and 7B. The results indicate that MANGO significantly improves the code pass rate based on the strong baselines. Meanwhile, the robustness of the logical comment decoding strategy is notably higher than the Chain-of-thoughts prompting. | [
"Chen, Yijie",
"Liu, Yijin",
"Meng, F",
"ong",
"Chen, Yufeng",
"Xu, Jinan",
"Zhou, Jie"
]
| Comments as Natural Logic Pivots: Improve Code Generation via Comment Perspective | findings-acl.420 | Poster | 2404.07549v1 |
https://aclanthology.org/2024.findings-acl.421.bib | @inproceedings{dai-etal-2024-cocktail,
title = "Cocktail: A Comprehensive Information Retrieval Benchmark with {LLM}-Generated Documents Integration",
author = "Dai, Sunhao and
Liu, Weihao and
Zhou, Yuqi and
Pang, Liang and
Ruan, Rongju and
Wang, Gang and
Dong, Zhenhua and
Xu, Jun and
Wen, Ji-Rong",
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.421",
pages = "7052--7074",
abstract = "The proliferation of Large Language Models (LLMs) has led to an influx of AI-generated content (AIGC) on the internet, transforming the corpus of Information Retrieval (IR) systems from solely human-written to a coexistence with LLM-generated content. The impact of this surge in AIGC on IR systems remains an open question, with the primary challenge being the lack of a dedicated benchmark for researchers. In this paper, we introduce Cocktail, a comprehensive benchmark tailored for evaluating IR models in this mixed-sourced data landscape of the LLM era. Cocktail consists of 16 diverse datasets with mixed human-written and LLM-generated corpora across various text retrieval tasks and domains. Additionally, to avoid the potential bias from previously included dataset information in LLMs, we also introduce an up-to-date dataset, named NQ-UTD, with queries derived from recent events. Through conducting over 1,000 experiments to assess state-of-the-art retrieval models against the benchmarked datasets in Cocktail, we uncover a clear trade-off between ranking performance and source bias in neural retrieval models, highlighting the necessity for a balanced approach in designing future IR systems. We hope Cocktail can serve as a foundational resource for IR research in the LLM era, with all data and code publicly available at https://github.com/KID-22/Cocktail.",
}
| The proliferation of Large Language Models (LLMs) has led to an influx of AI-generated content (AIGC) on the internet, transforming the corpus of Information Retrieval (IR) systems from solely human-written to a coexistence with LLM-generated content. The impact of this surge in AIGC on IR systems remains an open question, with the primary challenge being the lack of a dedicated benchmark for researchers. In this paper, we introduce Cocktail, a comprehensive benchmark tailored for evaluating IR models in this mixed-sourced data landscape of the LLM era. Cocktail consists of 16 diverse datasets with mixed human-written and LLM-generated corpora across various text retrieval tasks and domains. Additionally, to avoid the potential bias from previously included dataset information in LLMs, we also introduce an up-to-date dataset, named NQ-UTD, with queries derived from recent events. Through conducting over 1,000 experiments to assess state-of-the-art retrieval models against the benchmarked datasets in Cocktail, we uncover a clear trade-off between ranking performance and source bias in neural retrieval models, highlighting the necessity for a balanced approach in designing future IR systems. We hope Cocktail can serve as a foundational resource for IR research in the LLM era, with all data and code publicly available at https://github.com/KID-22/Cocktail. | [
"Dai, Sunhao",
"Liu, Weihao",
"Zhou, Yuqi",
"Pang, Liang",
"Ruan, Rongju",
"Wang, Gang",
"Dong, Zhenhua",
"Xu, Jun",
"Wen, Ji-Rong"
]
| Cocktail: A Comprehensive Information Retrieval Benchmark with {LLM}-Generated Documents Integration | findings-acl.421 | Poster | 2405.16546v2 |
https://aclanthology.org/2024.findings-acl.422.bib | @inproceedings{feng-etal-2024-continual,
title = "Continual Dialogue State Tracking via Reason-of-Select Distillation",
author = "Feng, Yujie and
Liu, Bo and
Dong, Xiaoyu and
Lu, Zexin and
Zhan, Li-Ming and
Wu, Xiao-Ming and
Lam, Albert",
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.422",
pages = "7075--7087",
abstract = "An ideal dialogue system requires continuous skill acquisition and adaptation to new tasks while retaining prior knowledge. Dialogue State Tracking (DST), vital in these systems, often involves learning new services, confronting catastrophic forgetting and a critical capability loss termed the {``}Value Selection Quandary{''}. To address these challenges, we introduce the Reason-of-Select (RoS) distillation method by enhancing smaller models with a novel {``}meta-reasoning{''} capability. Meta-reasoning, employing an enhanced multi-domain perspective, combines fragments of meta-knowledge from domain-specific dialogues during continual learning, transcending traditional single-perspective reasoning. This domain bootstrapping process enhances the model{'}s ability to dissect intricate dialogues from multiple possible values, and its domain-agnostic property aligns data distribution across different domains, effectively mitigating forgetting. Besides, two novel improvements, {``}multi-value resolution{''} strategy and Semantic Contrastive Reasoning Selection method, significantly enhance RoS by generating DST-specific selection chains and mitigating hallucinations in teachers{'} reasoning, ensuring effective and reliable knowledge transfer. Extensive experiments validate the exceptional performance and robust generalization capabilities of our method.",
}
| An ideal dialogue system requires continuous skill acquisition and adaptation to new tasks while retaining prior knowledge. Dialogue State Tracking (DST), vital in these systems, often involves learning new services, confronting catastrophic forgetting and a critical capability loss termed the {``}Value Selection Quandary{''}. To address these challenges, we introduce the Reason-of-Select (RoS) distillation method by enhancing smaller models with a novel {``}meta-reasoning{''} capability. Meta-reasoning, employing an enhanced multi-domain perspective, combines fragments of meta-knowledge from domain-specific dialogues during continual learning, transcending traditional single-perspective reasoning. This domain bootstrapping process enhances the model{'}s ability to dissect intricate dialogues from multiple possible values, and its domain-agnostic property aligns data distribution across different domains, effectively mitigating forgetting. Besides, two novel improvements, {``}multi-value resolution{''} strategy and Semantic Contrastive Reasoning Selection method, significantly enhance RoS by generating DST-specific selection chains and mitigating hallucinations in teachers{'} reasoning, ensuring effective and reliable knowledge transfer. Extensive experiments validate the exceptional performance and robust generalization capabilities of our method. | [
"Feng, Yujie",
"Liu, Bo",
"Dong, Xiaoyu",
"Lu, Zexin",
"Zhan, Li-Ming",
"Wu, Xiao-Ming",
"Lam, Albert"
]
| Continual Dialogue State Tracking via Reason-of-Select Distillation | findings-acl.422 | Poster | 2302.08220v2 |
https://aclanthology.org/2024.findings-acl.423.bib | @inproceedings{li-etal-2024-spotting,
title = "Spotting {AI}{'}s Touch: Identifying {LLM}-Paraphrased Spans in Text",
author = "Li, Yafu and
Wang, Zhilin and
Cui, Leyang and
Bi, Wei and
Shi, Shuming and
Zhang, Yue",
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.423",
pages = "7088--7107",
abstract = "AI-generated text detection has attracted increasing attention as powerful language models approach human-level generation. Limited work is devoted to detecting (partially) AI-paraphrased texts. However, AI paraphrasing is commonly employed in various application scenarios for text refinement and diversity. To this end, we propose a novel detection framework, paraphrased text span detection (PTD), aiming to identify paraphrased text spans within a text. Different from text-level detection, PTD takes in the full text and assigns each of the sentences with a score indicating the paraphrasing degree. We construct a dedicated dataset, PASTED, for paraphrased text span detection. Both in-distribution and out-of-distribution results demonstrate the effectiveness of PTD models in identifying AI-paraphrased text spans. Statistical and model analysis explains the crucial role of the surrounding context of the paraphrased text spans. Extensive experiments show that PTD models can generalize to versatile paraphrasing prompts as well as multiple paraphrased text spans.",
}
| AI-generated text detection has attracted increasing attention as powerful language models approach human-level generation. Limited work is devoted to detecting (partially) AI-paraphrased texts. However, AI paraphrasing is commonly employed in various application scenarios for text refinement and diversity. To this end, we propose a novel detection framework, paraphrased text span detection (PTD), aiming to identify paraphrased text spans within a text. Different from text-level detection, PTD takes in the full text and assigns each of the sentences with a score indicating the paraphrasing degree. We construct a dedicated dataset, PASTED, for paraphrased text span detection. Both in-distribution and out-of-distribution results demonstrate the effectiveness of PTD models in identifying AI-paraphrased text spans. Statistical and model analysis explains the crucial role of the surrounding context of the paraphrased text spans. Extensive experiments show that PTD models can generalize to versatile paraphrasing prompts as well as multiple paraphrased text spans. | [
"Li, Yafu",
"Wang, Zhilin",
"Cui, Leyang",
"Bi, Wei",
"Shi, Shuming",
"Zhang, Yue"
]
| Spotting {AI}{'}s Touch: Identifying {LLM}-Paraphrased Spans in Text | findings-acl.423 | Poster | 2405.12689v2 |
https://aclanthology.org/2024.findings-acl.424.bib | @inproceedings{lu-etal-2024-sofa,
title = "{S}o{FA}: Shielded On-the-fly Alignment via Priority Rule Following",
author = "Lu, Xinyu and
Yu, Bowen and
Lu, Yaojie and
Lin, Hongyu and
Yu, Haiyang and
Sun, Le and
Han, Xianpei and
Li, Yongbin",
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.424",
pages = "7108--7136",
abstract = "The alignment problem in Large Language Models (LLMs) involves adapting them to the broad spectrum of human values. This requirement challenges existing alignment methods due to diversity of preferences and regulatory standards. This paper introduces a novel alignment paradigm, priority rule following, which defines rules as the primary control mechanism in each dialog, prioritizing them over user instructions. Our preliminary analysis reveals that even the advanced LLMs, such as GPT-4, exhibit shortcomings in understanding and prioritizing the rules. Therefore, we present PriorityDistill, a semi-automated approach for distilling priority following signals from LLM simulations to ensure robust rule integration and adherence. Our experiments show that this method not only effectively minimizes misalignments utilizing only one general rule but also adapts smoothly to various unseen rules, ensuring they are shielded from hijacking and that the model responds appropriately.",
}
| The alignment problem in Large Language Models (LLMs) involves adapting them to the broad spectrum of human values. This requirement challenges existing alignment methods due to diversity of preferences and regulatory standards. This paper introduces a novel alignment paradigm, priority rule following, which defines rules as the primary control mechanism in each dialog, prioritizing them over user instructions. Our preliminary analysis reveals that even the advanced LLMs, such as GPT-4, exhibit shortcomings in understanding and prioritizing the rules. Therefore, we present PriorityDistill, a semi-automated approach for distilling priority following signals from LLM simulations to ensure robust rule integration and adherence. Our experiments show that this method not only effectively minimizes misalignments utilizing only one general rule but also adapts smoothly to various unseen rules, ensuring they are shielded from hijacking and that the model responds appropriately. | [
"Lu, Xinyu",
"Yu, Bowen",
"Lu, Yaojie",
"Lin, Hongyu",
"Yu, Haiyang",
"Sun, Le",
"Han, Xianpei",
"Li, Yongbin"
]
| {S}o{FA}: Shielded On-the-fly Alignment via Priority Rule Following | findings-acl.424 | Poster | 2402.17358v1 |
https://aclanthology.org/2024.findings-acl.425.bib | @inproceedings{goldstein-stanovsky-2024-zombies,
title = "Do Zombies Understand? A Choose-Your-Own-Adventure Exploration of Machine Cognition",
author = "Goldstein, Ariel and
Stanovsky, Gabriel",
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.425",
pages = "7137--7143",
abstract = "Recent advances in LLMs have sparked a debate on whether they understand text. In this position paper, we argue that opponents in this debate hold different definitions for understanding, and particularly differ in their view on the role of consciousness. To substantiate this claim, we propose a thought experiment involving an open-source chatbot Z which excels on every possible benchmark, seemingly without subjective experience. We ask whether Z is capable of understanding, and show that different schools of thought within seminal AI research seem to answer this question differently, uncovering their terminological disagreement. Moving forward, we propose two distinct working definitions for understanding which explicitly acknowledge the question of consciousness, and draw connections with a rich literature in philosophy, psychology and neuroscience.",
}
| Recent advances in LLMs have sparked a debate on whether they understand text. In this position paper, we argue that opponents in this debate hold different definitions for understanding, and particularly differ in their view on the role of consciousness. To substantiate this claim, we propose a thought experiment involving an open-source chatbot Z which excels on every possible benchmark, seemingly without subjective experience. We ask whether Z is capable of understanding, and show that different schools of thought within seminal AI research seem to answer this question differently, uncovering their terminological disagreement. Moving forward, we propose two distinct working definitions for understanding which explicitly acknowledge the question of consciousness, and draw connections with a rich literature in philosophy, psychology and neuroscience. | [
"Goldstein, Ariel",
"Stanovsky, Gabriel"
]
| Do Zombies Understand? A Choose-Your-Own-Adventure Exploration of Machine Cognition | findings-acl.425 | Poster | 2403.00499v2 |
https://aclanthology.org/2024.findings-acl.426.bib | @inproceedings{christ-etal-2024-modeling,
title = "Modeling Emotional Trajectories in Written Stories Utilizing Transformers and Weakly-Supervised Learning",
author = {Christ, Lukas and
Amiriparian, Shahin and
Milling, Manuel and
Aslan, Ilhan and
Schuller, Bj{\"o}rn},
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.426",
pages = "7144--7159",
abstract = "Telling stories is an integral part of human communication which can evoke emotions and influence the affective states of the audience. Automatically modeling emotional trajectories in stories has thus attracted considerable scholarly interest. However, as most existing works have been limited to unsupervised dictionary-based approaches, there is no benchmark for this task. We address this gap by introducing continuous valence and arousal labels for an existing dataset of children{'}s stories originally annotated with discrete emotion categories. We collect additional annotations for this data and map the categorical labels to the continuous valence and arousal space. For predicting the thus obtained emotionality signals, we fine-tune a DeBERTa model and improve upon this baseline via a weakly supervised learning approach. The best configuration achieves a Concordance Correlation Coefficient (CCC) of .8221 for valence and .7125 for arousal on the test set, demonstrating the efficacy of our proposed approach. A detailed analysis shows the extent to which the results vary depending on factors such as the author, the individual story, or the section within the story. In addition, we uncover the weaknesses of our approach by investigating examples that prove to be difficult to predict.",
}
| Telling stories is an integral part of human communication which can evoke emotions and influence the affective states of the audience. Automatically modeling emotional trajectories in stories has thus attracted considerable scholarly interest. However, as most existing works have been limited to unsupervised dictionary-based approaches, there is no benchmark for this task. We address this gap by introducing continuous valence and arousal labels for an existing dataset of children{'}s stories originally annotated with discrete emotion categories. We collect additional annotations for this data and map the categorical labels to the continuous valence and arousal space. For predicting the thus obtained emotionality signals, we fine-tune a DeBERTa model and improve upon this baseline via a weakly supervised learning approach. The best configuration achieves a Concordance Correlation Coefficient (CCC) of .8221 for valence and .7125 for arousal on the test set, demonstrating the efficacy of our proposed approach. A detailed analysis shows the extent to which the results vary depending on factors such as the author, the individual story, or the section within the story. In addition, we uncover the weaknesses of our approach by investigating examples that prove to be difficult to predict. | [
"Christ, Lukas",
"Amiriparian, Shahin",
"Milling, Manuel",
"Aslan, Ilhan",
"Schuller, Bj{\\\"o}rn"
]
| Modeling Emotional Trajectories in Written Stories Utilizing Transformers and Weakly-Supervised Learning | findings-acl.426 | Poster | 2406.02251v1 |
https://aclanthology.org/2024.findings-acl.427.bib | @inproceedings{cao-etal-2024-rap,
title = "{RAP}: Efficient Text-Video Retrieval with Sparse-and-Correlated Adapter",
author = "Cao, Meng and
Tang, Haoran and
Huang, Jinfa and
Jin, Peng and
Zhang, Can and
Liu, Ruyang and
Chen, Long and
Liang, Xiaodan and
Yuan, Li and
Li, Ge",
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.427",
pages = "7160--7174",
abstract = "Text-Video Retrieval (TVR) aims to align relevant video content with natural language queries. To date, most of the state-of-the-art TVR methods learn image-to-video transfer learning based on the large-scale pre-trained vision-language models (e.g., CLIP). However, fully fine-tuning these pre-trained models for TVR incurs prohibitively expensive computation cost. To this end, we propose to conduct efficient text-video Retrieval with a salient-and-correlated AdaPter (RAP), i.e., fine-tuning the pre-trained model with a few parameterized layers. To accommodate the text-video scenario, we equip our RAP with two indispensable characteristics including temporal sparsity and correlation. Specifically, we propose a low-rank modulation module to refine the per-image features from frozen CLIP backbone, which accentuates silent frames within the video features while alleviating temporal redundancy. Besides, we introduce an asynchronous self-attention mechanism which firstly selects top responsive visual patch and augments the correlation modeling between them with learnable temporal and patch offsets. Extensive experiments on four TVR datasets demonstrate that our RAP achieves superior or comparable performance compared to the fully fine-tuned counterpart and other parameter-efficient finetuning methods.",
}
| Text-Video Retrieval (TVR) aims to align relevant video content with natural language queries. To date, most of the state-of-the-art TVR methods learn image-to-video transfer learning based on the large-scale pre-trained vision-language models (e.g., CLIP). However, fully fine-tuning these pre-trained models for TVR incurs prohibitively expensive computation cost. To this end, we propose to conduct efficient text-video Retrieval with a salient-and-correlated AdaPter (RAP), i.e., fine-tuning the pre-trained model with a few parameterized layers. To accommodate the text-video scenario, we equip our RAP with two indispensable characteristics including temporal sparsity and correlation. Specifically, we propose a low-rank modulation module to refine the per-image features from frozen CLIP backbone, which accentuates silent frames within the video features while alleviating temporal redundancy. Besides, we introduce an asynchronous self-attention mechanism which firstly selects top responsive visual patch and augments the correlation modeling between them with learnable temporal and patch offsets. Extensive experiments on four TVR datasets demonstrate that our RAP achieves superior or comparable performance compared to the fully fine-tuned counterpart and other parameter-efficient finetuning methods. | [
"Cao, Meng",
"Tang, Haoran",
"Huang, Jinfa",
"Jin, Peng",
"Zhang, Can",
"Liu, Ruyang",
"Chen, Long",
"Liang, Xiaodan",
"Yuan, Li",
"Li, Ge"
]
| {RAP}: Efficient Text-Video Retrieval with Sparse-and-Correlated Adapter | findings-acl.427 | Poster | 2303.13220v1 |
https://aclanthology.org/2024.findings-acl.428.bib | @inproceedings{wang-etal-2024-benchmarking,
title = "Benchmarking and Improving Long-Text Translation with Large Language Models",
author = "Wang, Longyue and
Du, Zefeng and
Jiao, Wenxiang and
Lyu, Chenyang and
Pang, Jianhui and
Cui, Leyang and
Song, Kaiqiang and
Wong, Derek and
Shi, Shuming and
Tu, Zhaopeng",
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.428",
pages = "7175--7187",
abstract = "Recent studies have illuminated the promising capabilities of large language models (LLMs) in handling long texts. However, their performance in machine translation (MT) of long documents remains underexplored. This paper aims to shed light on how LLMs navigate this complex task, offering a comprehensive evaluation of their capabilities and limitations in long-text MT. First, we collect and construct an instruction-based benchmark dataset, specifically designed for the finetuning and evaluation of LLMs, encompassing multilingual, multi-domain, and document-level parallel data. Second, we conduct a comprehensive comparison between MT and LLM models concerning document-level translation. Our analysis uncovers that LLMs exhibit shortcomings in long-text domains, and their performance diminishes as document size escalates. By exploiting various extrapolation strategies, we enhance the capacity of LLMs to translate longer texts. We release data, code, and models at https://github.com/longyuewangdcu/Document-MT-LLM.",
}
| Recent studies have illuminated the promising capabilities of large language models (LLMs) in handling long texts. However, their performance in machine translation (MT) of long documents remains underexplored. This paper aims to shed light on how LLMs navigate this complex task, offering a comprehensive evaluation of their capabilities and limitations in long-text MT. First, we collect and construct an instruction-based benchmark dataset, specifically designed for the finetuning and evaluation of LLMs, encompassing multilingual, multi-domain, and document-level parallel data. Second, we conduct a comprehensive comparison between MT and LLM models concerning document-level translation. Our analysis uncovers that LLMs exhibit shortcomings in long-text domains, and their performance diminishes as document size escalates. By exploiting various extrapolation strategies, we enhance the capacity of LLMs to translate longer texts. We release data, code, and models at https://github.com/longyuewangdcu/Document-MT-LLM. | [
"Wang, Longyue",
"Du, Zefeng",
"Jiao, Wenxiang",
"Lyu, Chenyang",
"Pang, Jianhui",
"Cui, Leyang",
"Song, Kaiqiang",
"Wong, Derek",
"Shi, Shuming",
"Tu, Zhaopeng"
]
| Benchmarking and Improving Long-Text Translation with Large Language Models | findings-acl.428 | Poster | 2405.04164v1 |
https://aclanthology.org/2024.findings-acl.429.bib | @inproceedings{fan-etal-2024-personalized,
title = "Personalized Topic Selection Model for Topic-Grounded Dialogue",
author = "Fan, Shixuan and
Wei, Wei and
Wen, Xiaofei and
Mao, Xian-Ling and
Chen, Jixiong and
Chen, Dangyang",
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.429",
pages = "7188--7202",
abstract = "Recently, the topic-grounded dialogue (TGD) system has become increasingly popular as its powerful capability to actively guide users to accomplish specific tasks through topic-guided conversations. Most existing works utilize side information (e.g. topics or personas) in isolation to enhance the topic selection ability. However, due to disregarding the noise within these auxiliary information sources and their mutual influence, current models tend to predict user-uninteresting and contextually irrelevant topics. To build user-engaging and coherent dialogue agent, we propose a personalized topic selection model for topic-grounded dialogue, named PETD, which takes account of the interaction of side information to selectively aggregate such information for more accurately predicting subsequent topics. Specifically, we evaluate the correlation between global topics and personas and selectively incorporate the global topics aligned with user personas. Furthermore, we propose a contrastive learning based persona selector to filter relevant personas under the constraint of lacking pertinent persona annotations. Throughout the selection and generation, diverse relevant side information is considered. Extensive experiments demonstrate that our proposed method can generate engaging and diverse responses, outperforming state-of-the-art baselines across various evaluation metrics.",
}
| Recently, the topic-grounded dialogue (TGD) system has become increasingly popular as its powerful capability to actively guide users to accomplish specific tasks through topic-guided conversations. Most existing works utilize side information (e.g. topics or personas) in isolation to enhance the topic selection ability. However, due to disregarding the noise within these auxiliary information sources and their mutual influence, current models tend to predict user-uninteresting and contextually irrelevant topics. To build user-engaging and coherent dialogue agent, we propose a personalized topic selection model for topic-grounded dialogue, named PETD, which takes account of the interaction of side information to selectively aggregate such information for more accurately predicting subsequent topics. Specifically, we evaluate the correlation between global topics and personas and selectively incorporate the global topics aligned with user personas. Furthermore, we propose a contrastive learning based persona selector to filter relevant personas under the constraint of lacking pertinent persona annotations. Throughout the selection and generation, diverse relevant side information is considered. Extensive experiments demonstrate that our proposed method can generate engaging and diverse responses, outperforming state-of-the-art baselines across various evaluation metrics. | [
"Fan, Shixuan",
"Wei, Wei",
"Wen, Xiaofei",
"Mao, Xian-Ling",
"Chen, Jixiong",
"Chen, Dangyang"
]
| Personalized Topic Selection Model for Topic-Grounded Dialogue | findings-acl.429 | Poster | 2406.01988v1 |
https://aclanthology.org/2024.findings-acl.430.bib | @inproceedings{li-etal-2024-debiasing,
title = "Debiasing In-Context Learning by Instructing {LLM}s How to Follow Demonstrations",
author = "Li, Lvxue and
Chen, Jiaqi and
Lu, Xinyu and
Lu, Yaojie and
Lin, Hongyu and
Zhou, Shuheng and
Zhu, Huijia and
Wang, Weiqiang and
Liu, Zhongyi and
Han, Xianpei and
Sun, Le",
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.430",
pages = "7203--7215",
abstract = "In-context learning(ICL) has gained considerable attention due to its data efficiency and task adaptability. Unfortunately, ICL suffers from the demonstration bias, i.e., its performance and robustness are severely affected by the selection and ordering of demonstrations. In this paper, we identify that such demonstration bias may primarily stem from the semantic ambiguity induced by demonstrations, i.e., a demonstration may indicate multiple input-to-label mappings and its mapping can be interpreted differently in different contexts by LLMs. Such semantic ambiguity disrupts task comprehension during ICL and results in performance fluctuations. To resolve the semantic ambiguity problem, this paper further proposes two de-biasing strategies to mitigate demonstration bias in in-context learning. Experiments on six datasets show that our methods can effectively alleviate demonstration bias and significantly improve task performance.",
}
| In-context learning(ICL) has gained considerable attention due to its data efficiency and task adaptability. Unfortunately, ICL suffers from the demonstration bias, i.e., its performance and robustness are severely affected by the selection and ordering of demonstrations. In this paper, we identify that such demonstration bias may primarily stem from the semantic ambiguity induced by demonstrations, i.e., a demonstration may indicate multiple input-to-label mappings and its mapping can be interpreted differently in different contexts by LLMs. Such semantic ambiguity disrupts task comprehension during ICL and results in performance fluctuations. To resolve the semantic ambiguity problem, this paper further proposes two de-biasing strategies to mitigate demonstration bias in in-context learning. Experiments on six datasets show that our methods can effectively alleviate demonstration bias and significantly improve task performance. | [
"Li, Lvxue",
"Chen, Jiaqi",
"Lu, Xinyu",
"Lu, Yaojie",
"Lin, Hongyu",
"Zhou, Shuheng",
"Zhu, Huijia",
"Wang, Weiqiang",
"Liu, Zhongyi",
"Han, Xianpei",
"Sun, Le"
]
| Debiasing In-Context Learning by Instructing {LLM}s How to Follow Demonstrations | findings-acl.430 | Poster | 2407.02030v1 |
https://aclanthology.org/2024.findings-acl.431.bib | @inproceedings{vlachos-etal-2024-comparing,
title = "Comparing Data Augmentation Methods for End-to-End Task-Oriented Dialog Systems",
author = "Vlachos, Christos and
Stafylakis, Themos and
Androutsopoulos, Ion",
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.431",
pages = "7216--7240",
abstract = "Creating effective and reliable task-oriented dialog systems (ToDSs) is challenging, not only because of the complex structure of these systems, but also due to the scarcity of training data, especially when several modules need to be trained separately, each one with its own input/output training examples. Data augmentation (DA), whereby synthetic training examples are added to the training data, has been successful in other NLP systems, but has not been explored as extensively in ToDSs. We empirically evaluate the effectiveness of DA methods in an end-to-end ToDS setting, where a single system is trained to handle all processing stages, from user inputs to system outputs. We experiment with two ToDSs (UBAR, GALAXY) on two datasets (MultiWOZ, KVRET). We consider three types of DA methods (word-level, sentence-level, dialog-level), comparing eight DA methods that have shown promising results in ToDSs and other NLP systems. We show that all DA methods considered are beneficial, and we highlight the best ones, also providing advice to practitioners. We also introduce a more challenging few-shot cross-domain ToDS setting, reaching similar conclusions.",
}
| Creating effective and reliable task-oriented dialog systems (ToDSs) is challenging, not only because of the complex structure of these systems, but also due to the scarcity of training data, especially when several modules need to be trained separately, each one with its own input/output training examples. Data augmentation (DA), whereby synthetic training examples are added to the training data, has been successful in other NLP systems, but has not been explored as extensively in ToDSs. We empirically evaluate the effectiveness of DA methods in an end-to-end ToDS setting, where a single system is trained to handle all processing stages, from user inputs to system outputs. We experiment with two ToDSs (UBAR, GALAXY) on two datasets (MultiWOZ, KVRET). We consider three types of DA methods (word-level, sentence-level, dialog-level), comparing eight DA methods that have shown promising results in ToDSs and other NLP systems. We show that all DA methods considered are beneficial, and we highlight the best ones, also providing advice to practitioners. We also introduce a more challenging few-shot cross-domain ToDS setting, reaching similar conclusions. | [
"Vlachos, Christos",
"Stafylakis, Themos",
"Androutsopoulos, Ion"
]
| Comparing Data Augmentation Methods for End-to-End Task-Oriented Dialog Systems | findings-acl.431 | Poster | 2310.10380v1 |
https://aclanthology.org/2024.findings-acl.432.bib | @inproceedings{ma-etal-2024-ms2sl,
title = "{MS}2{SL}: Multimodal Spoken Data-Driven Continuous Sign Language Production",
author = "Ma, Jian and
Wang, Wenguan and
Yang, Yi and
Zheng, 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.432",
pages = "7241--7254",
abstract = "Sign language understanding has made significant strides; however, there is still no viable solution for generating sign sequences directlyfrom entire spoken content, e.g., text or speech. In this paper, we propose a unified framework for continuous sign language production, easing communication between sign and non-sign language users. In particular, a sequence diffusion model, utilizing embeddings extracted from text or speech, is crafted to generate sign predictions step by step. Moreover, by creating a joint embedding space for text, audio, and sign, we bind these modalities and leverage the semantic consistency among them to provide informative feedback for the model training. This embedding-consistency learning strategy minimizes the reliance on sign triplets and ensures continuous model refinement, evenwith a missing audio modality. Experiments on How2Sign and PHOENIX14T datasets demonstrate that our model achieves competitive performance in sign language production.",
}
| Sign language understanding has made significant strides; however, there is still no viable solution for generating sign sequences directlyfrom entire spoken content, e.g., text or speech. In this paper, we propose a unified framework for continuous sign language production, easing communication between sign and non-sign language users. In particular, a sequence diffusion model, utilizing embeddings extracted from text or speech, is crafted to generate sign predictions step by step. Moreover, by creating a joint embedding space for text, audio, and sign, we bind these modalities and leverage the semantic consistency among them to provide informative feedback for the model training. This embedding-consistency learning strategy minimizes the reliance on sign triplets and ensures continuous model refinement, evenwith a missing audio modality. Experiments on How2Sign and PHOENIX14T datasets demonstrate that our model achieves competitive performance in sign language production. | [
"Ma, Jian",
"Wang, Wenguan",
"Yang, Yi",
"Zheng, Feng"
]
| {MS}2{SL}: Multimodal Spoken Data-Driven Continuous Sign Language Production | findings-acl.432 | Poster | 2407.12842v1 |
https://aclanthology.org/2024.findings-acl.433.bib | @inproceedings{zhao-etal-2024-bba,
title = "{BBA}: Bi-Modal Behavioral Alignment for Reasoning with Large Vision-Language Models",
author = "Zhao, Xueliang and
Huang, Xinting and
Fu, Tingchen and
Li, Qintong and
Gong, Shansan and
Liu, Lemao and
Bi, Wei and
Kong, Lingpeng",
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.433",
pages = "7255--7279",
abstract = "Multimodal reasoning stands as a pivotal capability for large vision-language models (LVLMs). The integration with Domain-Specific Languages (DSL), offering precise visual representations, equips these models with the opportunity to execute more accurate reasoning in complex and professional domains. However, the vanilla Chain-of-Thought (CoT) prompting method faces challenges in effectively leveraging the unique strengths of visual and DSL representations, primarily due to their differing reasoning mechanisms. Additionally, it often falls short in addressing critical steps in multi-step reasoning tasks. To mitigate these challenges, we introduce the Bi-Modal Behavioral Alignment (BBA) prompting method, designed to maximize the potential of DSL in augmenting complex multi-modal reasoning tasks. This method initiates by guiding LVLMs to create separate reasoning chains for visual and DSL representations. Subsequently, it aligns these chains by addressing any inconsistencies, thus achieving a cohesive integration of behaviors from different modalities. Our experiments demonstrate that BBA substantially improves the performance of GPT-4V(ision) on geometry problem solving (28.34{\%} $\to$ 34.22{\%}), chess positional advantage prediction (42.08{\%} $\to$ 46.99{\%}) and molecular property prediction (77.47{\%} $\to$ 83.52{\%}).",
}
| Multimodal reasoning stands as a pivotal capability for large vision-language models (LVLMs). The integration with Domain-Specific Languages (DSL), offering precise visual representations, equips these models with the opportunity to execute more accurate reasoning in complex and professional domains. However, the vanilla Chain-of-Thought (CoT) prompting method faces challenges in effectively leveraging the unique strengths of visual and DSL representations, primarily due to their differing reasoning mechanisms. Additionally, it often falls short in addressing critical steps in multi-step reasoning tasks. To mitigate these challenges, we introduce the Bi-Modal Behavioral Alignment (BBA) prompting method, designed to maximize the potential of DSL in augmenting complex multi-modal reasoning tasks. This method initiates by guiding LVLMs to create separate reasoning chains for visual and DSL representations. Subsequently, it aligns these chains by addressing any inconsistencies, thus achieving a cohesive integration of behaviors from different modalities. Our experiments demonstrate that BBA substantially improves the performance of GPT-4V(ision) on geometry problem solving (28.34{\%} $\to$ 34.22{\%}), chess positional advantage prediction (42.08{\%} $\to$ 46.99{\%}) and molecular property prediction (77.47{\%} $\to$ 83.52{\%}). | [
"Zhao, Xueliang",
"Huang, Xinting",
"Fu, Tingchen",
"Li, Qintong",
"Gong, Shansan",
"Liu, Lemao",
"Bi, Wei",
"Kong, Lingpeng"
]
| {BBA}: Bi-Modal Behavioral Alignment for Reasoning with Large Vision-Language Models | findings-acl.433 | Poster | 2311.10947v2 |
https://aclanthology.org/2024.findings-acl.434.bib | @inproceedings{zheng-etal-2024-partialformer,
title = "{P}artial{F}ormer: Modeling Part Instead of Whole for Machine Translation",
author = "Zheng, Tong and
Li, Bei and
Bao, Huiwen and
Wang, Jiale and
Shan, Weiqiao and
Xiao, Tong and
Zhu, JingBo",
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.434",
pages = "7280--7294",
abstract = "The design choices in Transformer feed-forward neural networks have resulted in significant computational and parameter overhead. In this work, we emphasize the importance of hidden dimensions in designing lightweight FFNs, a factor often overlooked in previous architectures. Guided by this principle, we introduce PartialFormer, a parameter-efficient Transformer architecture utilizing multiple smaller FFNs to reduce parameters and computation while maintaining essential hidden dimensions. These smaller FFNs are integrated into a multi-head attention mechanism for effective collaboration. We also propose a tailored head scaling strategy to enhance PartialFormer{'}s capabilities. Furthermore, we present a residual-like attention calculation to improve depth scaling within PartialFormer. Extensive experiments on 9 translation tasks and 1 abstractive summarization task validate the effectiveness of our PartialFormer approach on machine translation and summarization tasks. Our code would be available at: https://github.com/zhengkid/PartialFormer.",
}
| The design choices in Transformer feed-forward neural networks have resulted in significant computational and parameter overhead. In this work, we emphasize the importance of hidden dimensions in designing lightweight FFNs, a factor often overlooked in previous architectures. Guided by this principle, we introduce PartialFormer, a parameter-efficient Transformer architecture utilizing multiple smaller FFNs to reduce parameters and computation while maintaining essential hidden dimensions. These smaller FFNs are integrated into a multi-head attention mechanism for effective collaboration. We also propose a tailored head scaling strategy to enhance PartialFormer{'}s capabilities. Furthermore, we present a residual-like attention calculation to improve depth scaling within PartialFormer. Extensive experiments on 9 translation tasks and 1 abstractive summarization task validate the effectiveness of our PartialFormer approach on machine translation and summarization tasks. Our code would be available at: https://github.com/zhengkid/PartialFormer. | [
"Zheng, Tong",
"Li, Bei",
"Bao, Huiwen",
"Wang, Jiale",
"Shan, Weiqiao",
"Xiao, Tong",
"Zhu, JingBo"
]
| {P}artial{F}ormer: Modeling Part Instead of Whole for Machine Translation | findings-acl.434 | Poster | 2310.14921v2 |
https://aclanthology.org/2024.findings-acl.435.bib | @inproceedings{kim-etal-2024-self-consistent,
title = "Self-Consistent Reasoning-based Aspect-Sentiment Quad Prediction with Extract-Then-Assign Strategy",
author = "Kim, Jieyong and
Heo, Ryang and
Seo, Yongsik and
Kang, SeongKu and
Yeo, Jinyoung and
Lee, Dongha",
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.435",
pages = "7295--7303",
abstract = "In the task of aspect sentiment quad prediction (ASQP), generative methods for predicting sentiment quads have shown promisingresults. However, they still suffer from imprecise predictions and limited interpretability, caused by data scarcity and inadequate modeling of the quadruplet composition process. In this paper, we propose Self-Consistent Reasoning-based Aspect sentiment quadruple Prediction (SCRAP), optimizing its model to generate reasonings and the corresponding sentiment quadruplets in sequence. SCRAP adopts the Extract-Then-Assign reasoning strategy, which closely mimics human cognition. In the end, SCRAP significantly improves the model{'}s ability to handle complex reasoning tasks and correctly predict quadruplets through consistency voting, resulting in enhanced interpretability and accuracy in ASQP.",
}
| In the task of aspect sentiment quad prediction (ASQP), generative methods for predicting sentiment quads have shown promisingresults. However, they still suffer from imprecise predictions and limited interpretability, caused by data scarcity and inadequate modeling of the quadruplet composition process. In this paper, we propose Self-Consistent Reasoning-based Aspect sentiment quadruple Prediction (SCRAP), optimizing its model to generate reasonings and the corresponding sentiment quadruplets in sequence. SCRAP adopts the Extract-Then-Assign reasoning strategy, which closely mimics human cognition. In the end, SCRAP significantly improves the model{'}s ability to handle complex reasoning tasks and correctly predict quadruplets through consistency voting, resulting in enhanced interpretability and accuracy in ASQP. | [
"Kim, Jieyong",
"Heo, Ryang",
"Seo, Yongsik",
"Kang, SeongKu",
"Yeo, Jinyoung",
"Lee, Dongha"
]
| Self-Consistent Reasoning-based Aspect-Sentiment Quad Prediction with Extract-Then-Assign Strategy | findings-acl.435 | Poster | 2403.00354v2 |
https://aclanthology.org/2024.findings-acl.436.bib | @inproceedings{dong-etal-2024-pace,
title = "{PACE}: Improving Prompt with Actor-Critic Editing for Large Language Model",
author = "Dong, Yihong and
Luo, Kangcheng and
Jiang, Xue and
Jin, Zhi and
Li, Ge",
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.436",
pages = "7304--7323",
abstract = "Large language models (LLMs) have showcased remarkable potential across various tasks by conditioning on prompts. However, the quality of different human-written prompts leads to substantial discrepancies in LLMs{'} performance, and improving prompts usually necessitates considerable human effort and expertise. To this end, this paper proposes Prompt with Actor-Critic Editing (PACE) for LLMs to enable automatic prompt editing. Drawing inspiration from the actor-critic algorithm in reinforcement learning, PACE leverages LLMs as the dual roles of actors and critics, conceptualizing prompt as a type of policy. PACE refines prompt, taking into account the feedback from both actors performing prompt and critics criticizing response. This process helps LLMs better align prompt to a specific task, thanks to real responses and thinking from LLMs.We conduct extensive experiments on 24 instruction induction tasks and 21 big-bench tasks. Experimental results indicate that PACE elevates the relative performance of medium/low-quality human-written prompts by up to 98{\%}, which has comparable performance to high-quality human-written prompts. Moreover, PACE also exhibits notable efficacy for prompt generation.",
}
| Large language models (LLMs) have showcased remarkable potential across various tasks by conditioning on prompts. However, the quality of different human-written prompts leads to substantial discrepancies in LLMs{'} performance, and improving prompts usually necessitates considerable human effort and expertise. To this end, this paper proposes Prompt with Actor-Critic Editing (PACE) for LLMs to enable automatic prompt editing. Drawing inspiration from the actor-critic algorithm in reinforcement learning, PACE leverages LLMs as the dual roles of actors and critics, conceptualizing prompt as a type of policy. PACE refines prompt, taking into account the feedback from both actors performing prompt and critics criticizing response. This process helps LLMs better align prompt to a specific task, thanks to real responses and thinking from LLMs.We conduct extensive experiments on 24 instruction induction tasks and 21 big-bench tasks. Experimental results indicate that PACE elevates the relative performance of medium/low-quality human-written prompts by up to 98{\%}, which has comparable performance to high-quality human-written prompts. Moreover, PACE also exhibits notable efficacy for prompt generation. | [
"Dong, Yihong",
"Luo, Kangcheng",
"Jiang, Xue",
"Jin, Zhi",
"Li, Ge"
]
| {PACE}: Improving Prompt with Actor-Critic Editing for Large Language Model | findings-acl.436 | Poster | 2308.10088v2 |
https://aclanthology.org/2024.findings-acl.437.bib | @inproceedings{xu-etal-2024-penetrative,
title = "Penetrative {AI}: Making {LLM}s Comprehend the Physical World",
author = "Xu, Huatao and
Han, Liying and
Yang, Qirui and
Li, Mo and
Srivastava, Mani",
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.437",
pages = "7324--7341",
abstract = "Recent developments in Large Language Models (LLMs) have demonstrated their remarkable capabilities across a range of tasks. Questions, however, persist about the nature of LLMs and their potential to integrate common-sense human knowledge when performing tasks involving information about the real physical world. This paper delves into these questions by exploring how LLMs can be extended to interact with and reason about the physical world through IoT sensors and actuators, a concept that we term {``}Penetrative AI{''}. The paper explores such an extension at two levels of LLMs{'} ability to penetrate into the physical world via the processing of sensory signals. Our preliminary findings indicate that LLMs, with ChatGPT being the representative example in our exploration, have considerable and unique proficiency in employing the embedded world knowledge for interpreting IoT sensor data and reasoning over them about tasks in the physical realm. Not only this opens up new applications for LLMs beyond traditional text-based tasks, but also enables new ways of incorporating human knowledge in cyber-physical systems.",
}
| Recent developments in Large Language Models (LLMs) have demonstrated their remarkable capabilities across a range of tasks. Questions, however, persist about the nature of LLMs and their potential to integrate common-sense human knowledge when performing tasks involving information about the real physical world. This paper delves into these questions by exploring how LLMs can be extended to interact with and reason about the physical world through IoT sensors and actuators, a concept that we term {``}Penetrative AI{''}. The paper explores such an extension at two levels of LLMs{'} ability to penetrate into the physical world via the processing of sensory signals. Our preliminary findings indicate that LLMs, with ChatGPT being the representative example in our exploration, have considerable and unique proficiency in employing the embedded world knowledge for interpreting IoT sensor data and reasoning over them about tasks in the physical realm. Not only this opens up new applications for LLMs beyond traditional text-based tasks, but also enables new ways of incorporating human knowledge in cyber-physical systems. | [
"Xu, Huatao",
"Han, Liying",
"Yang, Qirui",
"Li, Mo",
"Srivastava, Mani"
]
| Penetrative {AI}: Making {LLM}s Comprehend the Physical World | findings-acl.437 | Poster | 2310.09605v3 |
https://aclanthology.org/2024.findings-acl.438.bib | @inproceedings{zhang-etal-2024-impact,
title = "The Impact of Demonstrations on Multilingual In-Context Learning: A Multidimensional Analysis",
author = "Zhang, Miaoran and
Gautam, Vagrant and
Wang, Mingyang and
Alabi, Jesujoba and
Shen, Xiaoyu and
Klakow, Dietrich and
Mosbach, Marius",
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.438",
pages = "7342--7371",
abstract = "In-context learning is a popular inference strategy where large language models solve a task using only a few labeled demonstrations without needing any parameter updates. Although there have been extensive studies on English in-context learning, multilingual in-context learning remains under-explored, and we lack an in-depth understanding of the role of demonstrations in this context. To address this gap, we conduct a multidimensional analysis of multilingual in-context learning, experimenting with 5 models from different model families, 9 datasets covering classification and generation tasks, and 56 typologically diverse languages. Our results reveal that the effectiveness of demonstrations varies significantly across models, tasks, and languages. We also find that strong instruction-following models including Llama 2-Chat, GPT-3.5, and GPT-4 are largely insensitive to the quality of demonstrations. Instead, a carefully crafted template often eliminates the benefits of demonstrations for some tasks and languages altogether. These findings show that the importance of demonstrations might be overestimated. Our work highlights the need for granular evaluation across multiple axes towards a better understanding of in-context learning.",
}
| In-context learning is a popular inference strategy where large language models solve a task using only a few labeled demonstrations without needing any parameter updates. Although there have been extensive studies on English in-context learning, multilingual in-context learning remains under-explored, and we lack an in-depth understanding of the role of demonstrations in this context. To address this gap, we conduct a multidimensional analysis of multilingual in-context learning, experimenting with 5 models from different model families, 9 datasets covering classification and generation tasks, and 56 typologically diverse languages. Our results reveal that the effectiveness of demonstrations varies significantly across models, tasks, and languages. We also find that strong instruction-following models including Llama 2-Chat, GPT-3.5, and GPT-4 are largely insensitive to the quality of demonstrations. Instead, a carefully crafted template often eliminates the benefits of demonstrations for some tasks and languages altogether. These findings show that the importance of demonstrations might be overestimated. Our work highlights the need for granular evaluation across multiple axes towards a better understanding of in-context learning. | [
"Zhang, Miaoran",
"Gautam, Vagrant",
"Wang, Mingyang",
"Alabi, Jesujoba",
"Shen, Xiaoyu",
"Klakow, Dietrich",
"Mosbach, Marius"
]
| The Impact of Demonstrations on Multilingual In-Context Learning: A Multidimensional Analysis | findings-acl.438 | Poster | 2402.12976v2 |
https://aclanthology.org/2024.findings-acl.439.bib | @inproceedings{dong-etal-2024-rich,
title = "Rich Semantic Knowledge Enhanced Large Language Models for Few-shot {C}hinese Spell Checking",
author = "Dong, Ming and
Chen, Yujing and
Zhang, Miao and
Sun, Hao and
He, Tingting",
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.439",
pages = "7372--7383",
abstract = "Chinese Spell Checking (CSC) is a widely used technology, which plays a vital role in speech to text (STT) and optical character recognition (OCR). Most of the existing CSC approaches relying on BERT architecture achieve excellent performance. However, limited by the scale of the foundation model, BERT-based method does not work well in few-shot scenarios, showing certain limitations in practical applications. In this paper, we explore using an in-context learning method named RS-LLM $(\textbf{R}ich\ \textbf{S}emantic\ based\ LLMs\)$ to introduce large language models (LLMs) as the foundation model. Besides, we study the impact of introducing various Chinese rich semantic information in our framework. We found that by introducing a small number of specific Chinese rich semantic structures, LLMs achieve better performance than most of the BERT-based model on few-shot CSC task. Furthermore, we conduct experiments on multiple datasets, and the experimental results verified the superiority of our proposed framework.",
}
| Chinese Spell Checking (CSC) is a widely used technology, which plays a vital role in speech to text (STT) and optical character recognition (OCR). Most of the existing CSC approaches relying on BERT architecture achieve excellent performance. However, limited by the scale of the foundation model, BERT-based method does not work well in few-shot scenarios, showing certain limitations in practical applications. In this paper, we explore using an in-context learning method named RS-LLM $(\textbf{R}ich\ \textbf{S}emantic\ based\ LLMs\)$ to introduce large language models (LLMs) as the foundation model. Besides, we study the impact of introducing various Chinese rich semantic information in our framework. We found that by introducing a small number of specific Chinese rich semantic structures, LLMs achieve better performance than most of the BERT-based model on few-shot CSC task. Furthermore, we conduct experiments on multiple datasets, and the experimental results verified the superiority of our proposed framework. | [
"Dong, Ming",
"Chen, Yujing",
"Zhang, Miao",
"Sun, Hao",
"He, Tingting"
]
| Rich Semantic Knowledge Enhanced Large Language Models for Few-shot {C}hinese Spell Checking | findings-acl.439 | Poster | 2403.08492v2 |
https://aclanthology.org/2024.findings-acl.440.bib | @inproceedings{chitale-etal-2024-empirical,
title = "An Empirical Study of In-context Learning in {LLM}s for Machine Translation",
author = "Chitale, Pranjal and
Gala, Jay and
Dabre, Raj",
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.440",
pages = "7384--7406",
abstract = "Recent interest has surged in employing Large Language Models (LLMs) for machine translation (MT) via in-context learning (ICL) (Vilar et al., 2023). Most prior studies primarily focus on optimizing translation quality, with limited attention to understanding the specific aspects of ICL that influence the said quality. To this end, we perform the first of its kind, exhaustive study of in-context learning for machine translation (MT). We first establish that ICL is primarily example-driven and not instruction-driven. Following this, we conduct an extensive exploration of various aspects of the examples to understand their influence on downstream performance. Our analysis includes factors such as quality and quantity of demonstrations, spatial proximity, and source versus target originality. Further, we also investigate challenging scenarios involving indirectness and misalignment of examples to understand the limits of ICL. While we establish the significance of the quality of the target distribution over the source distribution of demonstrations, we further observe that perturbations sometimes act as regularizers, resulting in performance improvements. Surprisingly, ICL does not necessitate examples from the same task, and a related task with the same target distribution proves sufficient. We hope that our study acts as a guiding resource for considerations in utilizing ICL for MT. Our code is available on https://github.com/PranjalChitale/in-context-mt-analysis.",
}
| Recent interest has surged in employing Large Language Models (LLMs) for machine translation (MT) via in-context learning (ICL) (Vilar et al., 2023). Most prior studies primarily focus on optimizing translation quality, with limited attention to understanding the specific aspects of ICL that influence the said quality. To this end, we perform the first of its kind, exhaustive study of in-context learning for machine translation (MT). We first establish that ICL is primarily example-driven and not instruction-driven. Following this, we conduct an extensive exploration of various aspects of the examples to understand their influence on downstream performance. Our analysis includes factors such as quality and quantity of demonstrations, spatial proximity, and source versus target originality. Further, we also investigate challenging scenarios involving indirectness and misalignment of examples to understand the limits of ICL. While we establish the significance of the quality of the target distribution over the source distribution of demonstrations, we further observe that perturbations sometimes act as regularizers, resulting in performance improvements. Surprisingly, ICL does not necessitate examples from the same task, and a related task with the same target distribution proves sufficient. We hope that our study acts as a guiding resource for considerations in utilizing ICL for MT. Our code is available on https://github.com/PranjalChitale/in-context-mt-analysis. | [
"Chitale, Pranjal",
"Gala, Jay",
"Dabre, Raj"
]
| An Empirical Study of In-context Learning in {LLM}s for Machine Translation | findings-acl.440 | Poster | 2402.10207v5 |
https://aclanthology.org/2024.findings-acl.441.bib | @inproceedings{wang-etal-2024-answer-c,
title = "{``}My Answer is {C}{''}: First-Token Probabilities Do Not Match Text Answers in Instruction-Tuned Language Models",
author = {Wang, Xinpeng and
Ma, Bolei and
Hu, Chengzhi and
Weber-Genzel, Leon and
R{\"o}ttger, Paul and
Kreuter, Frauke and
Hovy, Dirk and
Plank, Barbara},
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.441",
pages = "7407--7416",
abstract = "The open-ended nature of language generation makes the evaluation of autoregressive large language models (LLMs) challenging. One common evaluation approach uses multiple-choice questions to limit the response space. The model is then evaluated by ranking the candidate answers by the log probability of the first token prediction. However, first-tokens may not consistently reflect the final response output, due to model{'}s diverse response styles such as starting with {``}Sure{''} or refusing to answer. Consequently, first-token evaluation is not indicative of model behaviour when interacting with users. But by how much? We evaluate how aligned first-token evaluation is with the text output along several dimensions, namely final option choice, refusal rate, choice distribution and robustness under prompt perturbation. Our results show that the two approaches are severely misaligned \textit{on all dimensions}, reaching mismatch rates over 60{\%}. Models heavily fine-tuned on conversational or safety data are especially impacted. Crucially, models remain misaligned even when we increasingly constrain prompts, i.e., force them to start with an option letter or example template. Our findings i) underscore the importance of inspecting the text output as well and ii) caution against relying solely on first-token evaluation.",
}
| The open-ended nature of language generation makes the evaluation of autoregressive large language models (LLMs) challenging. One common evaluation approach uses multiple-choice questions to limit the response space. The model is then evaluated by ranking the candidate answers by the log probability of the first token prediction. However, first-tokens may not consistently reflect the final response output, due to model{'}s diverse response styles such as starting with {``}Sure{''} or refusing to answer. Consequently, first-token evaluation is not indicative of model behaviour when interacting with users. But by how much? We evaluate how aligned first-token evaluation is with the text output along several dimensions, namely final option choice, refusal rate, choice distribution and robustness under prompt perturbation. Our results show that the two approaches are severely misaligned \textit{on all dimensions}, reaching mismatch rates over 60{\%}. Models heavily fine-tuned on conversational or safety data are especially impacted. Crucially, models remain misaligned even when we increasingly constrain prompts, i.e., force them to start with an option letter or example template. Our findings i) underscore the importance of inspecting the text output as well and ii) caution against relying solely on first-token evaluation. | [
"Wang, Xinpeng",
"Ma, Bolei",
"Hu, Chengzhi",
"Weber-Genzel, Leon",
"R{\\\"o}ttger, Paul",
"Kreuter, Frauke",
"Hovy, Dirk",
"Plank, Barbara"
]
| {``}My Answer is {C}{''}: First-Token Probabilities Do Not Match Text Answers in Instruction-Tuned Language Models | findings-acl.441 | Poster | 2404.08382v1 |
https://aclanthology.org/2024.findings-acl.442.bib | @inproceedings{sun-etal-2024-oda,
title = "{ODA}: Observation-Driven Agent for integrating {LLM}s and Knowledge Graphs",
author = "Sun, Lei and
Tao, Zhengwei and
Li, Youdi and
Arakawa, Hiroshi",
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.442",
pages = "7417--7431",
abstract = "The integration of Large Language Models (LLMs) and knowledge graphs (KGs) has achieved remarkable success in various natural language processing tasks. However, existing methodologies that integrate LLMs and KGs often navigate the task-solving process solely based on the LLM{'}s analysis of the question, overlooking the rich cognitive potential inherent in the vast knowledge encapsulated in KGs. To address this, we introduce Observation-Driven Agent (ODA), a novel AI agent framework tailored for tasks involving KGs. ODA incorporates KG reasoning abilities via global observation, which enhances reasoning capabilities through a cyclical paradigm of observation, action, and reflection. Confronting the exponential explosion of knowledge during observation, we innovatively design a recursive observation mechanism. Subsequently, we integrate the observed knowledge into the action and reflection modules. Through extensive experiments, ODA demonstrates state-of-the-art performance on several datasets, notably achieving accuracy improvements of 12.87{\%} and 8.9{\%}.",
}
| The integration of Large Language Models (LLMs) and knowledge graphs (KGs) has achieved remarkable success in various natural language processing tasks. However, existing methodologies that integrate LLMs and KGs often navigate the task-solving process solely based on the LLM{'}s analysis of the question, overlooking the rich cognitive potential inherent in the vast knowledge encapsulated in KGs. To address this, we introduce Observation-Driven Agent (ODA), a novel AI agent framework tailored for tasks involving KGs. ODA incorporates KG reasoning abilities via global observation, which enhances reasoning capabilities through a cyclical paradigm of observation, action, and reflection. Confronting the exponential explosion of knowledge during observation, we innovatively design a recursive observation mechanism. Subsequently, we integrate the observed knowledge into the action and reflection modules. Through extensive experiments, ODA demonstrates state-of-the-art performance on several datasets, notably achieving accuracy improvements of 12.87{\%} and 8.9{\%}. | [
"Sun, Lei",
"Tao, Zhengwei",
"Li, Youdi",
"Arakawa, Hiroshi"
]
| {ODA}: Observation-Driven Agent for integrating {LLM}s and Knowledge Graphs | findings-acl.442 | Poster | 2402.11163v1 |
https://aclanthology.org/2024.findings-acl.443.bib | @inproceedings{xu-etal-2024-comprehensive,
title = "A Comprehensive Study of Jailbreak Attack versus Defense for Large Language Models",
author = "Xu, Zihao and
Liu, Yi and
Deng, Gelei and
Li, Yuekang and
Picek, Stjepan",
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.443",
pages = "7432--7449",
abstract = "Large Language Models (LLMs) have increasingly become central to generating content with potential societal impacts. Notably, these models have demonstrated capabilities for generating content that could be deemed harmful. To mitigate these risks, researchers have adopted safety training techniques to align model outputs with societal values to curb the generation of malicious content. However, the phenomenon of {``}jailbreaking{''} {---} where carefully crafted prompts elicit harmful responses from models {---} persists as a significant challenge. This research conducts a comprehensive analysis of existing studies on jailbreaking LLMs and their defense techniques. We meticulously investigate nine attack techniques and seven defense techniques applied across three distinct language models: Vicuna, LLama, and GPT-3.5 Turbo. We aim to evaluate the effectiveness of these attack and defense techniques. Our findings reveal that existing white-box attacks underperform compared to universal techniques and that including special tokens in the input significantly affects the likelihood of successful attacks. This research highlights the need to concentrate on the security facets of LLMs. Additionally, we contribute to the field by releasing our datasets and testing framework, aiming to foster further research into LLM security. We believe these contributions will facilitate the exploration of security measures within this domain.",
}
| Large Language Models (LLMs) have increasingly become central to generating content with potential societal impacts. Notably, these models have demonstrated capabilities for generating content that could be deemed harmful. To mitigate these risks, researchers have adopted safety training techniques to align model outputs with societal values to curb the generation of malicious content. However, the phenomenon of {``}jailbreaking{''} {---} where carefully crafted prompts elicit harmful responses from models {---} persists as a significant challenge. This research conducts a comprehensive analysis of existing studies on jailbreaking LLMs and their defense techniques. We meticulously investigate nine attack techniques and seven defense techniques applied across three distinct language models: Vicuna, LLama, and GPT-3.5 Turbo. We aim to evaluate the effectiveness of these attack and defense techniques. Our findings reveal that existing white-box attacks underperform compared to universal techniques and that including special tokens in the input significantly affects the likelihood of successful attacks. This research highlights the need to concentrate on the security facets of LLMs. Additionally, we contribute to the field by releasing our datasets and testing framework, aiming to foster further research into LLM security. We believe these contributions will facilitate the exploration of security measures within this domain. | [
"Xu, Zihao",
"Liu, Yi",
"Deng, Gelei",
"Li, Yuekang",
"Picek, Stjepan"
]
| A Comprehensive Study of Jailbreak Attack versus Defense for Large Language Models | findings-acl.443 | Poster | 2401.16765v1 |
https://aclanthology.org/2024.findings-acl.444.bib | @inproceedings{kaliosis-etal-2024-data,
title = "A Data-Driven Guided Decoding Mechanism for Diagnostic Captioning",
author = "Kaliosis, Panagiotis and
Pavlopoulos, John and
Charalampakos, Foivos and
Moschovis, Georgios and
Androutsopoulos, Ion",
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.444",
pages = "7450--7466",
abstract = "Diagnostic Captioning (DC) automatically generates a diagnostic text from one or more medical images (e.g., X-rays, MRIs) of a patient. Treated as a draft, the generated text may assist clinicians, by providing an initial estimation of the patient{'}s condition, speeding up and helping safeguard the diagnostic process. The accuracy of a diagnostic text, however, strongly depends on how well the key medical conditions depicted in the images are expressed. We propose a new $\textit{data-driven}$ guided decoding method that incorporates medical information, in the form of existing tags capturing key conditions of the image(s), into the beam search of the diagnostic text generation process. We evaluate the proposed method on two medical datasets using four DC systems that range from generic image-to-text systems with CNN encoders and RNN decoders to pre-trained Large Language Models. The latter can also be used in few- and zero-shot learning scenarios. In most cases, the proposed mechanism improves performance with respect to all evaluation measures. We provide an open-source implementation of the proposed method at https://github.com/nlpaueb/dmmcs.",
}
| Diagnostic Captioning (DC) automatically generates a diagnostic text from one or more medical images (e.g., X-rays, MRIs) of a patient. Treated as a draft, the generated text may assist clinicians, by providing an initial estimation of the patient{'}s condition, speeding up and helping safeguard the diagnostic process. The accuracy of a diagnostic text, however, strongly depends on how well the key medical conditions depicted in the images are expressed. We propose a new $\textit{data-driven}$ guided decoding method that incorporates medical information, in the form of existing tags capturing key conditions of the image(s), into the beam search of the diagnostic text generation process. We evaluate the proposed method on two medical datasets using four DC systems that range from generic image-to-text systems with CNN encoders and RNN decoders to pre-trained Large Language Models. The latter can also be used in few- and zero-shot learning scenarios. In most cases, the proposed mechanism improves performance with respect to all evaluation measures. We provide an open-source implementation of the proposed method at https://github.com/nlpaueb/dmmcs. | [
"Kaliosis, Panagiotis",
"Pavlopoulos, John",
"Charalampakos, Foivos",
"Moschovis, Georgios",
"Androutsopoulos, Ion"
]
| A Data-Driven Guided Decoding Mechanism for Diagnostic Captioning | findings-acl.444 | Poster | 2406.14164v1 |
https://aclanthology.org/2024.findings-acl.445.bib | @inproceedings{zhang-etal-2024-balancing,
title = "Balancing Speciality and Versatility: a Coarse to Fine Framework for Supervised Fine-tuning Large Language Model",
author = "Zhang, Hengyuan and
Wu, Yanru and
Li, Dawei and
Yang, Sak and
Zhao, Rui and
Jiang, Yong and
Tan, 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.445",
pages = "7467--7509",
abstract = "Aligned Large Language Models (LLMs) showcase remarkable versatility, capable of handling diverse real-world tasks. Meanwhile, aligned LLMs are also expected to exhibit speciality, excelling in specific applications. However, fine-tuning with extra data, a common practice to gain speciality, often leads to catastrophic forgetting (CF) of previously acquired versatility, hindering the model{'}s performance across diverse tasks. In response to this challenge, we propose CoFiTune, a coarse to fine framework in an attempt to strike the balance between speciality and versatility. At the coarse-grained level, an empirical tree-search algorithm is utilized to pinpoint and update specific modules that are crucial for speciality, while keeping other parameters frozen; at the fine-grained level, a soft-masking mechanism regulates the update to the LLMs, mitigating the CF issue without harming speciality. In an overall evaluation of both speciality and versatility, CoFiTune consistently outperforms baseline methods across diverse tasks and model scales. Compared to the full-parameter SFT, CoFiTune leads to about 14{\%} versatility improvement and marginal speciality loss on a 13B model. Lastly, based on further analysis, we provide a speculative insight into the information forwarding process in LLMs, which helps explain the effectiveness of the proposed method. The code is available at https://github.com/rattlesnakey/CoFiTune.",
}
| Aligned Large Language Models (LLMs) showcase remarkable versatility, capable of handling diverse real-world tasks. Meanwhile, aligned LLMs are also expected to exhibit speciality, excelling in specific applications. However, fine-tuning with extra data, a common practice to gain speciality, often leads to catastrophic forgetting (CF) of previously acquired versatility, hindering the model{'}s performance across diverse tasks. In response to this challenge, we propose CoFiTune, a coarse to fine framework in an attempt to strike the balance between speciality and versatility. At the coarse-grained level, an empirical tree-search algorithm is utilized to pinpoint and update specific modules that are crucial for speciality, while keeping other parameters frozen; at the fine-grained level, a soft-masking mechanism regulates the update to the LLMs, mitigating the CF issue without harming speciality. In an overall evaluation of both speciality and versatility, CoFiTune consistently outperforms baseline methods across diverse tasks and model scales. Compared to the full-parameter SFT, CoFiTune leads to about 14{\%} versatility improvement and marginal speciality loss on a 13B model. Lastly, based on further analysis, we provide a speculative insight into the information forwarding process in LLMs, which helps explain the effectiveness of the proposed method. The code is available at https://github.com/rattlesnakey/CoFiTune. | [
"Zhang, Hengyuan",
"Wu, Yanru",
"Li, Dawei",
"Yang, Sak",
"Zhao, Rui",
"Jiang, Yong",
"Tan, Fei"
]
| Balancing Speciality and Versatility: a Coarse to Fine Framework for Supervised Fine-tuning Large Language Model | findings-acl.445 | Poster | 2404.10306v5 |
https://aclanthology.org/2024.findings-acl.446.bib | @inproceedings{xu-etal-2024-two,
title = "A Two-Agent Game for Zero-shot Relation Triplet Extraction",
author = "Xu, Ting and
Yang, Haiqin and
Zhao, Fei and
Wu, Zhen and
Dai, Xinyu",
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.446",
pages = "7510--7527",
abstract = "Relation triplet extraction is a fundamental task in natural language processing that aims to identify semantic relationships between entities in text. It is particularly challenging in the zero-shot setting, i.e., zero-shot relation triplet extraction (ZeroRTE), where the relation sets between training and test are disjoint. Existing methods deal with this task by integrating relations into prompts, which may lack sufficient understanding of the unseen relations. To address these limitations, this paper presents a novel Two-Agent Game (TAG) approach to deliberate and debate the semantics of unseen relations. TAG consists of two agents, a generator and an extractor. They iteratively interact in three key steps: attempting, criticizing, and rectifying. This enables the agents to fully debate and understand the unseen relations. Experimental results demonstrate consistent improvement over ALBERT-Large, BART, andGPT3.5, without incurring additional inference costs in all cases. Remarkably, our method outperforms strong baselines by a significant margin, achieving an impressive 6{\%}-16{\%} increase in F1 scores, particularly when dealingwith FewRel with five unseen relations.",
}
| Relation triplet extraction is a fundamental task in natural language processing that aims to identify semantic relationships between entities in text. It is particularly challenging in the zero-shot setting, i.e., zero-shot relation triplet extraction (ZeroRTE), where the relation sets between training and test are disjoint. Existing methods deal with this task by integrating relations into prompts, which may lack sufficient understanding of the unseen relations. To address these limitations, this paper presents a novel Two-Agent Game (TAG) approach to deliberate and debate the semantics of unseen relations. TAG consists of two agents, a generator and an extractor. They iteratively interact in three key steps: attempting, criticizing, and rectifying. This enables the agents to fully debate and understand the unseen relations. Experimental results demonstrate consistent improvement over ALBERT-Large, BART, andGPT3.5, without incurring additional inference costs in all cases. Remarkably, our method outperforms strong baselines by a significant margin, achieving an impressive 6{\%}-16{\%} increase in F1 scores, particularly when dealingwith FewRel with five unseen relations. | [
"Xu, Ting",
"Yang, Haiqin",
"Zhao, Fei",
"Wu, Zhen",
"Dai, Xinyu"
]
| A Two-Agent Game for Zero-shot Relation Triplet Extraction | findings-acl.446 | Poster | 2010.02609v3 |
https://aclanthology.org/2024.findings-acl.447.bib | @inproceedings{gu-etal-2024-light,
title = "Light-{PEFT}: Lightening Parameter-Efficient Fine-Tuning via Early Pruning",
author = "Gu, Naibin and
Fu, Peng and
Liu, Xiyu and
Shen, Bowen and
Lin, Zheng and
Wang, Weiping",
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.447",
pages = "7528--7541",
abstract = "Parameter-efficient fine-tuning (PEFT) has emerged as the predominant technique for fine-tuning in the era of large language models. However, existing PEFT methods still have inadequate training efficiency. Firstly, the utilization of large-scale foundation models during the training process is excessively redundant for certain fine-tuning tasks. Secondly, as the model size increases, the growth in trainable parameters of empirically added PEFT modules becomes non-negligible and redundant, leading to inefficiency. To achieve task-specific efficient fine-tuning, we propose the Light-PEFT framework, which includes two methods: Masked Early Pruning of the Foundation Model and Multi-Granularity Early Pruning of PEFT. The Light-PEFT framework allows for the simultaneous estimation of redundant parameters in both the foundation model and PEFT modules during the early stage of training. These parameters can then be pruned for more efficient fine-tuning. We validate our approach on GLUE, SuperGLUE, QA tasks, and various models. With Light-PEFT, parameters of the foundation model can be pruned by up to over 40{\%}, while still controlling trainable parameters to be only 25{\%} of the original PEFT method. Compared to utilizing the PEFT method directly, Light-PEFT achieves training and inference speedup, reduces memory usage, and maintains comparable performance and the plug-and-play feature of PEFT.",
}
| Parameter-efficient fine-tuning (PEFT) has emerged as the predominant technique for fine-tuning in the era of large language models. However, existing PEFT methods still have inadequate training efficiency. Firstly, the utilization of large-scale foundation models during the training process is excessively redundant for certain fine-tuning tasks. Secondly, as the model size increases, the growth in trainable parameters of empirically added PEFT modules becomes non-negligible and redundant, leading to inefficiency. To achieve task-specific efficient fine-tuning, we propose the Light-PEFT framework, which includes two methods: Masked Early Pruning of the Foundation Model and Multi-Granularity Early Pruning of PEFT. The Light-PEFT framework allows for the simultaneous estimation of redundant parameters in both the foundation model and PEFT modules during the early stage of training. These parameters can then be pruned for more efficient fine-tuning. We validate our approach on GLUE, SuperGLUE, QA tasks, and various models. With Light-PEFT, parameters of the foundation model can be pruned by up to over 40{\%}, while still controlling trainable parameters to be only 25{\%} of the original PEFT method. Compared to utilizing the PEFT method directly, Light-PEFT achieves training and inference speedup, reduces memory usage, and maintains comparable performance and the plug-and-play feature of PEFT. | [
"Gu, Naibin",
"Fu, Peng",
"Liu, Xiyu",
"Shen, Bowen",
"Lin, Zheng",
"Wang, Weiping"
]
| Light-{PEFT}: Lightening Parameter-Efficient Fine-Tuning via Early Pruning | findings-acl.447 | Poster | 2110.12007v1 |
https://aclanthology.org/2024.findings-acl.448.bib | @inproceedings{lardelli-etal-2024-building,
title = "Building Bridges: A Dataset for Evaluating Gender-Fair Machine Translation into {G}erman",
author = "Lardelli, Manuel and
Attanasio, Giuseppe and
Lauscher, Anne",
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.448",
pages = "7542--7550",
abstract = "The translation of gender-neutral person-referring terms (e.g.,the students) is often non-trivial.Translating from English into German poses an interesting case{---}in German, person-referring nouns are usually gender-specific, and if the gender of the referent(s) is unknown or diverse, the generic masculine (die Studenten (m.)) is commonly used. This solution, however, reduces the visibility of other genders, such as women and non-binary people. To counteract gender discrimination, a societal movement towards using gender-fair language exists (e.g., by adopting neosystems). However, gender-fair German is currently barely supported in machine translation (MT), requiring post-editing or manual translations. We address this research gap by studying gender-fair language in English-to-German MT. Concretely, we enrich a community-created gender-fair language dictionary and sample multi-sentence test instances from encyclopedic text and parliamentary speeches.Using these novel resources, we conduct the first benchmark study involving two commercial systems and six neural MT models for translating words in isolation and natural contexts across two domains. Our findings show that most systems produce mainly masculine forms and rarely gender-neutral variants, highlighting the need for future research. We release code and data at https://github.com/g8a9/building-bridges-gender-fair-german-mt.",
}
| The translation of gender-neutral person-referring terms (e.g.,the students) is often non-trivial.Translating from English into German poses an interesting case{---}in German, person-referring nouns are usually gender-specific, and if the gender of the referent(s) is unknown or diverse, the generic masculine (die Studenten (m.)) is commonly used. This solution, however, reduces the visibility of other genders, such as women and non-binary people. To counteract gender discrimination, a societal movement towards using gender-fair language exists (e.g., by adopting neosystems). However, gender-fair German is currently barely supported in machine translation (MT), requiring post-editing or manual translations. We address this research gap by studying gender-fair language in English-to-German MT. Concretely, we enrich a community-created gender-fair language dictionary and sample multi-sentence test instances from encyclopedic text and parliamentary speeches.Using these novel resources, we conduct the first benchmark study involving two commercial systems and six neural MT models for translating words in isolation and natural contexts across two domains. Our findings show that most systems produce mainly masculine forms and rarely gender-neutral variants, highlighting the need for future research. We release code and data at https://github.com/g8a9/building-bridges-gender-fair-german-mt. | [
"Lardelli, Manuel",
"Attanasio, Giuseppe",
"Lauscher, Anne"
]
| Building Bridges: A Dataset for Evaluating Gender-Fair Machine Translation into {G}erman | findings-acl.448 | Poster | 2406.06131v1 |
https://aclanthology.org/2024.findings-acl.449.bib | @inproceedings{sun-etal-2024-prompt,
title = "Prompt Chaining or Stepwise Prompt? Refinement in Text Summarization",
author = "Sun, Shichao and
Yuan, Ruifeng and
Cao, Ziqiang 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.449",
pages = "7551--7558",
abstract = "Large language models (LLMs) have demonstrated the capacity to improve summary quality by mirroring a human-like iterative process of critique and refinement starting from the initial draft. Two strategies are designed to perform this iterative process: $\textit{Prompt Chaining}$ and $\textit{Stepwise Prompt}$. Prompt chaining orchestrates the drafting, critiquing, and refining phases through a series of three discrete prompts, while Stepwise prompt integrates these phases within a single prompt. However, the relative effectiveness of the two methods has not been extensively studied. This paper is dedicated to examining and comparing these two methods in the context of text summarization to ascertain which method stands out as the most effective. Experimental results show that the prompt chaining method can produce a more favorable outcome. This might be because stepwise prompt might produce a simulated refinement process according to our various experiments. Since refinement is adaptable to diverse tasks, our conclusions have the potential to be extrapolated to other applications, thereby offering insights that may contribute to the broader development of LLMs.",
}
| Large language models (LLMs) have demonstrated the capacity to improve summary quality by mirroring a human-like iterative process of critique and refinement starting from the initial draft. Two strategies are designed to perform this iterative process: $\textit{Prompt Chaining}$ and $\textit{Stepwise Prompt}$. Prompt chaining orchestrates the drafting, critiquing, and refining phases through a series of three discrete prompts, while Stepwise prompt integrates these phases within a single prompt. However, the relative effectiveness of the two methods has not been extensively studied. This paper is dedicated to examining and comparing these two methods in the context of text summarization to ascertain which method stands out as the most effective. Experimental results show that the prompt chaining method can produce a more favorable outcome. This might be because stepwise prompt might produce a simulated refinement process according to our various experiments. Since refinement is adaptable to diverse tasks, our conclusions have the potential to be extrapolated to other applications, thereby offering insights that may contribute to the broader development of LLMs. | [
"Sun, Shichao",
"Yuan, Ruifeng",
"Cao, Ziqiang",
"Li, Wenjie",
"Liu, Pengfei"
]
| Prompt Chaining or Stepwise Prompt? Refinement in Text Summarization | findings-acl.449 | Poster | 2406.00507v1 |
https://aclanthology.org/2024.findings-acl.450.bib | @inproceedings{long-etal-2024-trust,
title = "Trust in Internal or External Knowledge? Generative Multi-Modal Entity Linking with Knowledge Retriever",
author = "Long, Xinwei and
Zeng, Jiali and
Meng, Fandong and
Zhou, Jie 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.450",
pages = "7559--7569",
abstract = "Multi-modal entity linking (MEL) is a challenging task that requires accurate prediction of entities within extensive search spaces, utilizing multi-modal contexts. Existing generative approaches struggle with the knowledge gap between visual entity information and the intrinsic parametric knowledge of LLMs. To address this knowledge gap, we introduce a novel approach called GELR, which incorporates a knowledge retriever to enhance visual entity information by leveraging external sources. Additionally, we devise a prioritization scheme that effectively handles noisy retrieval results and manages conflicts arising from the integration of external and internal knowledge. Moreover, we propose a noise-aware instruction tuning technique during training to finely adjust the model{'}s ability to leverage retrieved information effectively. Through extensive experiments conducted on three benchmarks, our approach showcases remarkable improvements, ranging from 3.0{\%} to 6.5{\%}, across all evaluation metrics compared to strong baselines. These results demonstrate the effectiveness and superiority of our proposed method in tackling the complexities of multi-modal entity linking.",
}
| Multi-modal entity linking (MEL) is a challenging task that requires accurate prediction of entities within extensive search spaces, utilizing multi-modal contexts. Existing generative approaches struggle with the knowledge gap between visual entity information and the intrinsic parametric knowledge of LLMs. To address this knowledge gap, we introduce a novel approach called GELR, which incorporates a knowledge retriever to enhance visual entity information by leveraging external sources. Additionally, we devise a prioritization scheme that effectively handles noisy retrieval results and manages conflicts arising from the integration of external and internal knowledge. Moreover, we propose a noise-aware instruction tuning technique during training to finely adjust the model{'}s ability to leverage retrieved information effectively. Through extensive experiments conducted on three benchmarks, our approach showcases remarkable improvements, ranging from 3.0{\%} to 6.5{\%}, across all evaluation metrics compared to strong baselines. These results demonstrate the effectiveness and superiority of our proposed method in tackling the complexities of multi-modal entity linking. | [
"Long, Xinwei",
"Zeng, Jiali",
"Meng, F",
"ong",
"Zhou, Jie",
"Zhou, Bowen"
]
| Trust in Internal or External Knowledge? Generative Multi-Modal Entity Linking with Knowledge Retriever | findings-acl.450 | Poster | 1810.10004v1 |
https://aclanthology.org/2024.findings-acl.451.bib | @inproceedings{aida-bollegala-2024-semantic,
title = "A Semantic Distance Metric Learning approach for Lexical Semantic Change Detection",
author = "Aida, Taichi and
Bollegala, Danushka",
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.451",
pages = "7570--7584",
abstract = "Detecting temporal semantic changes of words is an important task for various NLP applications that must make time-sensitive predictions.Lexical Semantic Change Detection (SCD) task involves predicting whether a given target word, $w$, changes its meaning between two different text corpora, $C_1$ and $C_2$.For this purpose, we propose a supervised two-staged SCD method that uses existing Word-in-Context (WiC) datasets.In the first stage, for a target word $w$, we learn two sense-aware encoders that represent the meaning of $w$ in a given sentence selected from a corpus.Next, in the second stage, we learn a sense-aware distance metric that compares the semantic representations of a target word across all of its occurrences in $C_1$ and $C_2$.Experimental results on multiple benchmark datasets for SCD show that our proposed method achieves strong performance in multiple languages.Additionally, our method achieves significant improvements on WiC benchmarks compared to a sense-aware encoder with conventional distance functions.",
}
| Detecting temporal semantic changes of words is an important task for various NLP applications that must make time-sensitive predictions.Lexical Semantic Change Detection (SCD) task involves predicting whether a given target word, $w$, changes its meaning between two different text corpora, $C_1$ and $C_2$.For this purpose, we propose a supervised two-staged SCD method that uses existing Word-in-Context (WiC) datasets.In the first stage, for a target word $w$, we learn two sense-aware encoders that represent the meaning of $w$ in a given sentence selected from a corpus.Next, in the second stage, we learn a sense-aware distance metric that compares the semantic representations of a target word across all of its occurrences in $C_1$ and $C_2$.Experimental results on multiple benchmark datasets for SCD show that our proposed method achieves strong performance in multiple languages.Additionally, our method achieves significant improvements on WiC benchmarks compared to a sense-aware encoder with conventional distance functions. | [
"Aida, Taichi",
"Bollegala, Danushka"
]
| A Semantic Distance Metric Learning approach for Lexical Semantic Change Detection | findings-acl.451 | Poster | 2403.00226v3 |
https://aclanthology.org/2024.findings-acl.452.bib | @inproceedings{li-etal-2024-achieved,
title = "What Have We Achieved on Non-autoregressive Translation?",
author = "Li, Yafu and
Zhang, Huajian and
Yan, Jianhao and
Yin, Yongjing and
Zhang, Yue",
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.452",
pages = "7585--7606",
abstract = "Recent advances have made non-autoregressive (NAT) translation comparable to autoregressive methods (AT). However, their evaluation using BLEU has been shown to weakly correlate with human annotations. Limited research compares non-autoregressive translation and autoregressive translation comprehensively, leaving uncertainty about the true proximity of NAT to AT. To address this gap, we systematically evaluate four representative NAT methods across various dimensions, including human evaluation. Our empirical results demonstrate that despite narrowing the performance gap, state-of-the-art NAT still underperforms AT under more reliable evaluation metrics. Furthermore, we discover that explicitly modeling dependencies is crucial for generating natural language and generalizing to out-of-distribution sequences.",
}
| Recent advances have made non-autoregressive (NAT) translation comparable to autoregressive methods (AT). However, their evaluation using BLEU has been shown to weakly correlate with human annotations. Limited research compares non-autoregressive translation and autoregressive translation comprehensively, leaving uncertainty about the true proximity of NAT to AT. To address this gap, we systematically evaluate four representative NAT methods across various dimensions, including human evaluation. Our empirical results demonstrate that despite narrowing the performance gap, state-of-the-art NAT still underperforms AT under more reliable evaluation metrics. Furthermore, we discover that explicitly modeling dependencies is crucial for generating natural language and generalizing to out-of-distribution sequences. | [
"Li, Yafu",
"Zhang, Huajian",
"Yan, Jianhao",
"Yin, Yongjing",
"Zhang, Yue"
]
| What Have We Achieved on Non-autoregressive Translation? | findings-acl.452 | Poster | 2406.14267v1 |
https://aclanthology.org/2024.findings-acl.453.bib | @inproceedings{reiss-etal-2024-zero,
title = "From Zero to Hero: Cold-Start Anomaly Detection",
author = "Reiss, Tal and
Kour, George and
Zwerdling, Naama and
Anaby Tavor, Ateret and
Hoshen, Yedid",
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.453",
pages = "7607--7617",
abstract = "When first deploying an anomaly detection system, e.g., to detect out-of-scope queries in chatbots, there are no observed data, making data-driven approaches ineffective. Zero-shot anomaly detection methods offer a solution to such {``}cold-start{''} cases, but unfortunately they are often not accurate enough. This paper studies the realistic but underexplored $\textit{cold-start}$ setting where an anomaly detection model is initialized using zero-shot guidance, but subsequently receives a small number of contaminated observations (namely, that may include anomalies). The goal is to make efficient use of both the zero-shot guidance and the observations. We propose ColdFusion, a method that effectively adapts the zero-shot anomaly detector to contaminated observations. To support future development of this new setting, we propose an evaluation suite consisting of evaluation protocols and metrics.",
}
| When first deploying an anomaly detection system, e.g., to detect out-of-scope queries in chatbots, there are no observed data, making data-driven approaches ineffective. Zero-shot anomaly detection methods offer a solution to such {``}cold-start{''} cases, but unfortunately they are often not accurate enough. This paper studies the realistic but underexplored $\textit{cold-start}$ setting where an anomaly detection model is initialized using zero-shot guidance, but subsequently receives a small number of contaminated observations (namely, that may include anomalies). The goal is to make efficient use of both the zero-shot guidance and the observations. We propose ColdFusion, a method that effectively adapts the zero-shot anomaly detector to contaminated observations. To support future development of this new setting, we propose an evaluation suite consisting of evaluation protocols and metrics. | [
"Reiss, Tal",
"Kour, George",
"Zwerdling, Naama",
"Anaby Tavor, Ateret",
"Hoshen, Yedid"
]
| From Zero to Hero: Cold-Start Anomaly Detection | findings-acl.453 | Poster | 2306.09067v2 |
https://aclanthology.org/2024.findings-acl.454.bib | @inproceedings{zhao-etal-2024-large,
title = "Large Language Models Fall Short: Understanding Complex Relationships in Detective Narratives",
author = "Zhao, Runcong and
Zhu, Qinglin and
Xu, Hainiu and
Li, Jiazheng and
Zhou, Yuxiang and
He, Yulan and
Gui, Lin",
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.454",
pages = "7618--7638",
abstract = "Existing datasets for narrative understanding often fail to represent the complexity and uncertainty of relationships in real-life social scenarios. To address this gap, we introduce a new benchmark, Conan, designed for extracting and analysing intricate character relation graphs from detective narratives. Specifically, we designed hierarchical relationship categories and manually extracted and annotated role-oriented relationships from the perspectives of various characters, incorporating both public relationships known to most characters and secret ones known to only a few. Our experiments with advanced Large Language Models (LLMs) like GPT-3.5, GPT-4, and Llama2 reveal their limitations in inferencing complex relationships and handling longer narratives. The combination of the Conan dataset and our pipeline strategy is geared towards understanding the ability of LLMs to comprehend nuanced relational dynamics in narrative contexts.",
}
| Existing datasets for narrative understanding often fail to represent the complexity and uncertainty of relationships in real-life social scenarios. To address this gap, we introduce a new benchmark, Conan, designed for extracting and analysing intricate character relation graphs from detective narratives. Specifically, we designed hierarchical relationship categories and manually extracted and annotated role-oriented relationships from the perspectives of various characters, incorporating both public relationships known to most characters and secret ones known to only a few. Our experiments with advanced Large Language Models (LLMs) like GPT-3.5, GPT-4, and Llama2 reveal their limitations in inferencing complex relationships and handling longer narratives. The combination of the Conan dataset and our pipeline strategy is geared towards understanding the ability of LLMs to comprehend nuanced relational dynamics in narrative contexts. | [
"Zhao, Runcong",
"Zhu, Qinglin",
"Xu, Hainiu",
"Li, Jiazheng",
"Zhou, Yuxiang",
"He, Yulan",
"Gui, Lin"
]
| Large Language Models Fall Short: Understanding Complex Relationships in Detective Narratives | findings-acl.454 | Poster | 2402.11051v1 |
https://aclanthology.org/2024.findings-acl.455.bib | @inproceedings{qiao-etal-2024-distillmike,
title = "{D}istill{MIKE}: Editing Distillation of Massive In-Context Knowledge Editing in Large Language Models",
author = "Qiao, Shanbao and
Liu, Xuebing and
Na, Seung-Hoon",
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.455",
pages = "7639--7654",
abstract = "Among the recently emerged knowledge editing methods, in-context knowledge editing (IKE) has shown respectable abilities on knowledge editing in terms of generalization and specificity. Noting the promising advantages but unexplored issues of IKE, we propose **DistillMIKE** as a novel extension of IKE, i.e., editing **distill**ation of ''**M**assive{''} **I**n-context **K**nowledge **E**diting in large language models (LLMs), mainly consisting of two expansions; 1) *Massive in-context knowledge editing (MIKE)*, which extends IKE to a massive editing task, aiming to inject not a single edit but a set of massive edits to LLMs; To preserve specificity, our key novel extension is a {``}selective{''} retrieval augmentation, where the retrieval-augmented IKE is only applied to {``}in-scope{''} examples, whereas the unedited model without IKE is employed for {``}out-of-scope{''} ones. 2) *Editing distillation* of MIKE using low-rank adaptation (LoRA), which distills editing abilities of MIKE to parameters of LLMs in a manner of eliminating the need of lengthy in-context demonstrations, thus removing the computational overhead encountered at the inference time. Experimental results on the zsRE and CounterFact datasets demonstrate that MIKE shows the state-of-the-art perfomrances and DistilMIKE show comparable performances with MIKE. Our code is available at https://github.com/JoveReCode/DistillMIKE.git.",
}
| Among the recently emerged knowledge editing methods, in-context knowledge editing (IKE) has shown respectable abilities on knowledge editing in terms of generalization and specificity. Noting the promising advantages but unexplored issues of IKE, we propose **DistillMIKE** as a novel extension of IKE, i.e., editing **distill**ation of ''**M**assive{''} **I**n-context **K**nowledge **E**diting in large language models (LLMs), mainly consisting of two expansions; 1) *Massive in-context knowledge editing (MIKE)*, which extends IKE to a massive editing task, aiming to inject not a single edit but a set of massive edits to LLMs; To preserve specificity, our key novel extension is a {``}selective{''} retrieval augmentation, where the retrieval-augmented IKE is only applied to {``}in-scope{''} examples, whereas the unedited model without IKE is employed for {``}out-of-scope{''} ones. 2) *Editing distillation* of MIKE using low-rank adaptation (LoRA), which distills editing abilities of MIKE to parameters of LLMs in a manner of eliminating the need of lengthy in-context demonstrations, thus removing the computational overhead encountered at the inference time. Experimental results on the zsRE and CounterFact datasets demonstrate that MIKE shows the state-of-the-art perfomrances and DistilMIKE show comparable performances with MIKE. Our code is available at https://github.com/JoveReCode/DistillMIKE.git. | [
"Qiao, Shanbao",
"Liu, Xuebing",
"Na, Seung-Hoon"
]
| {D}istill{MIKE}: Editing Distillation of Massive In-Context Knowledge Editing in Large Language Models | findings-acl.455 | Poster | 2310.10322v1 |
https://aclanthology.org/2024.findings-acl.456.bib | @inproceedings{xia-etal-2024-unlocking,
title = "Unlocking Efficiency in Large Language Model Inference: A Comprehensive Survey of Speculative Decoding",
author = "Xia, Heming and
Yang, Zhe and
Dong, Qingxiu and
Wang, Peiyi and
Li, Yongqi and
Ge, Tao and
Liu, Tianyu and
Li, Wenjie and
Sui, Zhifang",
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.456",
pages = "7655--7671",
abstract = "To mitigate the high inference latency stemming from autoregressive decoding in Large Language Models (LLMs), Speculative Decoding has emerged as a novel decoding paradigm for LLM inference. In each decoding step, this method first drafts several future tokens efficiently and then verifies them in parallel. Unlike autoregressive decoding, Speculative Decoding facilitates the simultaneous decoding of multiple tokens per step, thereby accelerating inference. This paper presents a comprehensive overview and analysis of this promising decoding paradigm. We begin by providing a formal definition and formulation of Speculative Decoding. Then, we organize in-depth discussions on its key facets, such as drafter selection and verification strategies. Furthermore, we present a comparative analysis of leading methods under third-party testing environments. We aim for this work to serve as a catalyst for further research on Speculative Decoding, ultimately contributing to more efficient LLM inference.",
}
| To mitigate the high inference latency stemming from autoregressive decoding in Large Language Models (LLMs), Speculative Decoding has emerged as a novel decoding paradigm for LLM inference. In each decoding step, this method first drafts several future tokens efficiently and then verifies them in parallel. Unlike autoregressive decoding, Speculative Decoding facilitates the simultaneous decoding of multiple tokens per step, thereby accelerating inference. This paper presents a comprehensive overview and analysis of this promising decoding paradigm. We begin by providing a formal definition and formulation of Speculative Decoding. Then, we organize in-depth discussions on its key facets, such as drafter selection and verification strategies. Furthermore, we present a comparative analysis of leading methods under third-party testing environments. We aim for this work to serve as a catalyst for further research on Speculative Decoding, ultimately contributing to more efficient LLM inference. | [
"Xia, Heming",
"Yang, Zhe",
"Dong, Qingxiu",
"Wang, Peiyi",
"Li, Yongqi",
"Ge, Tao",
"Liu, Tianyu",
"Li, Wenjie",
"Sui, Zhifang"
]
| Unlocking Efficiency in Large Language Model Inference: A Comprehensive Survey of Speculative Decoding | findings-acl.456 | Poster | 2401.07851v3 |
https://aclanthology.org/2024.findings-acl.457.bib | @inproceedings{kim-etal-2024-hierarchy,
title = "Hierarchy-aware Biased Bound Margin Loss Function for Hierarchical Text Classification",
author = "Kim, Gibaeg and
Im, SangHun and
Oh, Heung-Seon",
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.457",
pages = "7672--7682",
abstract = "Hierarchical text classification (HTC) is a challenging problem with two key issues: utilizing structural information and mitigating label imbalance. Recently, the unit-based approach generating unit-based feature representations has outperformed the global approach focusing on a global feature representation. Nevertheless, unit-based models using BCE and ZLPR losses still face static thresholding and label imbalance challenges. Those challenges become more critical in large-scale hierarchies. This paper introduces a novel hierarchy-aware loss function for unit-based HTC models: Hierarchy-aware Biased Bound Margin (HBM) loss. HBM integrates learnable bounds, biases, and a margin to address static thresholding and mitigate label imbalance adaptively. Experimental results on benchmark datasets demonstrate the superior performance of HBM compared to competitive HTC models.",
}
| Hierarchical text classification (HTC) is a challenging problem with two key issues: utilizing structural information and mitigating label imbalance. Recently, the unit-based approach generating unit-based feature representations has outperformed the global approach focusing on a global feature representation. Nevertheless, unit-based models using BCE and ZLPR losses still face static thresholding and label imbalance challenges. Those challenges become more critical in large-scale hierarchies. This paper introduces a novel hierarchy-aware loss function for unit-based HTC models: Hierarchy-aware Biased Bound Margin (HBM) loss. HBM integrates learnable bounds, biases, and a margin to address static thresholding and mitigate label imbalance adaptively. Experimental results on benchmark datasets demonstrate the superior performance of HBM compared to competitive HTC models. | [
"Kim, Gibaeg",
"Im, SangHun",
"Oh, Heung-Seon"
]
| Hierarchy-aware Biased Bound Margin Loss Function for Hierarchical Text Classification | findings-acl.457 | Poster | 2306.09132v1 |
https://aclanthology.org/2024.findings-acl.458.bib | @inproceedings{chen-etal-2024-improving-retrieval,
title = "Improving Retrieval Augmented Open-Domain Question-Answering with Vectorized Contexts",
author = "Chen, Zhuo and
Wang, Xinyu and
Jiang, Yong and
Xie, Pengjun and
Huang, Fei and
Tu, Kewei",
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.458",
pages = "7683--7694",
abstract = "In the era of large language models, applying techniques such as Retrieval Augmented Generation can better address Open-Domain Question-Answering problems. Due to constraints including model sizes and computing resources, the length of context is often limited, and it becomes challenging to empower the model to cover overlong contexts while answering questions from open domains. This paper proposes a general and convenient method to cover longer contexts in Open-Domain Question-Answering tasks. {\%}It leverages a small encoder language model that effectively encodes contexts, and the encoding applies cross-attention with origin inputs.It leverages a small encoder and cross-attention mechanism and effectively encodes contexts. With our method, the original language models can cover several times longer contexts while keeping the computing requirements close to the baseline. Our experiments demonstrate that after fine-tuning, there is improved performance across two held-in datasets, four held-out datasets, and also in two In Context Learning settings. Our code will be released at https://github.com/Alibaba-NLP/Vec-RA-ODQA.",
}
| In the era of large language models, applying techniques such as Retrieval Augmented Generation can better address Open-Domain Question-Answering problems. Due to constraints including model sizes and computing resources, the length of context is often limited, and it becomes challenging to empower the model to cover overlong contexts while answering questions from open domains. This paper proposes a general and convenient method to cover longer contexts in Open-Domain Question-Answering tasks. {\%}It leverages a small encoder language model that effectively encodes contexts, and the encoding applies cross-attention with origin inputs.It leverages a small encoder and cross-attention mechanism and effectively encodes contexts. With our method, the original language models can cover several times longer contexts while keeping the computing requirements close to the baseline. Our experiments demonstrate that after fine-tuning, there is improved performance across two held-in datasets, four held-out datasets, and also in two In Context Learning settings. Our code will be released at https://github.com/Alibaba-NLP/Vec-RA-ODQA. | [
"Chen, Zhuo",
"Wang, Xinyu",
"Jiang, Yong",
"Xie, Pengjun",
"Huang, Fei",
"Tu, Kewei"
]
| Improving Retrieval Augmented Open-Domain Question-Answering with Vectorized Contexts | findings-acl.458 | Poster | 2310.03184v2 |
https://aclanthology.org/2024.findings-acl.459.bib | @inproceedings{randl-etal-2024-cicle,
title = "{CICL}e: Conformal In-Context Learning for Largescale Multi-Class Food Risk Classification",
author = "Randl, Korbinian and
Pavlopoulos, John and
Henriksson, Aron and
Lindgren, Tony",
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.459",
pages = "7695--7715",
abstract = "Contaminated or adulterated food poses a substantial risk to human health. Given sets of labeled web texts for training, Machine Learning and Natural Language Processing can be applied to automatically detect such risks. We publish a dataset of 7,546 short texts describing public food recall announcements. Each text is manually labeled, on two granularity levels (coarse and fine), for food products and hazards that the recall corresponds to. We describe the dataset and benchmark naive, traditional, and Transformer models. Based on our analysis, Logistic Regression based on a TF-IDF representation outperforms RoBERTa and XLM-R on classes with low support. Finally, we discuss different prompting strategies and present an LLM-in-the-loop framework, based on Conformal Prediction, which boosts the performance of the base classifier while reducing energy consumption compared to normal prompting.",
}
| Contaminated or adulterated food poses a substantial risk to human health. Given sets of labeled web texts for training, Machine Learning and Natural Language Processing can be applied to automatically detect such risks. We publish a dataset of 7,546 short texts describing public food recall announcements. Each text is manually labeled, on two granularity levels (coarse and fine), for food products and hazards that the recall corresponds to. We describe the dataset and benchmark naive, traditional, and Transformer models. Based on our analysis, Logistic Regression based on a TF-IDF representation outperforms RoBERTa and XLM-R on classes with low support. Finally, we discuss different prompting strategies and present an LLM-in-the-loop framework, based on Conformal Prediction, which boosts the performance of the base classifier while reducing energy consumption compared to normal prompting. | [
"R",
"l, Korbinian",
"Pavlopoulos, John",
"Henriksson, Aron",
"Lindgren, Tony"
]
| {CICL}e: Conformal In-Context Learning for Largescale Multi-Class Food Risk Classification | findings-acl.459 | Poster | 2403.11904v3 |
https://aclanthology.org/2024.findings-acl.460.bib | @inproceedings{liu-etal-2024-intactkv,
title = "{I}ntact{KV}: Improving Large Language Model Quantization by Keeping Pivot Tokens Intact",
author = "Liu, Ruikang and
Bai, Haoli and
Lin, Haokun and
Li, Yuening and
Gao, Han and
Xu, Zhengzhuo and
Hou, Lu and
Yao, Jun and
Yuan, Chun",
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.460",
pages = "7716--7741",
abstract = "Large language models (LLMs) excel in natural language processing but demand intensive computation. To mitigate this, various quantization methods have been explored, yet they compromise LLM performance. This paper unveils a previously overlooked type of outliers in LLMs. Such outliers are found to allocate most of the attention scores on initial tokens of input, termed as pivot tokens, which are crucial to the performance of quantized LLMs. Given that, we propose IntactKV to generate the KV cache of pivot tokens losslessly from the full-precision model. The approach is simple and easy to combine with existing quantization solutions with no extra inference overhead. Besides, IntactKV can be calibrated as additional LLM parameters to boost the quantized LLMs further with minimal training costs. Mathematical analysis also proves that IntactKV effectively reduces the upper bound of quantization error. Empirical results show that IntactKV brings consistent improvement over various quantization methods across different LLMs and downstream tasks, leading to the new state-of-the-art for LLM quantization. The codes are available at https://github.com/ruikangliu/IntactKV.",
}
| Large language models (LLMs) excel in natural language processing but demand intensive computation. To mitigate this, various quantization methods have been explored, yet they compromise LLM performance. This paper unveils a previously overlooked type of outliers in LLMs. Such outliers are found to allocate most of the attention scores on initial tokens of input, termed as pivot tokens, which are crucial to the performance of quantized LLMs. Given that, we propose IntactKV to generate the KV cache of pivot tokens losslessly from the full-precision model. The approach is simple and easy to combine with existing quantization solutions with no extra inference overhead. Besides, IntactKV can be calibrated as additional LLM parameters to boost the quantized LLMs further with minimal training costs. Mathematical analysis also proves that IntactKV effectively reduces the upper bound of quantization error. Empirical results show that IntactKV brings consistent improvement over various quantization methods across different LLMs and downstream tasks, leading to the new state-of-the-art for LLM quantization. The codes are available at https://github.com/ruikangliu/IntactKV. | [
"Liu, Ruikang",
"Bai, Haoli",
"Lin, Haokun",
"Li, Yuening",
"Gao, Han",
"Xu, Zhengzhuo",
"Hou, Lu",
"Yao, Jun",
"Yuan, Chun"
]
| {I}ntact{KV}: Improving Large Language Model Quantization by Keeping Pivot Tokens Intact | findings-acl.460 | Poster | 2403.01241v2 |
https://aclanthology.org/2024.findings-acl.461.bib | @inproceedings{taniguchi-etal-2024-learning,
title = "Learning Adverbs with Spectral Mixture Kernels",
author = "Taniguchi, Tomoe and
Mochihashi, Daichi and
Kobayashi, Ichiro",
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.461",
pages = "7742--7752",
abstract = "For humans and robots to collaborate more in the real world, robots need to understand human intentions from the different manner of their behaviors. In our study, we focus on the meaning of adverbs which describe human motions. We propose a topic model, Hierarchical Dirichlet Process-Spectral Mixture Latent Dirichlet Allocation, which concurrently learns the relationship between those human motions and those adverbs by capturing the frequency kernels that represent motion characteristics and the shared topics of adverbs that depict such motions. We trained the model on datasets we made from movies about {``}walking{''} and {``}dancing{''}, and found that our model outperforms representative neural network models in terms of perplexity score. We also demonstrate our model{'}s ability to determine the adverbs for a given motion and confirmed that the model predicts more appropriate adverbs.",
}
| For humans and robots to collaborate more in the real world, robots need to understand human intentions from the different manner of their behaviors. In our study, we focus on the meaning of adverbs which describe human motions. We propose a topic model, Hierarchical Dirichlet Process-Spectral Mixture Latent Dirichlet Allocation, which concurrently learns the relationship between those human motions and those adverbs by capturing the frequency kernels that represent motion characteristics and the shared topics of adverbs that depict such motions. We trained the model on datasets we made from movies about {``}walking{''} and {``}dancing{''}, and found that our model outperforms representative neural network models in terms of perplexity score. We also demonstrate our model{'}s ability to determine the adverbs for a given motion and confirmed that the model predicts more appropriate adverbs. | [
"Taniguchi, Tomoe",
"Mochihashi, Daichi",
"Kobayashi, Ichiro"
]
| Learning Adverbs with Spectral Mixture Kernels | findings-acl.461 | Poster | 2309.15086v1 |
https://aclanthology.org/2024.findings-acl.462.bib | @inproceedings{hou-etal-2024-e,
title = "{E}-{EVAL}: A Comprehensive {C}hinese K-12 Education Evaluation Benchmark for Large Language Models",
author = "Hou, Jinchang and
Ao, Chang and
Wu, Haihong and
Kong, Xiangtao and
Zheng, Zhigang and
Tang, Daijia and
Li, Chengming and
Hu, Xiping and
Xu, Ruifeng and
Ni, Shiwen and
Yang, 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.462",
pages = "7753--7774",
abstract = "The rapid development of Large Language Models (LLMs) has led to their increasing utilization in Chinese K-12 education. Despite the growing integration of LLMs and education, the absence of a dedicated benchmark for evaluating LLMs within this domain presents a pressing concern. Consequently, there is an urgent need for a comprehensive natural language processing benchmark to precisely assess the capabilities of various LLMs in Chinese K-12 education. In response, we introduce E-EVAL, the first comprehensive evaluation benchmark specifically tailored for Chinese K-12 education. E-EVAL comprises 4,351 multiple-choice questions spanning primary, middle, and high school levels, covering a diverse array of subjects. Through meticulous evaluation, we find that Chinese-dominant models often outperform English-dominant ones, with many exceeding GPT 4.0. However, most struggle with complex subjects like mathematics. Additionally, our analysis indicates that most Chinese-dominant LLMs do not achieve higher scores at the primary school level compared to the middle school level, highlighting the nuanced relationship between proficiency in higher-order and lower-order knowledge domains. Furthermore, experimental results highlight the effectiveness of the Chain of Thought (CoT) technique in scientific subjects and Few-shot prompting in liberal arts. Through E-EVAL, we aim to conduct a rigorous analysis delineating the strengths and limitations of LLMs in educational applications, thereby contributing significantly to the advancement of Chinese K-12 education and LLMs.",
}
| The rapid development of Large Language Models (LLMs) has led to their increasing utilization in Chinese K-12 education. Despite the growing integration of LLMs and education, the absence of a dedicated benchmark for evaluating LLMs within this domain presents a pressing concern. Consequently, there is an urgent need for a comprehensive natural language processing benchmark to precisely assess the capabilities of various LLMs in Chinese K-12 education. In response, we introduce E-EVAL, the first comprehensive evaluation benchmark specifically tailored for Chinese K-12 education. E-EVAL comprises 4,351 multiple-choice questions spanning primary, middle, and high school levels, covering a diverse array of subjects. Through meticulous evaluation, we find that Chinese-dominant models often outperform English-dominant ones, with many exceeding GPT 4.0. However, most struggle with complex subjects like mathematics. Additionally, our analysis indicates that most Chinese-dominant LLMs do not achieve higher scores at the primary school level compared to the middle school level, highlighting the nuanced relationship between proficiency in higher-order and lower-order knowledge domains. Furthermore, experimental results highlight the effectiveness of the Chain of Thought (CoT) technique in scientific subjects and Few-shot prompting in liberal arts. Through E-EVAL, we aim to conduct a rigorous analysis delineating the strengths and limitations of LLMs in educational applications, thereby contributing significantly to the advancement of Chinese K-12 education and LLMs. | [
"Hou, Jinchang",
"Ao, Chang",
"Wu, Haihong",
"Kong, Xiangtao",
"Zheng, Zhigang",
"Tang, Daijia",
"Li, Chengming",
"Hu, Xiping",
"Xu, Ruifeng",
"Ni, Shiwen",
"Yang, Min"
]
| {E}-{EVAL}: A Comprehensive {C}hinese K-12 Education Evaluation Benchmark for Large Language Models | findings-acl.462 | Poster | 2401.15927v1 |
https://aclanthology.org/2024.findings-acl.463.bib | @inproceedings{meng-etal-2024-chartassistant,
title = "{C}hart{A}ssistant: A Universal Chart Multimodal Language Model via Chart-to-Table Pre-training and Multitask Instruction Tuning",
author = "Meng, Fanqing and
Shao, Wenqi and
Lu, Quanfeng and
Gao, Peng and
Zhang, Kaipeng and
Qiao, Yu and
Luo, 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.463",
pages = "7775--7803",
abstract = "Charts play a vital role in data visualization, understanding data patterns, and informed decision-making. However, their unique combination of graphical elements (e.g., bars, lines) and textual components (e.g., labels, legends) poses challenges for general-purpose multimodal models. While vision-language models trained on chart data excel in comprehension, they struggle with generalization. To address these challenges, we propose ChartAssistant, a chart-based vision-language model for universal chart comprehension and reasoning. ChartAssistant leverages ChartSFT, a comprehensive dataset covering diverse chart-related tasks with basic (e.g. bars and pies) and specialized (e.g. radars, and bubbles) chart types. It undergoes a two-stage training process, starting with pre-training on chart-to-table parsing to align chart and text, followed by multitask instruction-following fine-tuning. This approach enables ChartAssistant to achieve competitive performance across various chart tasks. Experimental results demonstrate significant performance gains over the state-of-the-art UniChart and ChartLlama methods, especially outperforming them on real-world chart data with zero-shot setting. The code and data are available at https://github.com/OpenGVLab/ChartAst.",
}
| Charts play a vital role in data visualization, understanding data patterns, and informed decision-making. However, their unique combination of graphical elements (e.g., bars, lines) and textual components (e.g., labels, legends) poses challenges for general-purpose multimodal models. While vision-language models trained on chart data excel in comprehension, they struggle with generalization. To address these challenges, we propose ChartAssistant, a chart-based vision-language model for universal chart comprehension and reasoning. ChartAssistant leverages ChartSFT, a comprehensive dataset covering diverse chart-related tasks with basic (e.g. bars and pies) and specialized (e.g. radars, and bubbles) chart types. It undergoes a two-stage training process, starting with pre-training on chart-to-table parsing to align chart and text, followed by multitask instruction-following fine-tuning. This approach enables ChartAssistant to achieve competitive performance across various chart tasks. Experimental results demonstrate significant performance gains over the state-of-the-art UniChart and ChartLlama methods, especially outperforming them on real-world chart data with zero-shot setting. The code and data are available at https://github.com/OpenGVLab/ChartAst. | [
"Meng, Fanqing",
"Shao, Wenqi",
"Lu, Quanfeng",
"Gao, Peng",
"Zhang, Kaipeng",
"Qiao, Yu",
"Luo, Ping"
]
| {C}hart{A}ssistant: A Universal Chart Multimodal Language Model via Chart-to-Table Pre-training and Multitask Instruction Tuning | findings-acl.463 | Poster | 2401.02384v3 |
https://aclanthology.org/2024.findings-acl.464.bib | @inproceedings{li-etal-2024-teaching,
title = "Teaching Small Language Models to Reason for Knowledge-Intensive Multi-Hop Question Answering",
author = "Li, Xiang and
He, Shizhu and
Lei, Fangyu and
JunYang, JunYang and
Su, Tianhuang and
Liu, Kang and
Zhao, 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.464",
pages = "7804--7816",
abstract = "Large Language Models (LLMs) can teach small language models (SLMs) to solve complex reasoning tasks (e.g., mathematical question answering) by Chain-of-thought Distillation (CoTD). Specifically, CoTD fine-tunes SLMs by utilizing rationales generated from LLMs such as ChatGPT. However, CoTD has certain limitations that make it unsuitable for knowledge-intensive multi-hop question answering: 1) SLMs have a very limited capacity in memorizing required knowledge compared to LLMs. 2) SLMs do not possess the same powerful integrated abilities in question understanding and knowledge reasoning as LLMs. To address the above limitations, we introduce Decompose-and-Response Distillation (D{\&}R Distillation), which distills two student models, namely Decomposer and Responser separately. The two models solve a knowledge-intensive multi-hop question through an interactive process of asking and answering subquestions. Our method offers two advantages: 1) SLMs have the capability to access external knowledge to address subquestions, which provides more comprehensive knowledge for multi-hop questions. 2) By employing simpler subquestions instead of complex CoT reasoning, SLMs effectively mitigate task complexity and decrease data prerequisites. Experimental results on three knowledge-intensive multi-hop question answering datasets demonstrate that D{\&}R Distillation can surpass previous CoTD methods, even with much less training data.",
}
| Large Language Models (LLMs) can teach small language models (SLMs) to solve complex reasoning tasks (e.g., mathematical question answering) by Chain-of-thought Distillation (CoTD). Specifically, CoTD fine-tunes SLMs by utilizing rationales generated from LLMs such as ChatGPT. However, CoTD has certain limitations that make it unsuitable for knowledge-intensive multi-hop question answering: 1) SLMs have a very limited capacity in memorizing required knowledge compared to LLMs. 2) SLMs do not possess the same powerful integrated abilities in question understanding and knowledge reasoning as LLMs. To address the above limitations, we introduce Decompose-and-Response Distillation (D{\&}R Distillation), which distills two student models, namely Decomposer and Responser separately. The two models solve a knowledge-intensive multi-hop question through an interactive process of asking and answering subquestions. Our method offers two advantages: 1) SLMs have the capability to access external knowledge to address subquestions, which provides more comprehensive knowledge for multi-hop questions. 2) By employing simpler subquestions instead of complex CoT reasoning, SLMs effectively mitigate task complexity and decrease data prerequisites. Experimental results on three knowledge-intensive multi-hop question answering datasets demonstrate that D{\&}R Distillation can surpass previous CoTD methods, even with much less training data. | [
"Li, Xiang",
"He, Shizhu",
"Lei, Fangyu",
"JunYang, JunYang",
"Su, Tianhuang",
"Liu, Kang",
"Zhao, Jun"
]
| Teaching Small Language Models to Reason for Knowledge-Intensive Multi-Hop Question Answering | findings-acl.464 | Poster | 2305.03453v4 |
https://aclanthology.org/2024.findings-acl.465.bib | @inproceedings{lai-etal-2024-alarm,
title = "{AL}a{RM}: Align Language Models via Hierarchical Rewards Modeling",
author = "Lai, Yuhang and
Wang, Siyuan and
Liu, Shujun and
Huang, Xuanjing and
Wei, Zhongyu",
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.465",
pages = "7817--7831",
abstract = "We introduce ALaRM, the first framework modeling hierarchical rewards in reinforcement learning from human feedback (RLHF), which is designed to enhance the alignment of large language models (LLMs) with human preferences. The framework addresses the limitations of current alignment approaches, which often struggle with the inconsistency and sparsity of human supervision signals, by integrating holistic rewards with aspect-specific rewards. This integration enables more precise and consistent guidance of language models towards desired outcomes, particularly in complex and open text generation tasks. By employing a methodology that filters and combines multiple rewards based on their consistency, the framework provides a reliable mechanism for improving model alignment. We validate our approach through applications in long-form question answering and machine translation tasks, employing gpt-3.5-turbo for pairwise comparisons, and demonstrate improvements over existing baselines. Our work underscores the effectiveness of hierarchical rewards modeling in refining LLM training processes for better human preference alignment. We release our code at https://ALaRM-fdu.github.io.",
}
| We introduce ALaRM, the first framework modeling hierarchical rewards in reinforcement learning from human feedback (RLHF), which is designed to enhance the alignment of large language models (LLMs) with human preferences. The framework addresses the limitations of current alignment approaches, which often struggle with the inconsistency and sparsity of human supervision signals, by integrating holistic rewards with aspect-specific rewards. This integration enables more precise and consistent guidance of language models towards desired outcomes, particularly in complex and open text generation tasks. By employing a methodology that filters and combines multiple rewards based on their consistency, the framework provides a reliable mechanism for improving model alignment. We validate our approach through applications in long-form question answering and machine translation tasks, employing gpt-3.5-turbo for pairwise comparisons, and demonstrate improvements over existing baselines. Our work underscores the effectiveness of hierarchical rewards modeling in refining LLM training processes for better human preference alignment. We release our code at https://ALaRM-fdu.github.io. | [
"Lai, Yuhang",
"Wang, Siyuan",
"Liu, Shujun",
"Huang, Xuanjing",
"Wei, Zhongyu"
]
| {AL}a{RM}: Align Language Models via Hierarchical Rewards Modeling | findings-acl.465 | Poster | 2403.06754v2 |
https://aclanthology.org/2024.findings-acl.466.bib | @inproceedings{liu-etal-2024-lstprompt,
title = "{LSTP}rompt: Large Language Models as Zero-Shot Time Series Forecasters by Long-Short-Term Prompting",
author = "Liu, Haoxin and
Zhao, Zhiyuan and
Wang, Jindong and
Kamarthi, Harshavardhan and
Prakash, B. Aditya",
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.466",
pages = "7832--7840",
abstract = "Time-series forecasting (TSF) finds broad applications in real-world scenarios. Prompting off-the-shelf Large Language Models (LLMs) demonstrates strong zero-shot TSF capabilities while preserving computational efficiency. However, existing prompting methods oversimplify TSF as language next-token predictions, overlooking its dynamic nature and lack of integration with state-of-the-art prompt strategies such as Chain-of-Thought. Thus, we propose LSTPrompt, a novel approach for prompting LLMs in zero-shot TSF tasks. LSTPrompt decomposes TSF into short-term and long-term forecasting sub-tasks, tailoring prompts to each. LSTPrompt guides LLMs to regularly reassess forecasting mechanisms to enhance adaptability. Extensive evaluations demonstrate consistently better performance of LSTPrompt than existing prompting methods, and competitive results compared to foundation TSF models.",
}
| Time-series forecasting (TSF) finds broad applications in real-world scenarios. Prompting off-the-shelf Large Language Models (LLMs) demonstrates strong zero-shot TSF capabilities while preserving computational efficiency. However, existing prompting methods oversimplify TSF as language next-token predictions, overlooking its dynamic nature and lack of integration with state-of-the-art prompt strategies such as Chain-of-Thought. Thus, we propose LSTPrompt, a novel approach for prompting LLMs in zero-shot TSF tasks. LSTPrompt decomposes TSF into short-term and long-term forecasting sub-tasks, tailoring prompts to each. LSTPrompt guides LLMs to regularly reassess forecasting mechanisms to enhance adaptability. Extensive evaluations demonstrate consistently better performance of LSTPrompt than existing prompting methods, and competitive results compared to foundation TSF models. | [
"Liu, Haoxin",
"Zhao, Zhiyuan",
"Wang, Jindong",
"Kamarthi, Harshavardhan",
"Prakash, B. Aditya"
]
| {LSTP}rompt: Large Language Models as Zero-Shot Time Series Forecasters by Long-Short-Term Prompting | findings-acl.466 | Poster | 2210.08964v5 |
https://aclanthology.org/2024.findings-acl.467.bib | @inproceedings{lu-etal-2024-mitigating,
title = "Mitigating Boundary Ambiguity and Inherent Bias for Text Classification in the Era of Large Language Models",
author = "Lu, Zhenyi and
Tian, Jie and
Wei, Wei and
Qu, Xiaoye and
Cheng, Yu and
Xie, Wenfeng and
Chen, Dangyang",
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.467",
pages = "7841--7864",
abstract = "Text classification is a crucial task encountered frequently in practical scenarios, yet it is still under-explored in the era of large language models (LLMs). This study shows that LLMs are vulnerable to changes in the number and arrangement of options in text classification. Our extensive empirical analyses reveal that the key bottleneck arises from ambiguous decision boundaries and inherent biases towards specific tokens and positions.To mitigate these issues, we make the first attempt and propose a novel two-stage classification framework for LLMs. Our approach is grounded in the empirical observation that pairwise comparisons can effectively alleviate boundary ambiguity and inherent bias. Specifically, we begin with a self-reduction technique to efficiently narrow down numerous options, which contributes to reduced decision space and a faster comparison process. Subsequently, pairwise contrastive comparisons are employed in a chain-of-thought manner to draw out nuances and distinguish confusable options, thus refining the ambiguous decision boundary.Extensive experiments on four datasets (Banking77, HWU64, LIU54, and Clinic150) verify the effectiveness of our framework. Furthermore, benefitting from our framework, various LLMs can achieve consistent improvements. Our code and data are available in \url{https://github.com/Chuge0335/PC-CoT}.",
}
| Text classification is a crucial task encountered frequently in practical scenarios, yet it is still under-explored in the era of large language models (LLMs). This study shows that LLMs are vulnerable to changes in the number and arrangement of options in text classification. Our extensive empirical analyses reveal that the key bottleneck arises from ambiguous decision boundaries and inherent biases towards specific tokens and positions.To mitigate these issues, we make the first attempt and propose a novel two-stage classification framework for LLMs. Our approach is grounded in the empirical observation that pairwise comparisons can effectively alleviate boundary ambiguity and inherent bias. Specifically, we begin with a self-reduction technique to efficiently narrow down numerous options, which contributes to reduced decision space and a faster comparison process. Subsequently, pairwise contrastive comparisons are employed in a chain-of-thought manner to draw out nuances and distinguish confusable options, thus refining the ambiguous decision boundary.Extensive experiments on four datasets (Banking77, HWU64, LIU54, and Clinic150) verify the effectiveness of our framework. Furthermore, benefitting from our framework, various LLMs can achieve consistent improvements. Our code and data are available in \url{https://github.com/Chuge0335/PC-CoT}. | [
"Lu, Zhenyi",
"Tian, Jie",
"Wei, Wei",
"Qu, Xiaoye",
"Cheng, Yu",
"Xie, Wenfeng",
"Chen, Dangyang"
]
| Mitigating Boundary Ambiguity and Inherent Bias for Text Classification in the Era of Large Language Models | findings-acl.467 | Poster | 2406.07001v1 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.