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.268.bib | @inproceedings{shen-etal-2024-empathicstories,
title = "{E}mpathic{S}tories++: A Multimodal Dataset for Empathy Towards Personal Experiences",
author = "Shen, Jocelyn and
Kim, Yubin and
Hulse, Mohit and
Zulfikar, Wazeer and
Alghowinem, Sharifa and
Breazeal, Cynthia and
Park, Hae",
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.268",
pages = "4525--4536",
abstract = "Modeling empathy is a complex endeavor that is rooted in interpersonal and experiential dimensions of human interaction, and remains an open problem within AI. Existing empathy datasets fall short in capturing the richness of empathy responses, often being confined to in-lab or acted scenarios, lacking longitudinal data, and missing self-reported labels. We introduce a new multimodal dataset for empathy during personal experience sharing: the EmpathicStories++ dataset containing 53 hours of video, audio, and text data of 41 participants sharing vulnerable experiences and reading empathically resonant stories with an AI agent. EmpathicStories++ is the first longitudinal dataset on empathy, collected over a month-long deployment of social robots in participants{'} homes, as participants engage in natural, empathic storytelling interactions with AI agents. We then introduce a novel task of predicting individuals{'} empathy toward others{'} stories based on their personal experiences, evaluated in two contexts: participants{'} own personal shared story context and their reflections on stories they read. We benchmark this task using state-of-the-art models to pave the way for future improvements in contextualized and longitudinal empathy modeling. Our work provides a valuable resource for further research in developing empathetic AI systems and understanding the intricacies of human empathy within genuine, real-world settings.",
}
| Modeling empathy is a complex endeavor that is rooted in interpersonal and experiential dimensions of human interaction, and remains an open problem within AI. Existing empathy datasets fall short in capturing the richness of empathy responses, often being confined to in-lab or acted scenarios, lacking longitudinal data, and missing self-reported labels. We introduce a new multimodal dataset for empathy during personal experience sharing: the EmpathicStories++ dataset containing 53 hours of video, audio, and text data of 41 participants sharing vulnerable experiences and reading empathically resonant stories with an AI agent. EmpathicStories++ is the first longitudinal dataset on empathy, collected over a month-long deployment of social robots in participants{'} homes, as participants engage in natural, empathic storytelling interactions with AI agents. We then introduce a novel task of predicting individuals{'} empathy toward others{'} stories based on their personal experiences, evaluated in two contexts: participants{'} own personal shared story context and their reflections on stories they read. We benchmark this task using state-of-the-art models to pave the way for future improvements in contextualized and longitudinal empathy modeling. Our work provides a valuable resource for further research in developing empathetic AI systems and understanding the intricacies of human empathy within genuine, real-world settings. | [
"Shen, Jocelyn",
"Kim, Yubin",
"Hulse, Mohit",
"Zulfikar, Wazeer",
"Alghowinem, Sharifa",
"Breazeal, Cynthia",
"Park, Hae"
]
| {E}mpathic{S}tories++: A Multimodal Dataset for Empathy Towards Personal Experiences | findings-acl.268 | Poster | 2405.15708v1 |
https://aclanthology.org/2024.findings-acl.269.bib | @inproceedings{shmidman-etal-2024-mrl,
title = "{MRL} Parsing Without Tears: The Case of {H}ebrew",
author = "Shmidman, Shaltiel and
Shmidman, Avi and
Koppel, Moshe 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.269",
pages = "4537--4550",
abstract = "Syntactic parsing remains a critical tool for relation extraction and information extraction, especially in resource-scarce languages where LLMs are lacking. Yet in morphologically rich languages (MRLs), where parsers need to identify multiple lexical units in each token, existing systems suffer in latency and setup complexity. Some use a pipeline to peel away the layers: first segmentation, then morphology tagging, and then syntax parsing; however, errors in earlier layers are then propagated forward. Others use a joint architecture to evaluate all permutations at once; while this improves accuracy, it is notoriously slow. In contrast, and taking Hebrew as a test case, we present a new {``}flipped pipeline{''}: decisions are made directly on the whole-token units by expert classifiers, each one dedicated to one specific task. The classifier predictions are independent of one another, and only at the end do we synthesize their predictions. This blazingly fast approach requires only a single huggingface call, without the need for recourse to lexicons or linguistic resources. When trained on the same training set used in previous studies, our model achieves near-SOTA performance on a wide array of Hebrew NLP tasks. Furthermore, when trained on a newly enlarged training corpus, our model achieves a new SOTA for Hebrew POS tagging and dependency parsing. We release this new SOTA model to the community. Because our architecture does not rely on any language-specific resources, it can serve as a model to develop similar parsers for other MRLs.",
}
| Syntactic parsing remains a critical tool for relation extraction and information extraction, especially in resource-scarce languages where LLMs are lacking. Yet in morphologically rich languages (MRLs), where parsers need to identify multiple lexical units in each token, existing systems suffer in latency and setup complexity. Some use a pipeline to peel away the layers: first segmentation, then morphology tagging, and then syntax parsing; however, errors in earlier layers are then propagated forward. Others use a joint architecture to evaluate all permutations at once; while this improves accuracy, it is notoriously slow. In contrast, and taking Hebrew as a test case, we present a new {``}flipped pipeline{''}: decisions are made directly on the whole-token units by expert classifiers, each one dedicated to one specific task. The classifier predictions are independent of one another, and only at the end do we synthesize their predictions. This blazingly fast approach requires only a single huggingface call, without the need for recourse to lexicons or linguistic resources. When trained on the same training set used in previous studies, our model achieves near-SOTA performance on a wide array of Hebrew NLP tasks. Furthermore, when trained on a newly enlarged training corpus, our model achieves a new SOTA for Hebrew POS tagging and dependency parsing. We release this new SOTA model to the community. Because our architecture does not rely on any language-specific resources, it can serve as a model to develop similar parsers for other MRLs. | [
"Shmidman, Shaltiel",
"Shmidman, Avi",
"Koppel, Moshe",
"Tsarfaty, Reut"
]
| {MRL} Parsing Without Tears: The Case of {H}ebrew | findings-acl.269 | Poster | 2403.06970v1 |
https://aclanthology.org/2024.findings-acl.270.bib | @inproceedings{amara-etal-2024-syntaxshap,
title = "{S}yntax{S}hap: Syntax-aware Explainability Method for Text Generation",
author = "Amara, Kenza and
Sevastjanova, Rita and
El-Assady, Mennatallah",
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.270",
pages = "4551--4566",
abstract = "To harness the power of large language models in safety-critical domains, we need to ensure the explainability of their predictions. However, despite the significant attention to model interpretability, there remains an unexplored domain in explaining sequence-to-sequence tasks using methods tailored for textual data. This paper introduces *SyntaxShap*, a local, model-agnostic explainability method for text generation that takes into consideration the syntax in the text data. The presented work extends Shapley values to account for parsing-based syntactic dependencies. Taking a game theoric approach, SyntaxShap only considers coalitions constraint by the dependency tree. We adopt a model-based evaluation to compare SyntaxShap and its weighted form to state-of-the-art explainability methods adapted to text generation tasks, using diverse metrics including faithfulness, coherency, and semantic alignment of the explanations to the model. We show that our syntax-aware method produces explanations that help build more faithful and coherent explanations for predictions by autoregressive models. Confronted with the misalignment of human and AI model reasoning, this paper also highlights the need for cautious evaluation strategies in explainable AI.",
}
| To harness the power of large language models in safety-critical domains, we need to ensure the explainability of their predictions. However, despite the significant attention to model interpretability, there remains an unexplored domain in explaining sequence-to-sequence tasks using methods tailored for textual data. This paper introduces *SyntaxShap*, a local, model-agnostic explainability method for text generation that takes into consideration the syntax in the text data. The presented work extends Shapley values to account for parsing-based syntactic dependencies. Taking a game theoric approach, SyntaxShap only considers coalitions constraint by the dependency tree. We adopt a model-based evaluation to compare SyntaxShap and its weighted form to state-of-the-art explainability methods adapted to text generation tasks, using diverse metrics including faithfulness, coherency, and semantic alignment of the explanations to the model. We show that our syntax-aware method produces explanations that help build more faithful and coherent explanations for predictions by autoregressive models. Confronted with the misalignment of human and AI model reasoning, this paper also highlights the need for cautious evaluation strategies in explainable AI. | [
"Amara, Kenza",
"Sevastjanova, Rita",
"El-Assady, Mennatallah"
]
| {S}yntax{S}hap: Syntax-aware Explainability Method for Text Generation | findings-acl.270 | Poster | 2402.09259v2 |
https://aclanthology.org/2024.findings-acl.271.bib | @inproceedings{srinath-etal-2024-automated,
title = "Automated Detection and Analysis of Data Practices Using A Real-World Corpus",
author = "Srinath, Mukund and
Narayanan Venkit, Pranav and
Badillo, Maria and
Schaub, Florian and
Giles, C. and
Wilson, Shomir",
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.271",
pages = "4567--4574",
abstract = "Privacy policies are crucial for informing users about data practices, yet their length and complexity often deter users from reading them. In this paper, we propose an automated approach to identify and visualize data practices within privacy policies at different levels of detail. Leveraging crowd-sourced annotations from the ToS;DR platform, we experiment with various methods to match policy excerpts with predefined data practice descriptions. We further conduct a case study to evaluate our approach on a real-world policy, demonstrating its effectiveness in simplifying complex policies. Experiments show that our approach accurately matches data practice descriptions with policy excerpts, facilitating the presentation of simplified privacy information to users.",
}
| Privacy policies are crucial for informing users about data practices, yet their length and complexity often deter users from reading them. In this paper, we propose an automated approach to identify and visualize data practices within privacy policies at different levels of detail. Leveraging crowd-sourced annotations from the ToS;DR platform, we experiment with various methods to match policy excerpts with predefined data practice descriptions. We further conduct a case study to evaluate our approach on a real-world policy, demonstrating its effectiveness in simplifying complex policies. Experiments show that our approach accurately matches data practice descriptions with policy excerpts, facilitating the presentation of simplified privacy information to users. | [
"Srinath, Mukund",
"Narayanan Venkit, Pranav",
"Badillo, Maria",
"Schaub, Florian",
"Giles, C.",
"Wilson, Shomir"
]
| Automated Detection and Analysis of Data Practices Using A Real-World Corpus | findings-acl.271 | Poster | 2401.17996v1 |
https://aclanthology.org/2024.findings-acl.272.bib | @inproceedings{fan-etal-2024-enhancing,
title = "Enhancing Hyperbolic Knowledge Graph Embeddings via Lorentz Transformations",
author = "Fan, Xiran and
Xu, Minghua and
Chen, Huiyuan and
Chen, Yuzhong and
Das, Mahashweta and
Yang, Hao",
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.272",
pages = "4575--4589",
abstract = "Knowledge Graph Embedding (KGE) is a powerful technique for predicting missing links in Knowledge Graphs (KGs) by learning the entities and relations. Hyperbolic space has emerged as a promising embedding space for KGs due to its ability to represent hierarchical data. Nevertheless, most existing hyperbolic KGE methods rely on tangent approximation and are not fully hyperbolic, resulting in distortions and inaccuracies. To overcome this limitation, we propose LorentzKG, a fully hyperbolic KGE method that represents entities as points in the Lorentz model and represents relations as the intrinsic transformation{---}the Lorentz transformations between entities. We demonstrate that the Lorentz transformation, which can be decomposed into Lorentz rotation/reflection and Lorentz boost, captures various types of relations including hierarchical structures. Experimental results show that our LorentzKG achieves state-of-the-art performance.",
}
| Knowledge Graph Embedding (KGE) is a powerful technique for predicting missing links in Knowledge Graphs (KGs) by learning the entities and relations. Hyperbolic space has emerged as a promising embedding space for KGs due to its ability to represent hierarchical data. Nevertheless, most existing hyperbolic KGE methods rely on tangent approximation and are not fully hyperbolic, resulting in distortions and inaccuracies. To overcome this limitation, we propose LorentzKG, a fully hyperbolic KGE method that represents entities as points in the Lorentz model and represents relations as the intrinsic transformation{---}the Lorentz transformations between entities. We demonstrate that the Lorentz transformation, which can be decomposed into Lorentz rotation/reflection and Lorentz boost, captures various types of relations including hierarchical structures. Experimental results show that our LorentzKG achieves state-of-the-art performance. | [
"Fan, Xiran",
"Xu, Minghua",
"Chen, Huiyuan",
"Chen, Yuzhong",
"Das, Mahashweta",
"Yang, Hao"
]
| Enhancing Hyperbolic Knowledge Graph Embeddings via Lorentz Transformations | findings-acl.272 | Poster | 2302.04545v2 |
https://aclanthology.org/2024.findings-acl.273.bib | @inproceedings{burns-etal-2024-tell,
title = "Tell Me What{'}s Next: Textual Foresight for Generic {UI} Representations",
author = "Burns, Andrea and
Saenko, Kate and
Plummer, Bryan",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.273",
pages = "4590--4611",
abstract = "Mobile app user interfaces (UIs) are rich with action, text, structure, and image content that can be utilized to learn generic UI representations for tasks like automating user commands, summarizing content, and evaluating the accessibility of user interfaces. Prior work has learned strong visual representations with local or global captioning losses, but fails to retain both granularities.To combat this, we propose Textual Foresight, a novel pretraining objective for learning UI screen representations. Textual Foresight generates global text descriptions of future UI states given a current UI and local action taken. Our approach requires joint reasoning over elements and entire screens, resulting in improved UI features: on generation tasks, UI agents trained with Textual Foresight outperform state-of-the-art by 2{\%} with 28x fewer images. We train with our newly constructed mobile app dataset, OpenApp, which results in the first public dataset for app UI representation learning. OpenApp enables new baselines, and we find Textual Foresight improves average task performance over them by 5.7{\%} while having access to 2x less data.",
}
| Mobile app user interfaces (UIs) are rich with action, text, structure, and image content that can be utilized to learn generic UI representations for tasks like automating user commands, summarizing content, and evaluating the accessibility of user interfaces. Prior work has learned strong visual representations with local or global captioning losses, but fails to retain both granularities.To combat this, we propose Textual Foresight, a novel pretraining objective for learning UI screen representations. Textual Foresight generates global text descriptions of future UI states given a current UI and local action taken. Our approach requires joint reasoning over elements and entire screens, resulting in improved UI features: on generation tasks, UI agents trained with Textual Foresight outperform state-of-the-art by 2{\%} with 28x fewer images. We train with our newly constructed mobile app dataset, OpenApp, which results in the first public dataset for app UI representation learning. OpenApp enables new baselines, and we find Textual Foresight improves average task performance over them by 5.7{\%} while having access to 2x less data. | [
"Burns, Andrea",
"Saenko, Kate",
"Plummer, Bryan"
]
| Tell Me What{'}s Next: Textual Foresight for Generic {UI} Representations | findings-acl.273 | Poster | 2406.07822v2 |
https://aclanthology.org/2024.findings-acl.274.bib | @inproceedings{zahid-etal-2024-probing,
title = "Probing the Uniquely Identifiable Linguistic Patterns of Conversational {AI} Agents",
author = "Zahid, Iqra and
Madusanka, Tharindu and
Batista-Navarro, Riza and
Sun, Youcheng",
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.274",
pages = "4612--4628",
abstract = "The proliferation of Conversational AI agents (CAAs) has emphasised the need to distinguish between human and machine-generated texts, with implications spanning digital forensics and cybersecurity. While prior research primarily focussed on distinguishing human from machine-generated text, our study takes a more refined approach by analysing different CAAs. We construct linguistic profiles for five CAAs, aiming to identify Uniquely Identifiable Linguistic Patterns (UILPs) for each model using authorship attribution techniques. Authorship attribution (AA) is the task of identifying the author of an unknown text from a pool of known authors. Our research seeks to answer crucial questions about the existence of UILPs in CAAs, the linguistic overlap between various text types generated by these models, and the feasibility of Authorship Attribution (AA) for CAAs based on UILPs. Promisingly, we are able to attribute CAAs based on their original texts with a weighted F1-score of 96.94{\%}. Further, we are able to attribute CAAs according to their writing style (as specified by prompts), yielding a weighted F1-score of 95.84{\%}, which sets the baseline for this task. By employing principal component analysis (PCA), we identify the top 100 most informative linguistic features for each CAA, achieving a weighted F1-score ranging from 86.04{\%} to 97.93{\%}, and an overall weighted F1-score of 93.86{\%}.",
}
| The proliferation of Conversational AI agents (CAAs) has emphasised the need to distinguish between human and machine-generated texts, with implications spanning digital forensics and cybersecurity. While prior research primarily focussed on distinguishing human from machine-generated text, our study takes a more refined approach by analysing different CAAs. We construct linguistic profiles for five CAAs, aiming to identify Uniquely Identifiable Linguistic Patterns (UILPs) for each model using authorship attribution techniques. Authorship attribution (AA) is the task of identifying the author of an unknown text from a pool of known authors. Our research seeks to answer crucial questions about the existence of UILPs in CAAs, the linguistic overlap between various text types generated by these models, and the feasibility of Authorship Attribution (AA) for CAAs based on UILPs. Promisingly, we are able to attribute CAAs based on their original texts with a weighted F1-score of 96.94{\%}. Further, we are able to attribute CAAs according to their writing style (as specified by prompts), yielding a weighted F1-score of 95.84{\%}, which sets the baseline for this task. By employing principal component analysis (PCA), we identify the top 100 most informative linguistic features for each CAA, achieving a weighted F1-score ranging from 86.04{\%} to 97.93{\%}, and an overall weighted F1-score of 93.86{\%}. | [
"Zahid, Iqra",
"Madusanka, Tharindu",
"Batista-Navarro, Riza",
"Sun, Youcheng"
]
| Probing the Uniquely Identifiable Linguistic Patterns of Conversational {AI} Agents | findings-acl.274 | Poster | 2305.09770v6 |
https://aclanthology.org/2024.findings-acl.275.bib | @inproceedings{salinas-morstatter-2024-butterfly,
title = "The Butterfly Effect of Altering Prompts: How Small Changes and Jailbreaks Affect Large Language Model Performance",
author = "Salinas, Abel and
Morstatter, Fred",
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.275",
pages = "4629--4651",
abstract = "Large Language Models (LLMs) are regularly being used to label data across many domains and for myriad tasks. By simply asking the LLM for an answer, or {``}prompting,{''} practitioners are able to use LLMs to quickly get a response for an arbitrary task. This prompting is done through a series of decisions by the practitioner, from simple wording of the prompt, to requesting the output in a certain data format, to jailbreaking in the case of prompts that address more sensitive topics. In this work, we ask: do variations in the way a prompt is constructed change the ultimate decision of the LLM? We answer this using a series of prompt variations across a variety of text classification tasks. We find that even the smallest of perturbations, such as adding a space at the end of a prompt, can cause the LLM to change its answer. Further, we find that requesting responses in XML and commonly used jailbreaks can have cataclysmic effects on the data labeled by LLMs.",
}
| Large Language Models (LLMs) are regularly being used to label data across many domains and for myriad tasks. By simply asking the LLM for an answer, or {``}prompting,{''} practitioners are able to use LLMs to quickly get a response for an arbitrary task. This prompting is done through a series of decisions by the practitioner, from simple wording of the prompt, to requesting the output in a certain data format, to jailbreaking in the case of prompts that address more sensitive topics. In this work, we ask: do variations in the way a prompt is constructed change the ultimate decision of the LLM? We answer this using a series of prompt variations across a variety of text classification tasks. We find that even the smallest of perturbations, such as adding a space at the end of a prompt, can cause the LLM to change its answer. Further, we find that requesting responses in XML and commonly used jailbreaks can have cataclysmic effects on the data labeled by LLMs. | [
"Salinas, Abel",
"Morstatter, Fred"
]
| The Butterfly Effect of Altering Prompts: How Small Changes and Jailbreaks Affect Large Language Model Performance | findings-acl.275 | Poster | 2401.03729v3 |
https://aclanthology.org/2024.findings-acl.276.bib | @inproceedings{xu-etal-2024-x,
title = "{X}-Shot: A Unified System to Handle Frequent, Few-shot and Zero-shot Learning Simultaneously in Classification",
author = "Xu, Hanzi and
Chen, Muhao and
Huang, Lifu and
Vucetic, Slobodan and
Yin, Wenpeng",
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.276",
pages = "4652--4665",
abstract = "In recent years, few-shot and zero-shot learning, which learn to predict labels with limited annotated instances, have garnered significant attention. Traditional approaches often treat frequent-shot (freq-shot; labels with abundant instances), few-shot, and zero-shot learning as distinct challenges, optimizing systems for just one of these scenarios. Yet, in real-world settings, label occurrences vary greatly. Some of them might appear thousands of times, while others might only appear sporadically or not at all. For practical deployment, it is crucial that a system can adapt to any label occurrence. We introduce a novel classification challenge: **X-shot**, reflecting a real-world context where freq-shot, few-shot, and zero-shot labels co-occur without predefined limits. Here, **X** can span from 0 to positive infinity. The crux of **X-shot** centers on open-domain generalization and devising a system versatile enough to manage various label scenarios. To solve **X-shot**, we propose **BinBin** (**B**inary **IN**ference **B**ased on **IN**struction following) that leverages the Indirect Supervision from a large collection of NLP tasks via instruction following, bolstered by Weak Supervision provided by large language models. **BinBin** surpasses previous state-of-the-art techniques on three benchmark datasets across multiple domains. To our knowledge, this is the first work addressing **X-shot** learning, where **X** remains variable.",
}
| In recent years, few-shot and zero-shot learning, which learn to predict labels with limited annotated instances, have garnered significant attention. Traditional approaches often treat frequent-shot (freq-shot; labels with abundant instances), few-shot, and zero-shot learning as distinct challenges, optimizing systems for just one of these scenarios. Yet, in real-world settings, label occurrences vary greatly. Some of them might appear thousands of times, while others might only appear sporadically or not at all. For practical deployment, it is crucial that a system can adapt to any label occurrence. We introduce a novel classification challenge: **X-shot**, reflecting a real-world context where freq-shot, few-shot, and zero-shot labels co-occur without predefined limits. Here, **X** can span from 0 to positive infinity. The crux of **X-shot** centers on open-domain generalization and devising a system versatile enough to manage various label scenarios. To solve **X-shot**, we propose **BinBin** (**B**inary **IN**ference **B**ased on **IN**struction following) that leverages the Indirect Supervision from a large collection of NLP tasks via instruction following, bolstered by Weak Supervision provided by large language models. **BinBin** surpasses previous state-of-the-art techniques on three benchmark datasets across multiple domains. To our knowledge, this is the first work addressing **X-shot** learning, where **X** remains variable. | [
"Xu, Hanzi",
"Chen, Muhao",
"Huang, Lifu",
"Vucetic, Slobodan",
"Yin, Wenpeng"
]
| {X}-Shot: A Unified System to Handle Frequent, Few-shot and Zero-shot Learning Simultaneously in Classification | findings-acl.276 | Poster | 2403.03863v1 |
https://aclanthology.org/2024.findings-acl.277.bib | @inproceedings{jiao-etal-2024-spin,
title = "{SPIN}: Sparsifying and Integrating Internal Neurons in Large Language Models for Text Classification",
author = {Jiao, Difan and
Liu, Yilun and
Tang, Zhenwei and
Matter, Daniel and
Pfeffer, J{\"u}rgen and
Anderson, Ashton},
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.277",
pages = "4666--4682",
abstract = "Among the many tasks that Large Language Models (LLMs) have revolutionized is text classification. Current text classification paradigms, however, rely solely on the output of the final layer in the LLM, with the rich information contained in internal neurons largely untapped. In this study, we present SPIN: a model-agnostic framework that sparsifies and integrates internal neurons of intermediate layers of LLMs for text classification. Specifically, SPIN sparsifies internal neurons by linear probing-based salient neuron selection layer by layer, avoiding noise from unrelated neurons and ensuring efficiency. The cross-layer salient neurons are then integrated to serve as multi-layered features for the classification head. Extensive experimental results show our proposed SPIN significantly improves text classification accuracy, efficiency, and interpretability.",
}
| Among the many tasks that Large Language Models (LLMs) have revolutionized is text classification. Current text classification paradigms, however, rely solely on the output of the final layer in the LLM, with the rich information contained in internal neurons largely untapped. In this study, we present SPIN: a model-agnostic framework that sparsifies and integrates internal neurons of intermediate layers of LLMs for text classification. Specifically, SPIN sparsifies internal neurons by linear probing-based salient neuron selection layer by layer, avoiding noise from unrelated neurons and ensuring efficiency. The cross-layer salient neurons are then integrated to serve as multi-layered features for the classification head. Extensive experimental results show our proposed SPIN significantly improves text classification accuracy, efficiency, and interpretability. | [
"Jiao, Difan",
"Liu, Yilun",
"Tang, Zhenwei",
"Matter, Daniel",
"Pfeffer, J{\\\"u}rgen",
"Anderson, Ashton"
]
| {SPIN}: Sparsifying and Integrating Internal Neurons in Large Language Models for Text Classification | findings-acl.277 | Poster | 2311.15983v2 |
https://aclanthology.org/2024.findings-acl.278.bib | @inproceedings{maeda-etal-2024-decomposing,
title = "Decomposing Co-occurrence Matrices into Interpretable Components as Formal Concepts",
author = "Maeda, Akihiro and
Torii, Takuma and
Hidaka, Shohei",
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.278",
pages = "4683--4700",
abstract = "This study addresses the interpretability of word representations through an investigation of a count-based co-occurrence matrix. Employing the mathematical methodology of Formal Concept Analysis, we reveal an underlying structure that is amenable to human interpretation. Furthermore, we unveil the emergence of hierarchical and geometrical structures within word vectors as consequences of word usage. Our experiments on the PPMI matrix demonstrate that the formal concepts that we identified align with interpretable categories, as shown in the category completion task.",
}
| This study addresses the interpretability of word representations through an investigation of a count-based co-occurrence matrix. Employing the mathematical methodology of Formal Concept Analysis, we reveal an underlying structure that is amenable to human interpretation. Furthermore, we unveil the emergence of hierarchical and geometrical structures within word vectors as consequences of word usage. Our experiments on the PPMI matrix demonstrate that the formal concepts that we identified align with interpretable categories, as shown in the category completion task. | [
"Maeda, Akihiro",
"Torii, Takuma",
"Hidaka, Shohei"
]
| Decomposing Co-occurrence Matrices into Interpretable Components as Formal Concepts | findings-acl.278 | Poster | 2403.16782v1 |
https://aclanthology.org/2024.findings-acl.279.bib | @inproceedings{yang-etal-2024-two,
title = "Two-Pronged Human Evaluation of {C}hat{GPT} Self-Correction in Radiology Report Simplification",
author = "Yang, Ziyu and
Cherian, Santhosh and
Vucetic, Slobodan",
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.279",
pages = "4701--4714",
abstract = "Radiology reports are highly technical documents aimed primarily at doctor-doctor communication. There has been an increasing interest in sharing those reports with patients, necessitating providing them patient-friendly simplifications of the original reports. This study explores the suitability of large language models in automatically generating those simplifications. We examine the usefulness of chain-of-thought and self-correction prompting mechanisms in this domain. We also propose a new evaluation protocol that employs radiologists and laypeople, where radiologists verify the factual correctness of simplifications, and laypeople assess simplicity and comprehension. Our experimental results demonstrate the effectiveness of self-correction prompting in producing high-quality simplifications. Our findings illuminate the preferences of radiologists and laypeople regarding text simplification, informing future research on this topic.",
}
| Radiology reports are highly technical documents aimed primarily at doctor-doctor communication. There has been an increasing interest in sharing those reports with patients, necessitating providing them patient-friendly simplifications of the original reports. This study explores the suitability of large language models in automatically generating those simplifications. We examine the usefulness of chain-of-thought and self-correction prompting mechanisms in this domain. We also propose a new evaluation protocol that employs radiologists and laypeople, where radiologists verify the factual correctness of simplifications, and laypeople assess simplicity and comprehension. Our experimental results demonstrate the effectiveness of self-correction prompting in producing high-quality simplifications. Our findings illuminate the preferences of radiologists and laypeople regarding text simplification, informing future research on this topic. | [
"Yang, Ziyu",
"Cherian, Santhosh",
"Vucetic, Slobodan"
]
| Two-Pronged Human Evaluation of {C}hat{GPT} Self-Correction in Radiology Report Simplification | findings-acl.279 | Poster | 2406.18859v1 |
https://aclanthology.org/2024.findings-acl.280.bib | @inproceedings{li-zhang-2024-planning,
title = "Planning First, Question Second: An {LLM}-Guided Method for Controllable Question Generation",
author = "Li, Kunze and
Zhang, Yu",
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.280",
pages = "4715--4729",
abstract = "In the field of education, for better assessment of students{'} abilities, generated questions often need to meet experts{'} requirements, indicating the need for controllable question generation (CQG). However, current CQG methods mainly focus on difficulty control, neglecting the control of question content and assessed abilities, which are also crucial in educational QG. In this paper, we propose an LLM-guided method PFQS (for Planning First, Question Second), which utilizes Llama 2 to generate an answer plan and then generates questions based on it. The plan not only includes candidate answers but also integrates LLM{'}s understanding and multiple requirements, which make question generation simple and controllable. We evaluate our approach on the FairytaleQA dataset, a well-structured QA dataset derived from child-friendly storybooks. In the dataset, the attribute label represents content control, while the local{\_}or{\_}sum and ex{\_}or{\_}im labels denote difficulty control. Experimental results demonstrate that our approach outperforms previous state-of-the-art results and achieves better consistency with requirements compared to prompt-based method. Further application of our method to Llama 2 and Mistral also leads to improved requirement consistency in a zero-shot setting.",
}
| In the field of education, for better assessment of students{'} abilities, generated questions often need to meet experts{'} requirements, indicating the need for controllable question generation (CQG). However, current CQG methods mainly focus on difficulty control, neglecting the control of question content and assessed abilities, which are also crucial in educational QG. In this paper, we propose an LLM-guided method PFQS (for Planning First, Question Second), which utilizes Llama 2 to generate an answer plan and then generates questions based on it. The plan not only includes candidate answers but also integrates LLM{'}s understanding and multiple requirements, which make question generation simple and controllable. We evaluate our approach on the FairytaleQA dataset, a well-structured QA dataset derived from child-friendly storybooks. In the dataset, the attribute label represents content control, while the local{\_}or{\_}sum and ex{\_}or{\_}im labels denote difficulty control. Experimental results demonstrate that our approach outperforms previous state-of-the-art results and achieves better consistency with requirements compared to prompt-based method. Further application of our method to Llama 2 and Mistral also leads to improved requirement consistency in a zero-shot setting. | [
"Li, Kunze",
"Zhang, Yu"
]
| Planning First, Question Second: An {LLM}-Guided Method for Controllable Question Generation | findings-acl.280 | Poster | 1807.03586v5 |
https://aclanthology.org/2024.findings-acl.281.bib | @inproceedings{liu-etal-2024-ra,
title = "{RA}-{ISF}: Learning to Answer and Understand from Retrieval Augmentation via Iterative Self-Feedback",
author = "Liu, Yanming and
Peng, Xinyue and
Zhang, Xuhong and
Liu, Weihao and
Yin, Jianwei and
Cao, Jiannan and
Du, Tianyu",
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.281",
pages = "4730--4749",
abstract = "Large language models (LLMs) demonstrate exceptional performance in numerous tasks but still heavily rely on knowledge stored in their parameters. Moreover, updating this knowledge incurs high training costs. Retrieval-augmented generation (RAG) methods address this issue by integrating external knowledge. The model can answer questions it couldn{'}t previously by retrieving knowledge relevant to the query. This approach improves performance in certain scenarios for specific tasks. However, if irrelevant texts are retrieved, it may impair model performance. In this paper, we propose Retrieval Augmented Iterative Self-Feedback (RA-ISF), a framework that iteratively decomposes tasks and processes them in three submodules to enhance the model{'}s problem-solving capabilities. Experiments show that our method outperforms existing benchmarks, performing well on models like GPT3.5, Llama2, significantly enhancing factual reasoning capabilities and reducing hallucinations.",
}
| Large language models (LLMs) demonstrate exceptional performance in numerous tasks but still heavily rely on knowledge stored in their parameters. Moreover, updating this knowledge incurs high training costs. Retrieval-augmented generation (RAG) methods address this issue by integrating external knowledge. The model can answer questions it couldn{'}t previously by retrieving knowledge relevant to the query. This approach improves performance in certain scenarios for specific tasks. However, if irrelevant texts are retrieved, it may impair model performance. In this paper, we propose Retrieval Augmented Iterative Self-Feedback (RA-ISF), a framework that iteratively decomposes tasks and processes them in three submodules to enhance the model{'}s problem-solving capabilities. Experiments show that our method outperforms existing benchmarks, performing well on models like GPT3.5, Llama2, significantly enhancing factual reasoning capabilities and reducing hallucinations. | [
"Liu, Yanming",
"Peng, Xinyue",
"Zhang, Xuhong",
"Liu, Weihao",
"Yin, Jianwei",
"Cao, Jiannan",
"Du, Tianyu"
]
| {RA}-{ISF}: Learning to Answer and Understand from Retrieval Augmentation via Iterative Self-Feedback | findings-acl.281 | Poster | 2403.06840v2 |
https://aclanthology.org/2024.findings-acl.282.bib | @inproceedings{khamnuansin-etal-2024-mrrank,
title = "{M}r{R}ank: Improving Question Answering Retrieval System through Multi-Result Ranking Model",
author = "Khamnuansin, Danupat and
Chalothorn, Tawunrat and
Chuangsuwanich, Ekapol",
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.282",
pages = "4750--4762",
abstract = "Large Language Models (LLMs) often struggle with hallucinations and outdated information. To address this, Information Retrieval (IR) systems can be employed to augment LLMs with up-to-date knowledge. However, existing IR techniques contain deficiencies, posing a performance bottleneck. Given the extensive array of IR systems, combining diverse approaches presents a viable strategy. Nevertheless, prior attempts have yielded restricted efficacy. In this work, we propose an approach that leverages learning-to-rank techniques to combine heterogeneous IR systems. We demonstrate the method on two Retrieval Question Answering (ReQA) tasks. Our empirical findings exhibit a significant performance enhancement, outperforming previous approaches and achieving state-of-the-art results on ReQA SQuAD.",
}
| Large Language Models (LLMs) often struggle with hallucinations and outdated information. To address this, Information Retrieval (IR) systems can be employed to augment LLMs with up-to-date knowledge. However, existing IR techniques contain deficiencies, posing a performance bottleneck. Given the extensive array of IR systems, combining diverse approaches presents a viable strategy. Nevertheless, prior attempts have yielded restricted efficacy. In this work, we propose an approach that leverages learning-to-rank techniques to combine heterogeneous IR systems. We demonstrate the method on two Retrieval Question Answering (ReQA) tasks. Our empirical findings exhibit a significant performance enhancement, outperforming previous approaches and achieving state-of-the-art results on ReQA SQuAD. | [
"Khamnuansin, Danupat",
"Chalothorn, Tawunrat",
"Chuangsuwanich, Ekapol"
]
| {M}r{R}ank: Improving Question Answering Retrieval System through Multi-Result Ranking Model | findings-acl.282 | Poster | 2403.17611v1 |
https://aclanthology.org/2024.findings-acl.283.bib | @inproceedings{yixing-etal-2024-chain,
title = "Chain-of-Question: A Progressive Question Decomposition Approach for Complex Knowledge Base Question Answering",
author = "Yixing, Peng and
Wang, Quan and
Zhang, Licheng and
Liu, Yi and
Mao, Zhendong",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.283",
pages = "4763--4776",
abstract = "Complex KBQA leverages the knowledge base (KB) to answer complex natural questions involving complicated semantics like multi-hop reasoning. Existing methods involve a question decomposition process, i.e., breaking a complex question into several simpler sub-questions, to assist obtaining logical forms for querying the KB. However, existing question decomposition process derives all sub-questions directly according to the original question, resulting in limitations when one sub-question relies on the answer from a previous one. In this work, we propose Chain-of-Question, a progressive question decomposition approach to address complex KBQA challenges. First, inspired by chain-of-thought, we design a prompt to guide LLM to sequentially decompose multiple semantically clear sub-questions and provide corresponding reference answers, where each step of the decomposition relies on the previous results. Next, we utilize the decomposition result to select relevant patterns (relation-entity pairs) as accurate and faithful auxiliary information for the following logical form generation. Finally, we jointly perform logical form generation and answer prediction, utilizing the predicted answer to supplement non-executable logical forms. Experimental results demonstrate that our method achieves state-of-the-art performance on multiple datasets.",
}
| Complex KBQA leverages the knowledge base (KB) to answer complex natural questions involving complicated semantics like multi-hop reasoning. Existing methods involve a question decomposition process, i.e., breaking a complex question into several simpler sub-questions, to assist obtaining logical forms for querying the KB. However, existing question decomposition process derives all sub-questions directly according to the original question, resulting in limitations when one sub-question relies on the answer from a previous one. In this work, we propose Chain-of-Question, a progressive question decomposition approach to address complex KBQA challenges. First, inspired by chain-of-thought, we design a prompt to guide LLM to sequentially decompose multiple semantically clear sub-questions and provide corresponding reference answers, where each step of the decomposition relies on the previous results. Next, we utilize the decomposition result to select relevant patterns (relation-entity pairs) as accurate and faithful auxiliary information for the following logical form generation. Finally, we jointly perform logical form generation and answer prediction, utilizing the predicted answer to supplement non-executable logical forms. Experimental results demonstrate that our method achieves state-of-the-art performance on multiple datasets. | [
"Yixing, Peng",
"Wang, Quan",
"Zhang, Licheng",
"Liu, Yi",
"Mao, Zhendong"
]
| Chain-of-Question: A Progressive Question Decomposition Approach for Complex Knowledge Base Question Answering | findings-acl.283 | Poster | 1803.06643v1 |
https://aclanthology.org/2024.findings-acl.284.bib | @inproceedings{zheng-etal-2024-instruction,
title = "Instruction Tuning with Retrieval-based Examples Ranking for Aspect-based Sentiment Analysis",
author = "Zheng, Guangmin and
Wang, Jin and
Yu, Liang-Chih and
Zhang, Xuejie",
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.284",
pages = "4777--4788",
abstract = "Aspect-based sentiment analysis (ABSA) identifies sentiment information related to specific aspects and provides deeper market insights to businesses and organizations. With the emergence of large language models (LMs), recent studies have proposed using fixed examples for instruction tuning to reformulate ABSA as a generation task. However, the performance is sensitive to the selection of in-context examples; several retrieval methods are based on surface similarity and are independent of the LM generative objective. This study proposes an instruction learning method with retrieval-based example ranking for ABSA tasks. For each target sample, an LM was applied as a scorer to estimate the likelihood of the output given the input and a candidate example as the prompt, and training examples were labeled as positive or negative by ranking the scores. An alternating training schema is proposed to train both the retriever and LM. Instructional prompts can be constructed using high-quality examples. The LM is used for both scoring and inference, improving the generation efficiency without incurring additional computational costs or training difficulties. Extensive experiments on three ABSA subtasks verified the effectiveness of the proposed method, demonstrating its superiority over various strong baseline models. Code and data are released at https://github.com/zgMin/IT-RER-ABSA.",
}
| Aspect-based sentiment analysis (ABSA) identifies sentiment information related to specific aspects and provides deeper market insights to businesses and organizations. With the emergence of large language models (LMs), recent studies have proposed using fixed examples for instruction tuning to reformulate ABSA as a generation task. However, the performance is sensitive to the selection of in-context examples; several retrieval methods are based on surface similarity and are independent of the LM generative objective. This study proposes an instruction learning method with retrieval-based example ranking for ABSA tasks. For each target sample, an LM was applied as a scorer to estimate the likelihood of the output given the input and a candidate example as the prompt, and training examples were labeled as positive or negative by ranking the scores. An alternating training schema is proposed to train both the retriever and LM. Instructional prompts can be constructed using high-quality examples. The LM is used for both scoring and inference, improving the generation efficiency without incurring additional computational costs or training difficulties. Extensive experiments on three ABSA subtasks verified the effectiveness of the proposed method, demonstrating its superiority over various strong baseline models. Code and data are released at https://github.com/zgMin/IT-RER-ABSA. | [
"Zheng, Guangmin",
"Wang, Jin",
"Yu, Liang-Chih",
"Zhang, Xuejie"
]
| Instruction Tuning with Retrieval-based Examples Ranking for Aspect-based Sentiment Analysis | findings-acl.284 | Poster | 2405.18035v2 |
https://aclanthology.org/2024.findings-acl.285.bib | @inproceedings{mou-etal-2024-unveiling,
title = "Unveiling the Truth and Facilitating Change: Towards Agent-based Large-scale Social Movement Simulation",
author = "Mou, Xinyi and
Wei, Zhongyu 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.285",
pages = "4789--4809",
abstract = "Social media has emerged as a cornerstone of social movements, wielding significant influence in driving societal change. Simulating the response of the public and forecasting the potential impact has become increasingly important. However, existing methods for simulating such phenomena encounter challenges concerning their efficacy and efficiency in capturing the behaviors of social movement participants. In this paper, we introduce a hybrid framework for social media user simulation, wherein users are categorized into two types. Core users are driven by Large Language Models, while numerous ordinary users are modeled by deductive agent-based models. We further construct a Twitter-like environment to replicate their response dynamics following trigger events. Subsequently, we develop a multi-faceted benchmark SoMoSiMu-Bench for evaluation and conduct comprehensive experiments across real-world datasets. Experimental results demonstrate the effectiveness and flexibility of our method.",
}
| Social media has emerged as a cornerstone of social movements, wielding significant influence in driving societal change. Simulating the response of the public and forecasting the potential impact has become increasingly important. However, existing methods for simulating such phenomena encounter challenges concerning their efficacy and efficiency in capturing the behaviors of social movement participants. In this paper, we introduce a hybrid framework for social media user simulation, wherein users are categorized into two types. Core users are driven by Large Language Models, while numerous ordinary users are modeled by deductive agent-based models. We further construct a Twitter-like environment to replicate their response dynamics following trigger events. Subsequently, we develop a multi-faceted benchmark SoMoSiMu-Bench for evaluation and conduct comprehensive experiments across real-world datasets. Experimental results demonstrate the effectiveness and flexibility of our method. | [
"Mou, Xinyi",
"Wei, Zhongyu",
"Huang, Xuanjing"
]
| Unveiling the Truth and Facilitating Change: Towards Agent-based Large-scale Social Movement Simulation | findings-acl.285 | Poster | 2402.16333v2 |
https://aclanthology.org/2024.findings-acl.286.bib | @inproceedings{kanayama-etal-2024-incorporating,
title = "Incorporating Syntax and Lexical Knowledge to Multilingual Sentiment Classification on Large Language Models",
author = "Kanayama, Hiroshi and
Zhao, Yang and
Iwamoto, Ran and
Ohko, Takuya",
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.286",
pages = "4810--4817",
abstract = "This paper exploits a sentiment extractor supported by syntactic and lexical resources to enhance multilingual sentiment classification solved through the generative approach, without retraining LLMs. By adding external information of words and phrases that have positive/negative polarities, the multilingual sentiment classification error was reduced by up to 33 points, and the combination of two approaches performed best especially in high-performing pairs of LLMs and languages.",
}
| This paper exploits a sentiment extractor supported by syntactic and lexical resources to enhance multilingual sentiment classification solved through the generative approach, without retraining LLMs. By adding external information of words and phrases that have positive/negative polarities, the multilingual sentiment classification error was reduced by up to 33 points, and the combination of two approaches performed best especially in high-performing pairs of LLMs and languages. | [
"Kanayama, Hiroshi",
"Zhao, Yang",
"Iwamoto, Ran",
"Ohko, Takuya"
]
| Incorporating Syntax and Lexical Knowledge to Multilingual Sentiment Classification on Large Language Models | findings-acl.286 | Poster | 2404.16627v1 |
https://aclanthology.org/2024.findings-acl.287.bib | @inproceedings{wang-etal-2024-locating,
title = "Locating and Extracting Relational Concepts in Large Language Models",
author = "Wang, Zijian and
Whyte, Britney and
Xu, Chang",
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.287",
pages = "4818--4832",
abstract = "Relational concepts are indeed foundational to the structure of knowledge representation, as they facilitate the association between various entity concepts, allowing us to express and comprehend complex world knowledge.By expressing relational concepts in natural language prompts, people can effortlessly interact with large language models (LLMs) and recall desired factual knowledge. However, the process of knowledge recall lacks interpretability, and representations of relational concepts within LLMs remain unknown to us. In this paper, we identify hidden states that can express entity and relational concepts through causal mediation analysis in fact recall processes. Our finding reveals that at the last token position of the input prompt, there are hidden states that solely express the causal effects of relational concepts. Based on this finding, we assume that these hidden states can be treated as relational representations and we can successfully extract them from LLMs. The experimental results demonstrate high credibility of the relational representations: they can be flexibly transplanted into other fact recall processes, and can also be used as robust entity connectors. Moreover, we also show that the relational representations exhibit significant potential for controllable fact recall through relation rewriting.",
}
| Relational concepts are indeed foundational to the structure of knowledge representation, as they facilitate the association between various entity concepts, allowing us to express and comprehend complex world knowledge.By expressing relational concepts in natural language prompts, people can effortlessly interact with large language models (LLMs) and recall desired factual knowledge. However, the process of knowledge recall lacks interpretability, and representations of relational concepts within LLMs remain unknown to us. In this paper, we identify hidden states that can express entity and relational concepts through causal mediation analysis in fact recall processes. Our finding reveals that at the last token position of the input prompt, there are hidden states that solely express the causal effects of relational concepts. Based on this finding, we assume that these hidden states can be treated as relational representations and we can successfully extract them from LLMs. The experimental results demonstrate high credibility of the relational representations: they can be flexibly transplanted into other fact recall processes, and can also be used as robust entity connectors. Moreover, we also show that the relational representations exhibit significant potential for controllable fact recall through relation rewriting. | [
"Wang, Zijian",
"Whyte, Britney",
"Xu, Chang"
]
| Locating and Extracting Relational Concepts in Large Language Models | findings-acl.287 | Poster | 2406.13184v1 |
https://aclanthology.org/2024.findings-acl.288.bib | @inproceedings{li-etal-2024-unraveling,
title = "Unraveling and Mitigating Retriever Inconsistencies in Retrieval-Augmented Large Language Models",
author = "Li, Mingda and
Li, Xinyu and
Chen, Yifan and
Xuan, Wenfeng and
Zhang, Weinan",
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.288",
pages = "4833--4850",
abstract = "Although Retrieval-Augmented Large Language Models (RALMs) demonstrate their superiority in terms of factuality, they do not consistently outperform the original retrieval-free Language Models (LMs). Our experiments reveal that this example-level performance inconsistency exists not only between retrieval-augmented and retrieval-free LM but also among different retrievers. To understand this phenomenon, we investigate the degeneration behavior of RALMs and theoretically decompose it into four categories. Further analysis based on our decomposition reveals that the innate difference in knowledge sources and the unpredictable degeneration of the reader model contribute most to the inconsistency. Drawing from our analysis, we introduce Ensemble of Retrievers (EoR), a trainable framework that can adaptively retrieve from different knowledge sources and effectively decrease unpredictable reader errors. Our experiments on Open Domain Question Answering show that EoR substantially improves performance over the RALM with a single retriever by considerably reducing inconsistent behaviors.",
}
| Although Retrieval-Augmented Large Language Models (RALMs) demonstrate their superiority in terms of factuality, they do not consistently outperform the original retrieval-free Language Models (LMs). Our experiments reveal that this example-level performance inconsistency exists not only between retrieval-augmented and retrieval-free LM but also among different retrievers. To understand this phenomenon, we investigate the degeneration behavior of RALMs and theoretically decompose it into four categories. Further analysis based on our decomposition reveals that the innate difference in knowledge sources and the unpredictable degeneration of the reader model contribute most to the inconsistency. Drawing from our analysis, we introduce Ensemble of Retrievers (EoR), a trainable framework that can adaptively retrieve from different knowledge sources and effectively decrease unpredictable reader errors. Our experiments on Open Domain Question Answering show that EoR substantially improves performance over the RALM with a single retriever by considerably reducing inconsistent behaviors. | [
"Li, Mingda",
"Li, Xinyu",
"Chen, Yifan",
"Xuan, Wenfeng",
"Zhang, Weinan"
]
| Unraveling and Mitigating Retriever Inconsistencies in Retrieval-Augmented Large Language Models | findings-acl.288 | Poster | 2405.20680v3 |
https://aclanthology.org/2024.findings-acl.289.bib | @inproceedings{zhang-etal-2024-senticvec,
title = "{S}entic{V}ec: Toward Robust and Human-Centric Neurosymbolic Sentiment Analysis",
author = "Zhang, Xulang and
Mao, Rui and
Cambria, Erik",
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.289",
pages = "4851--4863",
abstract = "The success of state-of-the-art Natural Language Processing (NLP) systems heavily depends on deep neural networks, which excel in various tasks through strong data fitting and latent feature modeling abilities. However, certain challenges linked to deep neural networks and supervised deep learning deserve considerations, e.g., extensive computing resources, knowledge forgetting, etc. Previous research attempted to tackle these challenges individually through irrelative techniques. However, they do not instigate fundamental shifts in the learning paradigm. In this work, we propose a novel neurosymbolic method for sentiment analysis to tackle these issues. We also propose a novel sentiment-pragmatic knowledge base that places emphasis on human subjectivity within varying domain annotations. We conducted extensive experiments to show that our neurosymbolic framework for sentiment analysis stands out for its lightweight nature, robustness across domains and languages, efficient few-shot training, and rapid convergence.",
}
| The success of state-of-the-art Natural Language Processing (NLP) systems heavily depends on deep neural networks, which excel in various tasks through strong data fitting and latent feature modeling abilities. However, certain challenges linked to deep neural networks and supervised deep learning deserve considerations, e.g., extensive computing resources, knowledge forgetting, etc. Previous research attempted to tackle these challenges individually through irrelative techniques. However, they do not instigate fundamental shifts in the learning paradigm. In this work, we propose a novel neurosymbolic method for sentiment analysis to tackle these issues. We also propose a novel sentiment-pragmatic knowledge base that places emphasis on human subjectivity within varying domain annotations. We conducted extensive experiments to show that our neurosymbolic framework for sentiment analysis stands out for its lightweight nature, robustness across domains and languages, efficient few-shot training, and rapid convergence. | [
"Zhang, Xulang",
"Mao, Rui",
"Cambria, Erik"
]
| {S}entic{V}ec: Toward Robust and Human-Centric Neurosymbolic Sentiment Analysis | findings-acl.289 | Poster | 2208.09329v1 |
https://aclanthology.org/2024.findings-acl.290.bib | @inproceedings{qian-etal-2024-towards,
title = "Towards Tracing Trustworthiness Dynamics: Revisiting Pre-training Period of Large Language Models",
author = "Qian, Chen and
Zhang, Jie and
Yao, Wei and
Liu, Dongrui and
Yin, Zhenfei and
Qiao, Yu and
Liu, Yong and
Shao, Jing",
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.290",
pages = "4864--4888",
abstract = "Ensuring the trustworthiness of large language models (LLMs) is crucial. Most studies concentrate on fully pre-trained LLMs to better understand and improve LLMs{'} trustworthiness. In this paper, to reveal the untapped potential of pre-training, we pioneer the exploration of LLMs{'} trustworthiness during this period, focusing on five key dimensions: reliability, privacy, toxicity, fairness, and robustness. To begin with, we apply linear probing to LLMs. The high probing accuracy suggests that \textit{LLMs in early pre-training can already distinguish concepts in each trustworthiness dimension}. Therefore, to further uncover the hidden possibilities of pre-training, we extract steering vectors from a LLM{'}s pre-training checkpoints to enhance the LLM{'}s trustworthiness. Finally, inspired by the theoretical result that mutual information estimation is bounded by linear probing accuracy, we also probe LLMs with mutual information to investigate the dynamics of trustworthiness during pre-training. We are the first to observe a similar two-phase phenomenon: fitting and compression. This research provides an initial exploration of trustworthiness modeling during LLM pre-training, seeking to unveil new insights and spur further developments in the field.",
}
| Ensuring the trustworthiness of large language models (LLMs) is crucial. Most studies concentrate on fully pre-trained LLMs to better understand and improve LLMs{'} trustworthiness. In this paper, to reveal the untapped potential of pre-training, we pioneer the exploration of LLMs{'} trustworthiness during this period, focusing on five key dimensions: reliability, privacy, toxicity, fairness, and robustness. To begin with, we apply linear probing to LLMs. The high probing accuracy suggests that \textit{LLMs in early pre-training can already distinguish concepts in each trustworthiness dimension}. Therefore, to further uncover the hidden possibilities of pre-training, we extract steering vectors from a LLM{'}s pre-training checkpoints to enhance the LLM{'}s trustworthiness. Finally, inspired by the theoretical result that mutual information estimation is bounded by linear probing accuracy, we also probe LLMs with mutual information to investigate the dynamics of trustworthiness during pre-training. We are the first to observe a similar two-phase phenomenon: fitting and compression. This research provides an initial exploration of trustworthiness modeling during LLM pre-training, seeking to unveil new insights and spur further developments in the field. | [
"Qian, Chen",
"Zhang, Jie",
"Yao, Wei",
"Liu, Dongrui",
"Yin, Zhenfei",
"Qiao, Yu",
"Liu, Yong",
"Shao, Jing"
]
| Towards Tracing Trustworthiness Dynamics: Revisiting Pre-training Period of Large Language Models | findings-acl.290 | Poster | 2402.19465v1 |
https://aclanthology.org/2024.findings-acl.291.bib | @inproceedings{xia-etal-2024-language,
title = "Language Models can Evaluate Themselves via Probability Discrepancy",
author = "Xia, Tingyu and
Yu, Bowen and
Wu, Yuan and
Chang, Yi and
Zhou, Chang",
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.291",
pages = "4889--4901",
abstract = "In this paper, we begin by illustrating that, when presented with a query, Large Language Models (LLMs) capable of providing accurate responses tend to exhibit a more uniform probability distribution compared to their less proficient counterparts. Building upon this observation, we introduce a novel self-assessment criterion termed ProbDiff for evaluating the performance of diverse LLMs. This method eliminates the need for training an additional evaluation model or relying on external proprietary models such as GPT-4 as a judger. Instead, it solely relies on the LLMs under evaluation to compute the probability discrepancy between the original response generation and its revised versions. A higher discrepancy in two LLMs for the same query suggests a relatively weaker ability. We discover that ProbDiff yields comparable results to mainstream GPT-4-based evaluations on various scenarios including NLG tasks like translation and summarization, as well as LLM evaluation benchmarks such as AlignBench, MT-Bench, and AlpacaEval, across LLMs of different sizes.",
}
| In this paper, we begin by illustrating that, when presented with a query, Large Language Models (LLMs) capable of providing accurate responses tend to exhibit a more uniform probability distribution compared to their less proficient counterparts. Building upon this observation, we introduce a novel self-assessment criterion termed ProbDiff for evaluating the performance of diverse LLMs. This method eliminates the need for training an additional evaluation model or relying on external proprietary models such as GPT-4 as a judger. Instead, it solely relies on the LLMs under evaluation to compute the probability discrepancy between the original response generation and its revised versions. A higher discrepancy in two LLMs for the same query suggests a relatively weaker ability. We discover that ProbDiff yields comparable results to mainstream GPT-4-based evaluations on various scenarios including NLG tasks like translation and summarization, as well as LLM evaluation benchmarks such as AlignBench, MT-Bench, and AlpacaEval, across LLMs of different sizes. | [
"Xia, Tingyu",
"Yu, Bowen",
"Wu, Yuan",
"Chang, Yi",
"Zhou, Chang"
]
| Language Models can Evaluate Themselves via Probability Discrepancy | findings-acl.291 | Poster | 2405.10516v2 |
https://aclanthology.org/2024.findings-acl.292.bib | @inproceedings{zhou-etal-2024-evaluating-validity,
title = "Evaluating the Validity of Word-level Adversarial Attacks with Large Language Models",
author = "Zhou, Huichi and
Wang, Zhaoyang and
Wang, Hongtao and
Chen, Dongping and
Mu, Wenhan and
Zhang, Fangyuan",
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.292",
pages = "4902--4922",
abstract = "Deep neural networks exhibit vulnerability to word-level adversarial attacks in natural language processing. Most of these attack methods adopt synonymous substitutions to perturb original samples for crafting adversarial examples while attempting to maintain semantic consistency with the originals. Some of them claim that they could achieve over 90{\%} attack success rate, thereby raising serious safety concerns. However, our investigation reveals that many purportedly successful adversarial examples are actually invalid due to significant changes in semantic meanings compared to their originals. Even when equipped with semantic constraints such as BERTScore, existing attack methods can generate up to 87.9{\%} invalid adversarial examples. Building on this insight, we first curate a 13K dataset for adversarial validity evaluation with the help of GPT-4. Then, an open-source large language model is fine-tuned to offer an interpretable validity score for assessing the semantic consistency between original and adversarial examples. Finally, this validity score can serve as a guide for existing adversarial attack methods to generate valid adversarial examples. Comprehensive experiments demonstrate the effectiveness of our method in evaluating and refining the quality of adversarial examples.",
}
| Deep neural networks exhibit vulnerability to word-level adversarial attacks in natural language processing. Most of these attack methods adopt synonymous substitutions to perturb original samples for crafting adversarial examples while attempting to maintain semantic consistency with the originals. Some of them claim that they could achieve over 90{\%} attack success rate, thereby raising serious safety concerns. However, our investigation reveals that many purportedly successful adversarial examples are actually invalid due to significant changes in semantic meanings compared to their originals. Even when equipped with semantic constraints such as BERTScore, existing attack methods can generate up to 87.9{\%} invalid adversarial examples. Building on this insight, we first curate a 13K dataset for adversarial validity evaluation with the help of GPT-4. Then, an open-source large language model is fine-tuned to offer an interpretable validity score for assessing the semantic consistency between original and adversarial examples. Finally, this validity score can serve as a guide for existing adversarial attack methods to generate valid adversarial examples. Comprehensive experiments demonstrate the effectiveness of our method in evaluating and refining the quality of adversarial examples. | [
"Zhou, Huichi",
"Wang, Zhaoyang",
"Wang, Hongtao",
"Chen, Dongping",
"Mu, Wenhan",
"Zhang, Fangyuan"
]
| Evaluating the Validity of Word-level Adversarial Attacks with Large Language Models | findings-acl.292 | Poster | 2311.11861v1 |
https://aclanthology.org/2024.findings-acl.293.bib | @inproceedings{zhao-etal-2024-language,
title = "On the Language Encoder of Contrastive Cross-modal Models",
author = "Zhao, Mengjie and
Ono, Junya and
Zhong, Zhi and
Lai, Chieh-Hsin and
Takida, Yuhta and
Murata, Naoki and
Liao, Wei-Hsiang and
Shibuya, Takashi and
Wakaki, Hiromi and
Mitsufuji, Yuki",
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.293",
pages = "4923--4940",
abstract = "Contrastive cross-modal models such as CLIP and CLAP aid various vision-language (VL) and audio-language (AL) tasks. However, there has been limited investigation of and improvement in their language encoder {--} the central component of encoding natural language descriptions of image/audio into vector representations. We extensively evaluate how unsupervised and supervised sentence embedding training affect language encoder quality and cross-modal task performance. In VL pretraining, we found that sentence embedding training enhances language encoder quality and aids in cross-modal tasks, improving contrastive VL models such as CyCLIP. Sentence embedding training benefits AL tasks when the amount of training data is large. We analyze the representation spaces to understand the strengths of sentence embedding training, and find that it improves text-space uniformity, at the cost of decreased cross-modal alignment.",
}
| Contrastive cross-modal models such as CLIP and CLAP aid various vision-language (VL) and audio-language (AL) tasks. However, there has been limited investigation of and improvement in their language encoder {--} the central component of encoding natural language descriptions of image/audio into vector representations. We extensively evaluate how unsupervised and supervised sentence embedding training affect language encoder quality and cross-modal task performance. In VL pretraining, we found that sentence embedding training enhances language encoder quality and aids in cross-modal tasks, improving contrastive VL models such as CyCLIP. Sentence embedding training benefits AL tasks when the amount of training data is large. We analyze the representation spaces to understand the strengths of sentence embedding training, and find that it improves text-space uniformity, at the cost of decreased cross-modal alignment. | [
"Zhao, Mengjie",
"Ono, Junya",
"Zhong, Zhi",
"Lai, Chieh-Hsin",
"Takida, Yuhta",
"Murata, Naoki",
"Liao, Wei-Hsiang",
"Shibuya, Takashi",
"Wakaki, Hiromi",
"Mitsufuji, Yuki"
]
| On the Language Encoder of Contrastive Cross-modal Models | findings-acl.293 | Poster | 2310.13267v1 |
https://aclanthology.org/2024.findings-acl.294.bib | @inproceedings{wu-etal-2024-co,
title = "Your Co-Workers Matter: Evaluating Collaborative Capabilities of Language Models in Blocks World",
author = "Wu, Guande and
Zhao, Chen and
Silva, Claudio and
He, He",
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.294",
pages = "4941--4957",
abstract = "Language agents that interact with the world on their own have great potential for automating digital tasks. While large language model (LLM) agents have made progress in understanding and executing tasks such as textual games and webpage control, many real-world tasks also require collaboration with humans or other LLMs in equal roles, which involves intent understanding, task coordination, and communication. To test LLM{'}s ability to collaborate, we design a blocks-world environment, where two agents, each having unique goals and skills, build a target structure together. To complete the goals, they can act in the world and communicate in natural language. Under this environment, we design increasingly challenging settings to evaluate different collaboration perspectives, from independent to more complex, dependent tasks. We further adopt chain-of-thought prompts that include intermediate reasoning steps to model the partner{'}s state and identify and correct execution errors. Both human-machine and machine-machine experiments show that LLM agents have strong grounding capacities, and our approach significantly improves the evaluation metric.",
}
| Language agents that interact with the world on their own have great potential for automating digital tasks. While large language model (LLM) agents have made progress in understanding and executing tasks such as textual games and webpage control, many real-world tasks also require collaboration with humans or other LLMs in equal roles, which involves intent understanding, task coordination, and communication. To test LLM{'}s ability to collaborate, we design a blocks-world environment, where two agents, each having unique goals and skills, build a target structure together. To complete the goals, they can act in the world and communicate in natural language. Under this environment, we design increasingly challenging settings to evaluate different collaboration perspectives, from independent to more complex, dependent tasks. We further adopt chain-of-thought prompts that include intermediate reasoning steps to model the partner{'}s state and identify and correct execution errors. Both human-machine and machine-machine experiments show that LLM agents have strong grounding capacities, and our approach significantly improves the evaluation metric. | [
"Wu, Gu",
"e",
"Zhao, Chen",
"Silva, Claudio",
"He, He"
]
| Your Co-Workers Matter: Evaluating Collaborative Capabilities of Language Models in Blocks World | findings-acl.294 | Poster | 2404.00246v1 |
https://aclanthology.org/2024.findings-acl.295.bib | @inproceedings{pang-etal-2024-anchor,
title = "Anchor-based Large Language Models",
author = "Pang, Jianhui and
Ye, Fanghua and
Wong, Derek and
He, Xin and
Chen, Wanshun and
Wang, Longyue",
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.295",
pages = "4958--4976",
abstract = "Large language models (LLMs) predominantly employ decoder-only transformer architectures, necessitating the retention of keys/values information for historical tokens to provide contextual information and avoid redundant computation. However, the substantial size and parameter volume of these LLMs require massive GPU memory. This memory demand increases with the length of the input text, leading to an urgent need for more efficient methods of information storage and processing. This study introduces Anchor-based LLMs (AnLLMs), which utilize an innovative anchor-based self-attention network (AnSAN) and also an anchor-based inference strategy. This approach enables LLMs to compress sequence information into an anchor token, reducing the keys/values cache and enhancing inference efficiency. Experiments on question-answering benchmarks reveal that AnLLMs maintain similar accuracy levels while achieving up to 99{\%} keys/values cache reduction and up to 3.5 times faster inference. Despite a minor compromise in accuracy, the substantial enhancements of AnLLMs employing the AnSAN technique in resource utilization and computational efficiency underscore their potential for practical LLM applications.",
}
| Large language models (LLMs) predominantly employ decoder-only transformer architectures, necessitating the retention of keys/values information for historical tokens to provide contextual information and avoid redundant computation. However, the substantial size and parameter volume of these LLMs require massive GPU memory. This memory demand increases with the length of the input text, leading to an urgent need for more efficient methods of information storage and processing. This study introduces Anchor-based LLMs (AnLLMs), which utilize an innovative anchor-based self-attention network (AnSAN) and also an anchor-based inference strategy. This approach enables LLMs to compress sequence information into an anchor token, reducing the keys/values cache and enhancing inference efficiency. Experiments on question-answering benchmarks reveal that AnLLMs maintain similar accuracy levels while achieving up to 99{\%} keys/values cache reduction and up to 3.5 times faster inference. Despite a minor compromise in accuracy, the substantial enhancements of AnLLMs employing the AnSAN technique in resource utilization and computational efficiency underscore their potential for practical LLM applications. | [
"Pang, Jianhui",
"Ye, Fanghua",
"Wong, Derek",
"He, Xin",
"Chen, Wanshun",
"Wang, Longyue"
]
| Anchor-based Large Language Models | findings-acl.295 | Poster | 2306.07377v1 |
https://aclanthology.org/2024.findings-acl.296.bib | @inproceedings{xu-etal-2024-mlevlm,
title = "{ML}e{VLM}: Improve Multi-level Progressive Capabilities based on Multimodal Large Language Model for Medical Visual Question Answering",
author = "Xu, Dexuan and
Chen, Yanyuan and
Wang, Jieyi and
Huang, Yue and
Wang, Hanpin and
Jin, Zhi and
Wang, Hongxing and
Yue, Weihua and
He, Jing and
Li, Hang and
Huang, Yu",
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.296",
pages = "4977--4997",
abstract = "Medical visual question answering (MVQA) requires in-depth understanding of medical images and questions to provide reliable answers. We summarize multi-level progressive capabilities that models need to focus on in MVQA: recognition, details, diagnosis, knowledge, and reasoning. Existing MVQA models tend to ignore the above capabilities due to unspecific data and plain architecture. To address these issues, this paper proposes Multi-level Visual Language Model (MLeVLM) for MVQA. On the data side, we construct a high-quality multi-level instruction dataset MLe-VQA via GPT-4, which covers multi-level questions and answers as well as reasoning processes from visual clues to semantic cognition. On the architecture side, we propose a multi-level feature alignment module, including attention-based token selector and context merger, which can efficiently align features at different levels from visual to semantic. To better evaluate the model{'}s capabilities, we manually construct a multi-level MVQA evaluation benchmark named MLe-Bench. Extensive experiments demonstrate the effectiveness of our constructed multi-level instruction dataset and the multi-level feature alignment module. It also proves that MLeVLM outperforms existing medical multimodal large language models.",
}
| Medical visual question answering (MVQA) requires in-depth understanding of medical images and questions to provide reliable answers. We summarize multi-level progressive capabilities that models need to focus on in MVQA: recognition, details, diagnosis, knowledge, and reasoning. Existing MVQA models tend to ignore the above capabilities due to unspecific data and plain architecture. To address these issues, this paper proposes Multi-level Visual Language Model (MLeVLM) for MVQA. On the data side, we construct a high-quality multi-level instruction dataset MLe-VQA via GPT-4, which covers multi-level questions and answers as well as reasoning processes from visual clues to semantic cognition. On the architecture side, we propose a multi-level feature alignment module, including attention-based token selector and context merger, which can efficiently align features at different levels from visual to semantic. To better evaluate the model{'}s capabilities, we manually construct a multi-level MVQA evaluation benchmark named MLe-Bench. Extensive experiments demonstrate the effectiveness of our constructed multi-level instruction dataset and the multi-level feature alignment module. It also proves that MLeVLM outperforms existing medical multimodal large language models. | [
"Xu, Dexuan",
"Chen, Yanyuan",
"Wang, Jieyi",
"Huang, Yue",
"Wang, Hanpin",
"Jin, Zhi",
"Wang, Hongxing",
"Yue, Weihua",
"He, Jing",
"Li, Hang",
"Huang, Yu"
]
| {ML}e{VLM}: Improve Multi-level Progressive Capabilities based on Multimodal Large Language Model for Medical Visual Question Answering | findings-acl.296 | Poster | 2408.03043v1 |
https://aclanthology.org/2024.findings-acl.297.bib | @inproceedings{park-etal-2024-disentangling,
title = "Disentangling Length from Quality in Direct Preference Optimization",
author = "Park, Ryan and
Rafailov, Rafael and
Ermon, Stefano and
Finn, Chelsea",
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.297",
pages = "4998--5017",
abstract = "Reinforcement Learning from Human Feedback (RLHF) has been a crucial component in the recent success of Large Language Models. However, RLHF is know to exploit biases in human preferences, such as verbosity. A well-formatted and eloquent answer is often more highly rated by users, even when it is less helpful and objective. A number of approaches have been developed to control those biases in the classical RLHF literature, but the problem remains relatively under-explored for Direct Alignment Algorithms such as Direct Preference Optimization (DPO). Unlike classical RLHF, DPO does not train a separate reward model or use reinforcement learning directly, so previous approaches developed to control verbosity cannot be directly applied to this setting. Our work makes several contributions. For the first time, we study the length problem in the DPO setting, showing significant exploitation in DPO and linking it to out-of-distribution bootstrapping. We then develop a principled but simple regularization strategy that prevents length exploitation, while still maintaining improvements in model quality. We demonstrate these affects across datasets on summarization and dialogue, where we achieve up to 20{\%} improvement in win rates when controlling for length, despite the GPT4 judge{'}s well-known verbosity bias.",
}
| Reinforcement Learning from Human Feedback (RLHF) has been a crucial component in the recent success of Large Language Models. However, RLHF is know to exploit biases in human preferences, such as verbosity. A well-formatted and eloquent answer is often more highly rated by users, even when it is less helpful and objective. A number of approaches have been developed to control those biases in the classical RLHF literature, but the problem remains relatively under-explored for Direct Alignment Algorithms such as Direct Preference Optimization (DPO). Unlike classical RLHF, DPO does not train a separate reward model or use reinforcement learning directly, so previous approaches developed to control verbosity cannot be directly applied to this setting. Our work makes several contributions. For the first time, we study the length problem in the DPO setting, showing significant exploitation in DPO and linking it to out-of-distribution bootstrapping. We then develop a principled but simple regularization strategy that prevents length exploitation, while still maintaining improvements in model quality. We demonstrate these affects across datasets on summarization and dialogue, where we achieve up to 20{\%} improvement in win rates when controlling for length, despite the GPT4 judge{'}s well-known verbosity bias. | [
"Park, Ryan",
"Rafailov, Rafael",
"Ermon, Stefano",
"Finn, Chelsea"
]
| Disentangling Length from Quality in Direct Preference Optimization | findings-acl.297 | Poster | 2403.19159v1 |
https://aclanthology.org/2024.findings-acl.298.bib | @inproceedings{li-etal-2024-mike,
title = "{MIKE}: A New Benchmark for Fine-grained Multimodal Entity Knowledge Editing",
author = "Li, Jiaqi and
Du, Miaozeng and
Zhang, Chuanyi and
Chen, Yongrui and
Hu, Nan and
Qi, Guilin and
Jiang, Haiyun and
Cheng, Siyuan and
Tian, Bozhong",
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.298",
pages = "5018--5029",
abstract = "Multimodal knowledge editing represents a critical advancement in enhancing the capabilities of Multimodal Large Language Models (MLLMs). Despite its potential, current benchmarks predominantly focus on coarse-grained knowledge, leaving the intricacies of fine-grained (FG) multimodal entity knowledge largely unexplored. This gap presents a notable challenge, as FG entity recognition is pivotal for the practical deployment and effectiveness of MLLMs in diverse real-world scenarios. To bridge this gap, we introduce MIKE, a comprehensive benchmark and dataset specifically designed for the FG multimodal entity knowledge editing. MIKE encompasses a suite of tasks tailored to assess different perspectives, including Vanilla Name Answering, Entity-Level Caption, and Complex-Scenario Recognition. In addition, a new form of knowledge editing, Multi-step Editing, is introduced to evaluate the editing efficiency. Through our extensive evaluations, we demonstrate that the current state-of-the-art methods face significant challenges in tackling our proposed benchmark, underscoring the complexity of FG knowledge editing in MLLMs. Our findings spotlight the urgent need for novel approaches in this domain, setting a clear agenda for future research and development efforts within the community.",
}
| Multimodal knowledge editing represents a critical advancement in enhancing the capabilities of Multimodal Large Language Models (MLLMs). Despite its potential, current benchmarks predominantly focus on coarse-grained knowledge, leaving the intricacies of fine-grained (FG) multimodal entity knowledge largely unexplored. This gap presents a notable challenge, as FG entity recognition is pivotal for the practical deployment and effectiveness of MLLMs in diverse real-world scenarios. To bridge this gap, we introduce MIKE, a comprehensive benchmark and dataset specifically designed for the FG multimodal entity knowledge editing. MIKE encompasses a suite of tasks tailored to assess different perspectives, including Vanilla Name Answering, Entity-Level Caption, and Complex-Scenario Recognition. In addition, a new form of knowledge editing, Multi-step Editing, is introduced to evaluate the editing efficiency. Through our extensive evaluations, we demonstrate that the current state-of-the-art methods face significant challenges in tackling our proposed benchmark, underscoring the complexity of FG knowledge editing in MLLMs. Our findings spotlight the urgent need for novel approaches in this domain, setting a clear agenda for future research and development efforts within the community. | [
"Li, Jiaqi",
"Du, Miaozeng",
"Zhang, Chuanyi",
"Chen, Yongrui",
"Hu, Nan",
"Qi, Guilin",
"Jiang, Haiyun",
"Cheng, Siyuan",
"Tian, Bozhong"
]
| {MIKE}: A New Benchmark for Fine-grained Multimodal Entity Knowledge Editing | findings-acl.298 | Poster | 2402.14835v1 |
https://aclanthology.org/2024.findings-acl.299.bib | @inproceedings{wan-etal-2024-reformulating,
title = "Reformulating Domain Adaptation of Large Language Models as Adapt-Retrieve-Revise: A Case Study on {C}hinese Legal Domain",
author = "Wan, Zhen and
Zhang, Yating and
Wang, Yexiang and
Cheng, Fei and
Kurohashi, Sadao",
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.299",
pages = "5030--5041",
abstract = "While large language models (LLMs) like GPT-4 have recently demonstrated astonishing zero-shot capabilities in general domain tasks, they often generate content with hallucinations in specific domains such as Chinese law, hindering their application in these areas. This is typically due to the absence of training data that encompasses such a specific domain, preventing GPT-4 from acquiring in-domain knowledge. A pressing challenge is that it{'}s not plausible to continue training LLMs of the GPT-4{'}s scale on in-domain data.This paper introduces a simple yet effective domain adaptation framework for GPT-4 by reformulating generation as an adapt-retrieve-revise process. The initial step is to adapt an affordable 7B LLM to the Chinese legal domain by continuing learning in-domain data. When solving an in-domain task, we leverage the adapted LLM to generate a draft answer given a task query. Then, the draft answer will be used to retrieve supporting evidence candidates from an external in-domain knowledge base. Finally, the draft answer and retrieved evidence are concatenated into a whole prompt to let GPT-4 assess the evidence and revise the draft answer to generate the final answer. Our proposal combines the advantages of the efficiency of adapting a smaller 7B model with the evidence-assessing capability of GPT-4 and effectively prevents GPT-4 from generating hallucinatory content. In the zero-shot setting of four Chinese legal tasks, our method improves the average score by +33.6 points, compared to GPT-4 direct generation. When compared to two stronger retrieval-based baselines, our method outperforms them by +17.0 and +23.5.",
}
| While large language models (LLMs) like GPT-4 have recently demonstrated astonishing zero-shot capabilities in general domain tasks, they often generate content with hallucinations in specific domains such as Chinese law, hindering their application in these areas. This is typically due to the absence of training data that encompasses such a specific domain, preventing GPT-4 from acquiring in-domain knowledge. A pressing challenge is that it{'}s not plausible to continue training LLMs of the GPT-4{'}s scale on in-domain data.This paper introduces a simple yet effective domain adaptation framework for GPT-4 by reformulating generation as an adapt-retrieve-revise process. The initial step is to adapt an affordable 7B LLM to the Chinese legal domain by continuing learning in-domain data. When solving an in-domain task, we leverage the adapted LLM to generate a draft answer given a task query. Then, the draft answer will be used to retrieve supporting evidence candidates from an external in-domain knowledge base. Finally, the draft answer and retrieved evidence are concatenated into a whole prompt to let GPT-4 assess the evidence and revise the draft answer to generate the final answer. Our proposal combines the advantages of the efficiency of adapting a smaller 7B model with the evidence-assessing capability of GPT-4 and effectively prevents GPT-4 from generating hallucinatory content. In the zero-shot setting of four Chinese legal tasks, our method improves the average score by +33.6 points, compared to GPT-4 direct generation. When compared to two stronger retrieval-based baselines, our method outperforms them by +17.0 and +23.5. | [
"Wan, Zhen",
"Zhang, Yating",
"Wang, Yexiang",
"Cheng, Fei",
"Kurohashi, Sadao"
]
| Reformulating Domain Adaptation of Large Language Models as Adapt-Retrieve-Revise: A Case Study on {C}hinese Legal Domain | findings-acl.299 | Poster | 2311.00333v2 |
https://aclanthology.org/2024.findings-acl.300.bib | @inproceedings{agarwal-etal-2024-mememqa,
title = "{M}eme{MQA}: Multimodal Question Answering for Memes via Rationale-Based Inferencing",
author = "Agarwal, Siddhant and
Sharma, Shivam and
Nakov, Preslav and
Chakraborty, Tanmoy",
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.300",
pages = "5042--5078",
abstract = "Memes have evolved as a prevalent medium for diverse communication, ranging from humour to propaganda. With the rising popularity of image-focused content, there is a growing need to explore its potential harm from different aspects. Previous studies have analyzed memes in closed settings - detecting harm, applying semantic labels, and offering natural language explanations. To extend this research, we introduce MemeMQA, a multimodal question-answering framework aiming to solicit accurate responses to structured questions while providing coherent explanations. We curate MemeMQACorpus, a new dataset featuring 1,880 questions related to 1,122 memes with corresponding answer-explanation pairs. We further propose ARSENAL, a novel two-stage multimodal framework that leverages the reasoning capabilities of LLMs to address MemeMQA. We benchmark MemeMQA using competitive baselines and demonstrate its superiority - {\textasciitilde}18{\%} enhanced answer prediction accuracy and distinct text generation lead across various metrics measuring lexical and semantic alignment over the best baseline. We analyze ARSENAL{'}s robustness through diversification of question-set, confounder-based evaluation regarding MemeMQA{'}s generalizability, and modality-specific assessment, enhancing our understanding of meme interpretation in the multimodal communication landscape.",
}
| Memes have evolved as a prevalent medium for diverse communication, ranging from humour to propaganda. With the rising popularity of image-focused content, there is a growing need to explore its potential harm from different aspects. Previous studies have analyzed memes in closed settings - detecting harm, applying semantic labels, and offering natural language explanations. To extend this research, we introduce MemeMQA, a multimodal question-answering framework aiming to solicit accurate responses to structured questions while providing coherent explanations. We curate MemeMQACorpus, a new dataset featuring 1,880 questions related to 1,122 memes with corresponding answer-explanation pairs. We further propose ARSENAL, a novel two-stage multimodal framework that leverages the reasoning capabilities of LLMs to address MemeMQA. We benchmark MemeMQA using competitive baselines and demonstrate its superiority - {\textasciitilde}18{\%} enhanced answer prediction accuracy and distinct text generation lead across various metrics measuring lexical and semantic alignment over the best baseline. We analyze ARSENAL{'}s robustness through diversification of question-set, confounder-based evaluation regarding MemeMQA{'}s generalizability, and modality-specific assessment, enhancing our understanding of meme interpretation in the multimodal communication landscape. | [
"Agarwal, Siddhant",
"Sharma, Shivam",
"Nakov, Preslav",
"Chakraborty, Tanmoy"
]
| {M}eme{MQA}: Multimodal Question Answering for Memes via Rationale-Based Inferencing | findings-acl.300 | Poster | 2405.11215v1 |
https://aclanthology.org/2024.findings-acl.301.bib | @inproceedings{li-etal-2024-improving-attributed,
title = "Improving Attributed Text Generation of Large Language Models via Preference Learning",
author = "Li, Dongfang and
Sun, Zetian and
Hu, Baotian and
Liu, Zhenyu and
Hu, Xinshuo and
Liu, Xuebo 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.301",
pages = "5079--5101",
abstract = "Large language models have been widely adopted in natural language processing, yet they face the challenge of generating unreliable content. Recent works aim to reduce misinformation and hallucinations by resorting to attribution as a means to provide evidence (i.e., citations). However, current attribution methods usually focus on the retrieval stage and automatic evaluation that neglect mirroring the citation mechanisms in human scholarly writing to bolster credibility. In this paper, we address these challenges by modelling the attribution task as preference learning and introducing an Automatic Preference Optimization (APO) framework. First, we create a curated collection for post-training with 6,330 examples by collecting and filtering from existing datasets. Second, considering the high cost of labelling preference data, we further propose an automatic method to synthesize attribution preference data resulting in 95,263 pairs. Moreover, inspired by the human citation process, we further propose a progressive preference optimization method by leveraging fine-grained information. Extensive experiments on three datasets (i.e., ASQA, StrategyQA, and ELI5) demonstrate that APO achieves state-of-the-art citation F1 with higher answer quality.",
}
| Large language models have been widely adopted in natural language processing, yet they face the challenge of generating unreliable content. Recent works aim to reduce misinformation and hallucinations by resorting to attribution as a means to provide evidence (i.e., citations). However, current attribution methods usually focus on the retrieval stage and automatic evaluation that neglect mirroring the citation mechanisms in human scholarly writing to bolster credibility. In this paper, we address these challenges by modelling the attribution task as preference learning and introducing an Automatic Preference Optimization (APO) framework. First, we create a curated collection for post-training with 6,330 examples by collecting and filtering from existing datasets. Second, considering the high cost of labelling preference data, we further propose an automatic method to synthesize attribution preference data resulting in 95,263 pairs. Moreover, inspired by the human citation process, we further propose a progressive preference optimization method by leveraging fine-grained information. Extensive experiments on three datasets (i.e., ASQA, StrategyQA, and ELI5) demonstrate that APO achieves state-of-the-art citation F1 with higher answer quality. | [
"Li, Dongfang",
"Sun, Zetian",
"Hu, Baotian",
"Liu, Zhenyu",
"Hu, Xinshuo",
"Liu, Xuebo",
"Zhang, Min"
]
| Improving Attributed Text Generation of Large Language Models via Preference Learning | findings-acl.301 | Poster | 2403.18381v1 |
https://aclanthology.org/2024.findings-acl.302.bib | @inproceedings{kim-etal-2024-kombo,
title = "{KOMBO}: {K}orean Character Representations Based on the Combination Rules of Subcharacters",
author = "Kim, SungHo and
Park, Juhyeong and
Kim, Yeachan and
Lee, SangKeun",
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.302",
pages = "5102--5119",
abstract = "The Korean writing system, Hangeul, has a unique character representation rigidly following the invention principles recorded in Hunminjeongeum. However, existing pre-trained language models (PLMs) for Korean have overlooked these principles. In this paper, we introduce a novel framework for Korean PLMs called KOMBO, which firstly brings the invention principles of Hangeul to represent character. Our proposed method, KOMBO, exhibits notable experimental proficiency across diverse NLP tasks. In particular, our method outperforms the state-of-the-art Korean PLM by an average of 2.11{\%} in five Korean natural language understanding tasks. Furthermore, extensive experiments demonstrate that our proposed method is suitable for comprehending the linguistic features of the Korean language. Consequently, we shed light on the superiority of using subcharacters over the typical subword-based approach for Korean PLMs. Our code is available at: https://github.com/SungHo3268/KOMBO.",
}
| The Korean writing system, Hangeul, has a unique character representation rigidly following the invention principles recorded in Hunminjeongeum. However, existing pre-trained language models (PLMs) for Korean have overlooked these principles. In this paper, we introduce a novel framework for Korean PLMs called KOMBO, which firstly brings the invention principles of Hangeul to represent character. Our proposed method, KOMBO, exhibits notable experimental proficiency across diverse NLP tasks. In particular, our method outperforms the state-of-the-art Korean PLM by an average of 2.11{\%} in five Korean natural language understanding tasks. Furthermore, extensive experiments demonstrate that our proposed method is suitable for comprehending the linguistic features of the Korean language. Consequently, we shed light on the superiority of using subcharacters over the typical subword-based approach for Korean PLMs. Our code is available at: https://github.com/SungHo3268/KOMBO. | [
"Kim, SungHo",
"Park, Juhyeong",
"Kim, Yeachan",
"Lee, SangKeun"
]
| {KOMBO}: {K}orean Character Representations Based on the Combination Rules of Subcharacters | findings-acl.302 | Poster | 1405.5474v1 |
https://aclanthology.org/2024.findings-acl.303.bib | @inproceedings{yoshida-etal-2024-tree,
title = "Tree-Planted Transformers: Unidirectional Transformer Language Models with Implicit Syntactic Supervision",
author = "Yoshida, Ryo and
Someya, Taiga and
Oseki, Yohei",
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.303",
pages = "5120--5134",
abstract = "Syntactic Language Models (SLMs) can be trained efficiently to reach relatively high performance; however, they have trouble with inference efficiency due to the explicit generation of syntactic structures. In this paper, we propose a new method dubbed tree-planting: instead of explicitly generating syntactic structures, we {``}plant{''} trees into attention weights of unidirectional Transformer LMs to implicitly reflect syntactic structures of natural language. Specifically, unidirectional Transformer LMs trained with tree-planting will be called Tree-Planted Transformers (TPT), which inherit the training efficiency from SLMs without changing the inference efficiency of their underlying Transformer LMs. Targeted syntactic evaluations on the SyntaxGym benchmark demonstrated that TPTs, despite the lack of explicit generation of syntactic structures, significantly outperformed not only vanilla Transformer LMs but also various SLMs that generate hundreds of syntactic structures in parallel. This result suggests that TPTs can learn human-like syntactic knowledge as data-efficiently as SLMs while maintaining the modeling space of Transformer LMs unchanged.",
}
| Syntactic Language Models (SLMs) can be trained efficiently to reach relatively high performance; however, they have trouble with inference efficiency due to the explicit generation of syntactic structures. In this paper, we propose a new method dubbed tree-planting: instead of explicitly generating syntactic structures, we {``}plant{''} trees into attention weights of unidirectional Transformer LMs to implicitly reflect syntactic structures of natural language. Specifically, unidirectional Transformer LMs trained with tree-planting will be called Tree-Planted Transformers (TPT), which inherit the training efficiency from SLMs without changing the inference efficiency of their underlying Transformer LMs. Targeted syntactic evaluations on the SyntaxGym benchmark demonstrated that TPTs, despite the lack of explicit generation of syntactic structures, significantly outperformed not only vanilla Transformer LMs but also various SLMs that generate hundreds of syntactic structures in parallel. This result suggests that TPTs can learn human-like syntactic knowledge as data-efficiently as SLMs while maintaining the modeling space of Transformer LMs unchanged. | [
"Yoshida, Ryo",
"Someya, Taiga",
"Oseki, Yohei"
]
| Tree-Planted Transformers: Unidirectional Transformer Language Models with Implicit Syntactic Supervision | findings-acl.303 | Poster | 2402.12691v2 |
https://aclanthology.org/2024.findings-acl.304.bib | @inproceedings{chang-etal-2024-play,
title = "Play Guessing Game with {LLM}: Indirect Jailbreak Attack with Implicit Clues",
author = "Chang, Zhiyuan and
Li, Mingyang and
Liu, Yi and
Wang, Junjie and
Wang, Qing and
Liu, Yang",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.304",
pages = "5135--5147",
abstract = "With the development of LLMs, the security threats of LLMs are getting more and more attention. Numerous jailbreak attacks have been proposed to assess the security defense of LLMs. Current jailbreak attacks primarily utilize scenario camouflage techniques. However their explicitly mention of malicious intent will be easily recognized and defended by LLMs. In this paper, we propose an indirect jailbreak attack approach, Puzzler, which can bypass the LLM{'}s defensive strategies and obtain malicious response by implicitly providing LLMs with some clues about the original malicious query. In addition, inspired by the wisdom of {``}When unable to attack, defend{''} from Sun Tzu{'}s Art of War, we adopt a defensive stance to gather clues about the original malicious query through LLMs. The experimental results indicate that the Query Success Rate of the Puzzler is 14.0{\%}-82.7{\%} higher than baselines on the most prominent LLMs. Furthermore, when tested against the state-of-the-art jailbreak detection approaches, Puzzler proves to be more effective at evading detection compared to baselines.",
}
| With the development of LLMs, the security threats of LLMs are getting more and more attention. Numerous jailbreak attacks have been proposed to assess the security defense of LLMs. Current jailbreak attacks primarily utilize scenario camouflage techniques. However their explicitly mention of malicious intent will be easily recognized and defended by LLMs. In this paper, we propose an indirect jailbreak attack approach, Puzzler, which can bypass the LLM{'}s defensive strategies and obtain malicious response by implicitly providing LLMs with some clues about the original malicious query. In addition, inspired by the wisdom of {``}When unable to attack, defend{''} from Sun Tzu{'}s Art of War, we adopt a defensive stance to gather clues about the original malicious query through LLMs. The experimental results indicate that the Query Success Rate of the Puzzler is 14.0{\%}-82.7{\%} higher than baselines on the most prominent LLMs. Furthermore, when tested against the state-of-the-art jailbreak detection approaches, Puzzler proves to be more effective at evading detection compared to baselines. | [
"Chang, Zhiyuan",
"Li, Mingyang",
"Liu, Yi",
"Wang, Junjie",
"Wang, Qing",
"Liu, Yang"
]
| Play Guessing Game with {LLM}: Indirect Jailbreak Attack with Implicit Clues | findings-acl.304 | Poster | 2402.09091v2 |
https://aclanthology.org/2024.findings-acl.305.bib | @inproceedings{kweon-etal-2024-publicly,
title = "Publicly Shareable Clinical Large Language Model Built on Synthetic Clinical Notes",
author = "Kweon, Sunjun and
Kim, Junu and
Kim, Jiyoun and
Im, Sujeong and
Cho, Eunbyeol and
Bae, Seongsu and
Oh, Jungwoo and
Lee, Gyubok and
Moon, Jong Hak and
You, Seng Chan and
Baek, Seungjin and
Han, Chang Hoon and
Jung, Yoon Bin and
Jo, Yohan and
Choi, Edward",
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.305",
pages = "5148--5168",
abstract = "The development of large language models tailored for handling patients{'} clinical notes is often hindered by the limited accessibility and usability of these notes due to strict privacy regulations.To address these challenges, we first create synthetic large-scale clinical notes using publicly available case reports extracted from biomedical literature.We then use these synthetic notes to train our specialized clinical large language model, Asclepius.While Asclepius is trained on synthetic data, we assess its potential performance in real-world applications by evaluating it using real clinical notes.We benchmark Asclepius against several other large language models, including GPT-3.5-turbo and other open-source alternatives. To further validate our approach using synthetic notes, we also compare Asclepius with its variants trained on real clinical notes. Our findings convincingly demonstrate that synthetic clinical notes can serve as viable substitutes for real ones when constructing high-performing clinical language models. This conclusion is supported by detailed evaluations conducted by both GPT-4 and medical professionals. All resources{---}including weights, codes, and data{---}used in the development of Asclepius will be made publicly accessible for future research.",
}
| The development of large language models tailored for handling patients{'} clinical notes is often hindered by the limited accessibility and usability of these notes due to strict privacy regulations.To address these challenges, we first create synthetic large-scale clinical notes using publicly available case reports extracted from biomedical literature.We then use these synthetic notes to train our specialized clinical large language model, Asclepius.While Asclepius is trained on synthetic data, we assess its potential performance in real-world applications by evaluating it using real clinical notes.We benchmark Asclepius against several other large language models, including GPT-3.5-turbo and other open-source alternatives. To further validate our approach using synthetic notes, we also compare Asclepius with its variants trained on real clinical notes. Our findings convincingly demonstrate that synthetic clinical notes can serve as viable substitutes for real ones when constructing high-performing clinical language models. This conclusion is supported by detailed evaluations conducted by both GPT-4 and medical professionals. All resources{---}including weights, codes, and data{---}used in the development of Asclepius will be made publicly accessible for future research. | [
"Kweon, Sunjun",
"Kim, Junu",
"Kim, Jiyoun",
"Im, Sujeong",
"Cho, Eunbyeol",
"Bae, Seongsu",
"Oh, Jungwoo",
"Lee, Gyubok",
"Moon, Jong Hak",
"You, Seng Chan",
"Baek, Seungjin",
"Han, Chang Hoon",
"Jung, Yoon Bin",
"Jo, Yohan",
"Choi, Edward"
]
| Publicly Shareable Clinical Large Language Model Built on Synthetic Clinical Notes | findings-acl.305 | Poster | 2309.00237v4 |
https://aclanthology.org/2024.findings-acl.306.bib | @inproceedings{fei-etal-2024-extending,
title = "Extending Context Window of Large Language Models via Semantic Compression",
author = "Fei, Weizhi and
Niu, Xueyan and
Zhou, Pingyi and
Hou, Lu and
Bai, Bo and
Deng, Lei and
Han, Wei",
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.306",
pages = "5169--5181",
abstract = "Transformer based Large Language Models (LLMs) often impose limitations on the length of the text input to ensure the generation of fluent and relevant responses due to the quadratic complexity. These constraints restrict their applicability in long text scenarios. In this paper, we propose a novel semantic compression method that enables generalization to texts that are 6-8 times longer without incurring significant computational costs or requiring fine-tuning. Our proposed framework draws inspiration from source coding in information theory and employs a pre-trained model to reduce the semantic redundancy of long inputs before passing them to the LLMs for downstream tasks. Experimental results demonstrate that our method effectively extends the context window of LLMs across a range of tasks including question answering, summarization, few-shot learning, and information retrieval. Furthermore, the proposed semantic compression method exhibits consistent fluency in text generation while reducing the associated computational overhead.",
}
| Transformer based Large Language Models (LLMs) often impose limitations on the length of the text input to ensure the generation of fluent and relevant responses due to the quadratic complexity. These constraints restrict their applicability in long text scenarios. In this paper, we propose a novel semantic compression method that enables generalization to texts that are 6-8 times longer without incurring significant computational costs or requiring fine-tuning. Our proposed framework draws inspiration from source coding in information theory and employs a pre-trained model to reduce the semantic redundancy of long inputs before passing them to the LLMs for downstream tasks. Experimental results demonstrate that our method effectively extends the context window of LLMs across a range of tasks including question answering, summarization, few-shot learning, and information retrieval. Furthermore, the proposed semantic compression method exhibits consistent fluency in text generation while reducing the associated computational overhead. | [
"Fei, Weizhi",
"Niu, Xueyan",
"Zhou, Pingyi",
"Hou, Lu",
"Bai, Bo",
"Deng, Lei",
"Han, Wei"
]
| Extending Context Window of Large Language Models via Semantic Compression | findings-acl.306 | Poster | 2312.09571v1 |
https://aclanthology.org/2024.findings-acl.307.bib | @inproceedings{wei-jie-etal-2024-plausible,
title = "Plausible Extractive Rationalization through Semi-Supervised Entailment Signal",
author = "Wei Jie, Yeo and
Satapathy, Ranjan and
Cambria, Erik",
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.307",
pages = "5182--5192",
abstract = "The increasing use of complex and opaque black box models requires the adoption of interpretable measures, one such option is extractive rationalizing models, which serve as a more interpretable alternative. These models, also known as Explain-Then-Predict models, employ an explainer model to extract rationales and subsequently condition the predictor with the extracted information. Their primary objective is to provide precise and faithful explanations, represented by the extracted rationales. In this paper, we take a semi-supervised approach to optimize for the plausibility of extracted rationales. We adopt a pre-trained natural language inference (NLI) model and further fine-tune it on a small set of supervised rationales (10{\%}). The NLI predictor is leveraged as a source of supervisory signals to the explainer via entailment alignment. We show that, by enforcing the alignment agreement between the explanation and answer in a question-answering task, the performance can be improved without access to ground truth labels. We evaluate our approach on the ERASER dataset and show that our approach achieves comparable results with supervised extractive models and outperforms unsupervised approaches by {\textgreater} 100{\%}.",
}
| The increasing use of complex and opaque black box models requires the adoption of interpretable measures, one such option is extractive rationalizing models, which serve as a more interpretable alternative. These models, also known as Explain-Then-Predict models, employ an explainer model to extract rationales and subsequently condition the predictor with the extracted information. Their primary objective is to provide precise and faithful explanations, represented by the extracted rationales. In this paper, we take a semi-supervised approach to optimize for the plausibility of extracted rationales. We adopt a pre-trained natural language inference (NLI) model and further fine-tune it on a small set of supervised rationales (10{\%}). The NLI predictor is leveraged as a source of supervisory signals to the explainer via entailment alignment. We show that, by enforcing the alignment agreement between the explanation and answer in a question-answering task, the performance can be improved without access to ground truth labels. We evaluate our approach on the ERASER dataset and show that our approach achieves comparable results with supervised extractive models and outperforms unsupervised approaches by {\textgreater} 100{\%}. | [
"Wei Jie, Yeo",
"Satapathy, Ranjan",
"Cambria, Erik"
]
| Plausible Extractive Rationalization through Semi-Supervised Entailment Signal | findings-acl.307 | Poster | 2402.08479v4 |
https://aclanthology.org/2024.findings-acl.308.bib | @inproceedings{park-etal-2024-translation,
title = "Translation Deserves Better: Analyzing Translation Artifacts in Cross-lingual Visual Question Answering",
author = "Park, ChaeHun and
Lee, Koanho and
Lim, Hyesu and
Kim, Jaeseok and
Park, Junmo and
Heo, Yu-Jung and
Chang, Du-Seong and
Choo, Jaegul",
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.308",
pages = "5193--5221",
abstract = "Building a reliable visual question answering (VQA) system across different languages is a challenging problem, primarily due to the lack of abundant samples for training. To address this challenge, recent studies have employed machine translation systems for the cross-lingual VQA task. This involves translating the evaluation samples into a source language (usually English) and using monolingual models (i.e., translate-test). However, our analysis reveals that translated texts contain unique characteristics distinct from human-written ones, referred to as translation artifacts. We find that these artifacts can significantly affect the models, confirmed by extensive experiments across diverse models, languages, and translation processes. In light of this, we present a simple data augmentation strategy that can alleviate the adverse impacts of translation artifacts.",
}
| Building a reliable visual question answering (VQA) system across different languages is a challenging problem, primarily due to the lack of abundant samples for training. To address this challenge, recent studies have employed machine translation systems for the cross-lingual VQA task. This involves translating the evaluation samples into a source language (usually English) and using monolingual models (i.e., translate-test). However, our analysis reveals that translated texts contain unique characteristics distinct from human-written ones, referred to as translation artifacts. We find that these artifacts can significantly affect the models, confirmed by extensive experiments across diverse models, languages, and translation processes. In light of this, we present a simple data augmentation strategy that can alleviate the adverse impacts of translation artifacts. | [
"Park, ChaeHun",
"Lee, Koanho",
"Lim, Hyesu",
"Kim, Jaeseok",
"Park, Junmo",
"Heo, Yu-Jung",
"Chang, Du-Seong",
"Choo, Jaegul"
]
| Translation Deserves Better: Analyzing Translation Artifacts in Cross-lingual Visual Question Answering | findings-acl.308 | Poster | 2406.02331v1 |
https://aclanthology.org/2024.findings-acl.309.bib | @inproceedings{yang-etal-2024-scented,
title = "Scented-{EAE}: Stage-Customized Entity Type Embedding for Event Argument Extraction",
author = "Yang, Yu and
Guo, Jinyu and
Shuang, Kai and
Mao, Chenrui",
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.309",
pages = "5222--5235",
abstract = "Existing methods for incorporating entities into EAE rely on prompts or NER. They typically fail to explicitly explore the role of entity types, which results in shallow argument comprehension and often encounter three issues: (1) weak semantic associations due to missing role-entity correspondence cues; (2) compromised semantic integrity from abandoning context after recognizing entities regardless of their types; (3) one-sided semantic understanding relying solely on argument role semantics. To tackle these issues, we propose Scented-EAE, an EAE model with stage-customized entity type embedding to explicitly underscore and explore the role of entity types, thus intervening in argument selection. Specifically, at the input stage, we strengthen semantic associations by prompting role-entity correspondence after extending a non-autoregressive decoder as part of the encoder. At the intermediate stage, we preserve semantic integrity by optimizing our proposed BIO-aware NER and EAE via a novel IPE joint learning. At the output stage, we expand semantic understanding dimensions by determining arguments using span selectors from argument roles and entity types. Experiments show that our model achieves state-of-the-art performance on mainstream benchmarks. In addition, it also exhibits robustness in low-resource settings with the help of prompts and entity types.",
}
| Existing methods for incorporating entities into EAE rely on prompts or NER. They typically fail to explicitly explore the role of entity types, which results in shallow argument comprehension and often encounter three issues: (1) weak semantic associations due to missing role-entity correspondence cues; (2) compromised semantic integrity from abandoning context after recognizing entities regardless of their types; (3) one-sided semantic understanding relying solely on argument role semantics. To tackle these issues, we propose Scented-EAE, an EAE model with stage-customized entity type embedding to explicitly underscore and explore the role of entity types, thus intervening in argument selection. Specifically, at the input stage, we strengthen semantic associations by prompting role-entity correspondence after extending a non-autoregressive decoder as part of the encoder. At the intermediate stage, we preserve semantic integrity by optimizing our proposed BIO-aware NER and EAE via a novel IPE joint learning. At the output stage, we expand semantic understanding dimensions by determining arguments using span selectors from argument roles and entity types. Experiments show that our model achieves state-of-the-art performance on mainstream benchmarks. In addition, it also exhibits robustness in low-resource settings with the help of prompts and entity types. | [
"Yang, Yu",
"Guo, Jinyu",
"Shuang, Kai",
"Mao, Chenrui"
]
| Scented-{EAE}: Stage-Customized Entity Type Embedding for Event Argument Extraction | findings-acl.309 | Poster | 2403.12523v1 |
https://aclanthology.org/2024.findings-acl.310.bib | @inproceedings{lei-etal-2024-fast,
title = "Fast Randomized Low-Rank Adaptation of Pre-trained Language Models with {PAC} Regularization",
author = "Lei, Zijian and
Qian, Dong and
Cheung, William",
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.310",
pages = "5236--5249",
abstract = "Low-rank adaptation (LoRA) achieves parameter efficient fine-tuning for large language models (LLMs) by decomposing the model weight update into a pair of low-rank projection matrices. Yet, the memory overhead restricts it to scale up when the model size increases. We propose Randomized LoRA (RLoRA) which adopts Randomized Walsh-Hadamard Transform to achieve significant reduction in the size of trainable parameters compared to LoRA. At the same time, it allows a PAC-Bayes regularizer to be efficiently incorporated to improve generalization. We evaluate the effectiveness of RLoRA on LLMs RoBERTa, GPT-2 and LLaMA-7B using GLUE, E2E and math reasoning benchmarks. With a much lower memory requirement, RLoRA can give similar performance as the SOTA low-rank adaptation methods for these three tasks and significantly better performance under few-shot settings.",
}
| Low-rank adaptation (LoRA) achieves parameter efficient fine-tuning for large language models (LLMs) by decomposing the model weight update into a pair of low-rank projection matrices. Yet, the memory overhead restricts it to scale up when the model size increases. We propose Randomized LoRA (RLoRA) which adopts Randomized Walsh-Hadamard Transform to achieve significant reduction in the size of trainable parameters compared to LoRA. At the same time, it allows a PAC-Bayes regularizer to be efficiently incorporated to improve generalization. We evaluate the effectiveness of RLoRA on LLMs RoBERTa, GPT-2 and LLaMA-7B using GLUE, E2E and math reasoning benchmarks. With a much lower memory requirement, RLoRA can give similar performance as the SOTA low-rank adaptation methods for these three tasks and significantly better performance under few-shot settings. | [
"Lei, Zijian",
"Qian, Dong",
"Cheung, William"
]
| Fast Randomized Low-Rank Adaptation of Pre-trained Language Models with {PAC} Regularization | findings-acl.310 | Poster | 2310.17588v1 |
https://aclanthology.org/2024.findings-acl.311.bib | @inproceedings{yang-etal-2024-sda,
title = "{SDA}: Semantic Discrepancy Alignment for Text-conditioned Image Retrieval",
author = "Yang, Yuchen and
Wang, Yu and
Wang, Yanfeng",
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.311",
pages = "5250--5261",
abstract = "In the realm of text-conditioned image retrieval, models utilize a query composed of a reference image and modification text to retrieve corresponding images. Despite its significance, this task is fraught with challenges, including small-scale datasets due to labeling costs and the complexity of attributes in modification texts. These challenges often result in models learning a generalized representation of the query, thereby missing the semantic correlations of image and text attributes.In this paper, we introduce a general boosting framework designed to address these issues by employing semantic discrepancy alignment. Our framework first leverages the ChatGPT to augment text data by modifying the original modification text{'}s attributes. The augmented text is then combined with the original reference image to create an augmented composed query. Then we generate corresponding images using GPT-4 for the augmented composed query.We realize the cross-modal semantic discrepancy alignment by formulating distance consistency and neighbor consistency between the image and text domains. Through this novel approach, attribute in the text domain can be more effectively transferred to the image domain, enhancing retrieval performance. Extensive experiments on three prominent datasets validate the effectiveness of our approach, with state-of-the-art results on a majority of evaluation metrics compared to various baseline methods.",
}
| In the realm of text-conditioned image retrieval, models utilize a query composed of a reference image and modification text to retrieve corresponding images. Despite its significance, this task is fraught with challenges, including small-scale datasets due to labeling costs and the complexity of attributes in modification texts. These challenges often result in models learning a generalized representation of the query, thereby missing the semantic correlations of image and text attributes.In this paper, we introduce a general boosting framework designed to address these issues by employing semantic discrepancy alignment. Our framework first leverages the ChatGPT to augment text data by modifying the original modification text{'}s attributes. The augmented text is then combined with the original reference image to create an augmented composed query. Then we generate corresponding images using GPT-4 for the augmented composed query.We realize the cross-modal semantic discrepancy alignment by formulating distance consistency and neighbor consistency between the image and text domains. Through this novel approach, attribute in the text domain can be more effectively transferred to the image domain, enhancing retrieval performance. Extensive experiments on three prominent datasets validate the effectiveness of our approach, with state-of-the-art results on a majority of evaluation metrics compared to various baseline methods. | [
"Yang, Yuchen",
"Wang, Yu",
"Wang, Yanfeng"
]
| {SDA}: Semantic Discrepancy Alignment for Text-conditioned Image Retrieval | findings-acl.311 | Poster | 2302.06081v2 |
https://aclanthology.org/2024.findings-acl.312.bib | @inproceedings{liu-etal-2024-se2,
title = "$Se^2$: Sequential Example Selection for In-Context Learning",
author = "Liu, Haoyu and
Liu, Jianfeng and
Huang, Shaohan and
Zhan, Yuefeng and
Sun, Hao and
Deng, Weiwei and
Wei, Furu and
Zhang, Qi",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.312",
pages = "5262--5284",
abstract = "The remarkable capability of large language models(LLMs) for in-context learning(ICL) needs to be activated by demonstration examples. Prior work has extensively explored the selection of examples for ICL, predominantly following the {``}select then organize{''} paradigm, such approaches often neglect the internal relationships between examples and exist an inconsistency between the training and inference. In this paper, we formulate the problem as a $Se$quential $Se$lection problem and introduce $Se^2$, a sequential-aware method that leverages the LLM{'}s feedback on varying context, aiding in capturing inter-relationships and sequential information among examples, significantly enriching the contextuality and relevance of ICL prompts. Meanwhile, we utilize beam search to seek and construct example sequences, enhancing both quality and diversity. Extensive experiments across 23 NLP tasks from 8 distinct categories illustrate that $Se^2$ markedly surpasses competitive baselines and achieves 42{\%} relative improvement over random selection. Further in-depth analysis shows the effectiveness of proposed strategies, highlighting $Se^2${`}s exceptional stability and adaptability across various scenarios. Code available at https://github.com/microsoft/LMOps.",
}
| The remarkable capability of large language models(LLMs) for in-context learning(ICL) needs to be activated by demonstration examples. Prior work has extensively explored the selection of examples for ICL, predominantly following the {``}select then organize{''} paradigm, such approaches often neglect the internal relationships between examples and exist an inconsistency between the training and inference. In this paper, we formulate the problem as a $Se$quential $Se$lection problem and introduce $Se^2$, a sequential-aware method that leverages the LLM{'}s feedback on varying context, aiding in capturing inter-relationships and sequential information among examples, significantly enriching the contextuality and relevance of ICL prompts. Meanwhile, we utilize beam search to seek and construct example sequences, enhancing both quality and diversity. Extensive experiments across 23 NLP tasks from 8 distinct categories illustrate that $Se^2$ markedly surpasses competitive baselines and achieves 42{\%} relative improvement over random selection. Further in-depth analysis shows the effectiveness of proposed strategies, highlighting $Se^2${`}s exceptional stability and adaptability across various scenarios. Code available at https://github.com/microsoft/LMOps. | [
"Liu, Haoyu",
"Liu, Jianfeng",
"Huang, Shaohan",
"Zhan, Yuefeng",
"Sun, Hao",
"Deng, Weiwei",
"Wei, Furu",
"Zhang, Qi"
]
| $Se^2$: Sequential Example Selection for In-Context Learning | findings-acl.312 | Poster | 2402.13874 |
https://aclanthology.org/2024.findings-acl.313.bib | @inproceedings{yi-etal-2024-generation,
title = "Generation Meets Verification: Accelerating Large Language Model Inference with Smart Parallel Auto-Correct Decoding",
author = "Yi, Hanling and
Lin, Feng and
Li, Hongbin and
Peiyang, Ning and
Yu, Xiaotian and
Xiao, 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.313",
pages = "5285--5299",
abstract = "This research aims to accelerate the inference speed of large language models (LLMs) with billions of parameters. We propose Smart Parallel Auto-Correct dEcoding (SPACE), an approach designed for achieving lossless acceleration of LLMs. By integrating semi-autoregressive inference and speculative decoding capabilities, SPACE uniquely enables autoregressive LLMs to parallelize token generation and verification. This is realized through a specialized semi-autoregressive supervised fine-tuning process that equips existing LLMs with the ability to simultaneously predict multiple tokens. Additionally, an auto-correct decoding algorithm facilitates the simultaneous generation and verification of token sequences within a single model invocation. Through extensive experiments on a range of LLMs, SPACE has demonstrated inference speedup ranging from 2.7x-4.0x on HumanEval-X while maintaining output quality.",
}
| This research aims to accelerate the inference speed of large language models (LLMs) with billions of parameters. We propose Smart Parallel Auto-Correct dEcoding (SPACE), an approach designed for achieving lossless acceleration of LLMs. By integrating semi-autoregressive inference and speculative decoding capabilities, SPACE uniquely enables autoregressive LLMs to parallelize token generation and verification. This is realized through a specialized semi-autoregressive supervised fine-tuning process that equips existing LLMs with the ability to simultaneously predict multiple tokens. Additionally, an auto-correct decoding algorithm facilitates the simultaneous generation and verification of token sequences within a single model invocation. Through extensive experiments on a range of LLMs, SPACE has demonstrated inference speedup ranging from 2.7x-4.0x on HumanEval-X while maintaining output quality. | [
"Yi, Hanling",
"Lin, Feng",
"Li, Hongbin",
"Peiyang, Ning",
"Yu, Xiaotian",
"Xiao, Rong"
]
| Generation Meets Verification: Accelerating Large Language Model Inference with Smart Parallel Auto-Correct Decoding | findings-acl.313 | Poster | 2402.11809v3 |
https://aclanthology.org/2024.findings-acl.314.bib | @inproceedings{cao-etal-2024-structeval,
title = "{S}truct{E}val: Deepen and Broaden Large Language Model Assessment via Structured Evaluation",
author = "Cao, Boxi and
Ren, Mengjie and
Lin, Hongyu and
Han, Xianpei and
Zhang, Feng and
Zhan, Junfeng 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.314",
pages = "5300--5318",
abstract = "Evaluation is the baton for the development of large language models. Current evaluations typically employ a single-item assessment paradigm for each atomic test objective, which struggle to discern whether a model genuinely possesses the required capabilities or merely memorizes/guesses the answers to specific questions. To this end, this paper proposes a novel evaluation framework referred to as StructEval. Starting from an atomic test objective, StructEval deepens and broadens the evaluation by conducting a structured assessment across multiple cognitive levels and critical concepts, and therefore offers a comprehensive, robust and consistent evaluations for large language models. Experiments on three widely-used benchmarks demonstrate that StructEval serves as a reliable tool for resisting the risk of data contamination, and reducing the interference of potential biases, thereby providing a more reliable and consistent conclusion regarding model capabilities. Our framework also sheds light on the design of future principled and trustworthy LLM evaluation protocols.",
}
| Evaluation is the baton for the development of large language models. Current evaluations typically employ a single-item assessment paradigm for each atomic test objective, which struggle to discern whether a model genuinely possesses the required capabilities or merely memorizes/guesses the answers to specific questions. To this end, this paper proposes a novel evaluation framework referred to as StructEval. Starting from an atomic test objective, StructEval deepens and broadens the evaluation by conducting a structured assessment across multiple cognitive levels and critical concepts, and therefore offers a comprehensive, robust and consistent evaluations for large language models. Experiments on three widely-used benchmarks demonstrate that StructEval serves as a reliable tool for resisting the risk of data contamination, and reducing the interference of potential biases, thereby providing a more reliable and consistent conclusion regarding model capabilities. Our framework also sheds light on the design of future principled and trustworthy LLM evaluation protocols. | [
"Cao, Boxi",
"Ren, Mengjie",
"Lin, Hongyu",
"Han, Xianpei",
"Zhang, Feng",
"Zhan, Junfeng",
"Sun, Le"
]
| {S}truct{E}val: Deepen and Broaden Large Language Model Assessment via Structured Evaluation | findings-acl.314 | Poster | 2408.03281v2 |
https://aclanthology.org/2024.findings-acl.315.bib | @inproceedings{wu-etal-2024-mitigating-privacy,
title = "Mitigating Privacy Seesaw in Large Language Models: Augmented Privacy Neuron Editing via Activation Patching",
author = "Wu, Xinwei and
Dong, Weilong and
Xu, Shaoyang and
Xiong, Deyi",
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.315",
pages = "5319--5332",
abstract = "Protecting privacy leakage in large language models remains a paramount challenge. In this paper, we reveal Privacy Seesaw in LLM privacy safeguarding, a phenomenon where measures to secure specific private information inadvertently heighten exposure risks for other privacy. Through comprehensive analysis, we identify the amount of targeted privacy data and the volume of edited privacy neurons as the two central triggers to this issue. To mitigate privacy seesaw, we propose Augmented Privacy Neuron Editing via Activation Patching (APNEAP), a novel framework designed to well balance model performance with privacy protection. The proposed APNEAP augments collected private data by automatically synthesizing new private data, which deactivates the first trigger to the privacy seesaw issue. Additionally, it adapts activation patching to privacy neuron editing for switching off the second trigger to the privacy seesaw problem. Experimental results show that the proposed APNEAP is capable of alleviating the privacy seesaw phenomenon and offers a more stable and reliable approach to privacy protection in LLMs than previous methods.",
}
| Protecting privacy leakage in large language models remains a paramount challenge. In this paper, we reveal Privacy Seesaw in LLM privacy safeguarding, a phenomenon where measures to secure specific private information inadvertently heighten exposure risks for other privacy. Through comprehensive analysis, we identify the amount of targeted privacy data and the volume of edited privacy neurons as the two central triggers to this issue. To mitigate privacy seesaw, we propose Augmented Privacy Neuron Editing via Activation Patching (APNEAP), a novel framework designed to well balance model performance with privacy protection. The proposed APNEAP augments collected private data by automatically synthesizing new private data, which deactivates the first trigger to the privacy seesaw issue. Additionally, it adapts activation patching to privacy neuron editing for switching off the second trigger to the privacy seesaw problem. Experimental results show that the proposed APNEAP is capable of alleviating the privacy seesaw phenomenon and offers a more stable and reliable approach to privacy protection in LLMs than previous methods. | [
"Wu, Xinwei",
"Dong, Weilong",
"Xu, Shaoyang",
"Xiong, Deyi"
]
| Mitigating Privacy Seesaw in Large Language Models: Augmented Privacy Neuron Editing via Activation Patching | findings-acl.315 | Poster | 2310.20138v2 |
https://aclanthology.org/2024.findings-acl.316.bib | @inproceedings{mascarell-etal-2024-information,
title = "Which Information Matters? Dissecting Human-written Multi-document Summaries with Partial Information Decomposition",
author = "Mascarell, Laura and
LHomme, Yan and
El Helou, Majed",
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.316",
pages = "5333--5338",
abstract = "Understanding the nature of high-quality summaries is crucial to further improve the performance of multi-document summarization. We propose an approach to characterize human-written summaries using partial information decomposition, which decomposes the mutual information provided by all source documents into union, redundancy, synergy, and unique information. Our empirical analysis on different MDS datasets shows that there is a direct dependency between the number of sources and their contribution to the summary.",
}
| Understanding the nature of high-quality summaries is crucial to further improve the performance of multi-document summarization. We propose an approach to characterize human-written summaries using partial information decomposition, which decomposes the mutual information provided by all source documents into union, redundancy, synergy, and unique information. Our empirical analysis on different MDS datasets shows that there is a direct dependency between the number of sources and their contribution to the summary. | [
"Mascarell, Laura",
"LHomme, Yan",
"El Helou, Majed"
]
| Which Information Matters? Dissecting Human-written Multi-document Summaries with Partial Information Decomposition | findings-acl.316 | Poster | 2405.14470v1 |
https://aclanthology.org/2024.findings-acl.317.bib | @inproceedings{yi-etal-2024-badacts,
title = "{B}ad{A}cts: A Universal Backdoor Defense in the Activation Space",
author = "Yi, Biao and
Chen, Sishuo and
Li, Yiming and
Li, Tong and
Zhang, Baolei and
Liu, Zheli",
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.317",
pages = "5339--5352",
abstract = "Backdoor attacks pose an increasingly severe security threat to Deep Neural Networks (DNNs) during their development stage. In response, backdoor sample purification has emerged as a promising defense mechanism, aiming to eliminate backdoor triggers while preserving the integrity of the clean content in the samples. However, existing approaches have been predominantly focused on the word space, which are ineffective against feature-space triggers and significantly impair performance on clean data. To address this, we introduce a universal backdoor defense that purifies backdoor samples in the activation space by drawing abnormal activations towards optimized minimum clean activation distribution intervals. The advantages of our approach are twofold: (1) By operating in the activation space, our method captures from surface-level information like words to higher-level semantic concepts such as syntax, thus counteracting diverse triggers; (2) the fine-grained continuous nature of the activation space allows for more precise preservation of clean content while removing triggers. Furthermore, we propose a detection module based on statistical information of abnormal activations, to achieve a better trade-off between clean accuracy and defending performance. Extensive experiments on diverse datasets and against diverse attacks (including syntax and style attacks) demonstrate that our defense achieves state-of-the-art performance.",
}
| Backdoor attacks pose an increasingly severe security threat to Deep Neural Networks (DNNs) during their development stage. In response, backdoor sample purification has emerged as a promising defense mechanism, aiming to eliminate backdoor triggers while preserving the integrity of the clean content in the samples. However, existing approaches have been predominantly focused on the word space, which are ineffective against feature-space triggers and significantly impair performance on clean data. To address this, we introduce a universal backdoor defense that purifies backdoor samples in the activation space by drawing abnormal activations towards optimized minimum clean activation distribution intervals. The advantages of our approach are twofold: (1) By operating in the activation space, our method captures from surface-level information like words to higher-level semantic concepts such as syntax, thus counteracting diverse triggers; (2) the fine-grained continuous nature of the activation space allows for more precise preservation of clean content while removing triggers. Furthermore, we propose a detection module based on statistical information of abnormal activations, to achieve a better trade-off between clean accuracy and defending performance. Extensive experiments on diverse datasets and against diverse attacks (including syntax and style attacks) demonstrate that our defense achieves state-of-the-art performance. | [
"Yi, Biao",
"Chen, Sishuo",
"Li, Yiming",
"Li, Tong",
"Zhang, Baolei",
"Liu, Zheli"
]
| {B}ad{A}cts: A Universal Backdoor Defense in the Activation Space | findings-acl.317 | Poster | 2405.11227v1 |
https://aclanthology.org/2024.findings-acl.318.bib | @inproceedings{liu-etal-2024-reactxt,
title = "{R}eact{XT}: Understanding Molecular {``}Reaction-ship{''} via Reaction-Contextualized Molecule-Text Pretraining",
author = "Liu, Zhiyuan and
Shi, Yaorui and
Zhang, An and
Li, Sihang and
Zhang, Enzhi and
Wang, Xiang and
Kawaguchi, Kenji and
Chua, Tat-Seng",
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.318",
pages = "5353--5377",
abstract = "Molecule-text modeling, which aims to facilitate molecule-relevant tasks with a textual interface and textual knowledge, is an emerging research direction. Beyond single molecules, studying reaction-text modeling holds promise for helping the synthesis of new materials and drugs. However, previous works mostly neglect reaction-text modeling: they primarily focus on modeling individual molecule-text pairs or learning chemical reactions without texts in context. Additionally, one key task of reaction-text modeling {--} experimental procedure prediction {--} is less explored due to the absence of an open-source dataset. The task is to predict step-by-step actions of conducting chemical experiments and is crucial to automating chemical synthesis. To resolve the challenges above, we propose a new pretraining method, ReactXT, for reaction-text modeling, and a new dataset, OpenExp, for experimental procedure prediction. Specifically, ReactXT features three types of input contexts to incrementally pretrain LMs. Each of the three input contexts corresponds to a pretraining task to improve the text-based understanding of either reactions or single molecules. ReactXT demonstrates consistent improvements in experimental procedure prediction and molecule captioning and offers competitive results in retrosynthesis. Our code is available at https://github.com/syr-cn/ReactXT.",
}
| Molecule-text modeling, which aims to facilitate molecule-relevant tasks with a textual interface and textual knowledge, is an emerging research direction. Beyond single molecules, studying reaction-text modeling holds promise for helping the synthesis of new materials and drugs. However, previous works mostly neglect reaction-text modeling: they primarily focus on modeling individual molecule-text pairs or learning chemical reactions without texts in context. Additionally, one key task of reaction-text modeling {--} experimental procedure prediction {--} is less explored due to the absence of an open-source dataset. The task is to predict step-by-step actions of conducting chemical experiments and is crucial to automating chemical synthesis. To resolve the challenges above, we propose a new pretraining method, ReactXT, for reaction-text modeling, and a new dataset, OpenExp, for experimental procedure prediction. Specifically, ReactXT features three types of input contexts to incrementally pretrain LMs. Each of the three input contexts corresponds to a pretraining task to improve the text-based understanding of either reactions or single molecules. ReactXT demonstrates consistent improvements in experimental procedure prediction and molecule captioning and offers competitive results in retrosynthesis. Our code is available at https://github.com/syr-cn/ReactXT. | [
"Liu, Zhiyuan",
"Shi, Yaorui",
"Zhang, An",
"Li, Sihang",
"Zhang, Enzhi",
"Wang, Xiang",
"Kawaguchi, Kenji",
"Chua, Tat-Seng"
]
| {R}eact{XT}: Understanding Molecular {``}Reaction-ship{''} via Reaction-Contextualized Molecule-Text Pretraining | findings-acl.318 | Poster | 2305.08238v1 |
https://aclanthology.org/2024.findings-acl.319.bib | @inproceedings{yan-etal-2024-multi,
title = "Multi-modal Concept Alignment Pre-training for Generative Medical Visual Question Answering",
author = "Yan, Quan and
Duan, Junwen and
Wang, Jianxin",
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.319",
pages = "5378--5389",
abstract = "Medical Visual Question Answering (Med-VQA) seeks to accurately respond to queries regarding medical images, a task particularly challenging for open-ended questions. This study unveils the Multi-modal Concept Alignment Pre-training (MMCAP) approach for generative Med-VQA, leveraging a knowledge graph sourced from medical image-caption datasets and the Unified Medical Language System. MMCAP advances the fusion of visual and textual medical knowledge via a graph attention network and a transformer decoder. Additionally, it incorporates a Type Conditional Prompt in the fine-tuning phase, markedly boosting the accuracy and relevance of answers to open-ended questions. Our tests on benchmark datasets illustrate MMCAP{'}s superiority over existing methods, demonstrating its high efficiency in data-limited settings and effective knowledge-image alignment capability.",
}
| Medical Visual Question Answering (Med-VQA) seeks to accurately respond to queries regarding medical images, a task particularly challenging for open-ended questions. This study unveils the Multi-modal Concept Alignment Pre-training (MMCAP) approach for generative Med-VQA, leveraging a knowledge graph sourced from medical image-caption datasets and the Unified Medical Language System. MMCAP advances the fusion of visual and textual medical knowledge via a graph attention network and a transformer decoder. Additionally, it incorporates a Type Conditional Prompt in the fine-tuning phase, markedly boosting the accuracy and relevance of answers to open-ended questions. Our tests on benchmark datasets illustrate MMCAP{'}s superiority over existing methods, demonstrating its high efficiency in data-limited settings and effective knowledge-image alignment capability. | [
"Yan, Quan",
"Duan, Junwen",
"Wang, Jianxin"
]
| Multi-modal Concept Alignment Pre-training for Generative Medical Visual Question Answering | findings-acl.319 | Poster | 2305.10415v5 |
https://aclanthology.org/2024.findings-acl.320.bib | @inproceedings{kasa-etal-2024-exploring,
title = "Exploring Ordinality in Text Classification: A Comparative Study of Explicit and Implicit Techniques",
author = "Kasa, Siva Rajesh and
Goel, Aniket and
Gupta, Karan and
Roychowdhury, Sumegh and
Priyatam, Pattisapu and
Bhanushali, Anish and
Srinivasa Murthy, Prasanna",
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.320",
pages = "5390--5404",
abstract = "Ordinal Classification (OC) is a widely encountered challenge in Natural Language Processing (NLP), with applications in various domains such as sentiment analysis, rating prediction, and more. Previous approaches to tackle OC have primarily focused on modifying existing or creating novel loss functions that explicitly account for the ordinal nature of labels. However, with the advent of Pre-trained Language Models (PLMs), it became possible to tackle ordinality through the implicit semantics of the labels as well. This paper provides a comprehensive theoretical and empirical examination of both these approaches. Furthermore, we also offer strategic recommendations regarding the most effective approach to adopt based on specific settings.",
}
| Ordinal Classification (OC) is a widely encountered challenge in Natural Language Processing (NLP), with applications in various domains such as sentiment analysis, rating prediction, and more. Previous approaches to tackle OC have primarily focused on modifying existing or creating novel loss functions that explicitly account for the ordinal nature of labels. However, with the advent of Pre-trained Language Models (PLMs), it became possible to tackle ordinality through the implicit semantics of the labels as well. This paper provides a comprehensive theoretical and empirical examination of both these approaches. Furthermore, we also offer strategic recommendations regarding the most effective approach to adopt based on specific settings. | [
"Kasa, Siva Rajesh",
"Goel, Aniket",
"Gupta, Karan",
"Roychowdhury, Sumegh",
"Priyatam, Pattisapu",
"Bhanushali, Anish",
"Srinivasa Murthy, Prasanna"
]
| Exploring Ordinality in Text Classification: A Comparative Study of Explicit and Implicit Techniques | findings-acl.320 | Poster | 2405.11775v1 |
https://aclanthology.org/2024.findings-acl.321.bib | @inproceedings{gao-etal-2024-evaluating-large,
title = "Evaluating Large Language Models on {W}ikipedia-Style Survey Generation",
author = "Gao, Fan and
Jiang, Hang and
Yang, Rui and
Zeng, Qingcheng and
Lu, Jinghui and
Blum, Moritz and
She, Tianwei and
Jiang, Yuang and
Li, Irene",
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.321",
pages = "5405--5418",
abstract = "Educational materials such as survey articles in specialized fields like computer science traditionally require tremendous expert inputs and are therefore expensive to create and update. Recently, Large Language Models (LLMs) have achieved significant success across various general tasks. However, their effectiveness and limitations in the education domain are yet to be fully explored. In this work, we examine the proficiency of LLMs in generating succinct survey articles specific to the niche field of NLP in computer science, focusing on a curated list of 99 topics. Automated benchmarks reveal that GPT-4 surpasses its predecessors, inluding GPT-3.5, PaLM2, and LLaMa2 by margins ranging from 2{\%} to 20{\%} in comparison to the established ground truth. We compare both human and GPT-based evaluation scores and provide in-depth analysis. While our findings suggest that GPT-created surveys are more contemporary and accessible than human-authored ones, certain limitations were observed. Notably, GPT-4, despite often delivering outstanding content, occasionally exhibited lapses like missing details or factual errors. At last, we compared the rating behavior between humans and GPT-4 and found systematic bias in using GPT evaluation.",
}
| Educational materials such as survey articles in specialized fields like computer science traditionally require tremendous expert inputs and are therefore expensive to create and update. Recently, Large Language Models (LLMs) have achieved significant success across various general tasks. However, their effectiveness and limitations in the education domain are yet to be fully explored. In this work, we examine the proficiency of LLMs in generating succinct survey articles specific to the niche field of NLP in computer science, focusing on a curated list of 99 topics. Automated benchmarks reveal that GPT-4 surpasses its predecessors, inluding GPT-3.5, PaLM2, and LLaMa2 by margins ranging from 2{\%} to 20{\%} in comparison to the established ground truth. We compare both human and GPT-based evaluation scores and provide in-depth analysis. While our findings suggest that GPT-created surveys are more contemporary and accessible than human-authored ones, certain limitations were observed. Notably, GPT-4, despite often delivering outstanding content, occasionally exhibited lapses like missing details or factual errors. At last, we compared the rating behavior between humans and GPT-4 and found systematic bias in using GPT evaluation. | [
"Gao, Fan",
"Jiang, Hang",
"Yang, Rui",
"Zeng, Qingcheng",
"Lu, Jinghui",
"Blum, Moritz",
"She, Tianwei",
"Jiang, Yuang",
"Li, Irene"
]
| Evaluating Large Language Models on {W}ikipedia-Style Survey Generation | findings-acl.321 | Poster | 2206.05395v1 |
https://aclanthology.org/2024.findings-acl.322.bib | @inproceedings{yang-etal-2024-butterfly,
title = "The Butterfly Effect of Model Editing: Few Edits Can Trigger Large Language Models Collapse",
author = "Yang, Wanli and
Sun, Fei and
Ma, Xinyu and
Liu, Xun and
Yin, Dawei and
Cheng, Xueqi",
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.322",
pages = "5419--5437",
abstract = "Although model editing has shown promise in revising knowledge in Large Language Models (LLMs), its impact on the inherent capabilities of LLMs is often overlooked. In this work, we reveal a critical phenomenon: even a single edit can trigger model collapse, manifesting as significant performance degradation in various benchmark tasks. However, benchmarking LLMs after each edit, while necessary to prevent such collapses, is impractically time-consuming and resource-intensive. To mitigate this, we propose using perplexity as a surrogate metric, validated by extensive experiments demonstrating changes in an edited model{'}s perplexity are strongly correlated with its downstream task performances. We further conduct an in-depth study on sequential editing, a practical setting for real-world scenarios, across various editing methods and LLMs, focusing on hard cases from our previous single edit studies. The results indicate that nearly all examined editing methods result in model collapse after only few edits. To facilitate further research, we have utilized GPT-3.5 to develop a new dataset, HardEdit, based on those hard cases. This dataset aims to establish the foundation for pioneering research in reliable model editing and the mechanisms underlying editing-induced model collapse. We hope this work can draw the community{'}s attention to the potential risks inherent in model editing practices.",
}
| Although model editing has shown promise in revising knowledge in Large Language Models (LLMs), its impact on the inherent capabilities of LLMs is often overlooked. In this work, we reveal a critical phenomenon: even a single edit can trigger model collapse, manifesting as significant performance degradation in various benchmark tasks. However, benchmarking LLMs after each edit, while necessary to prevent such collapses, is impractically time-consuming and resource-intensive. To mitigate this, we propose using perplexity as a surrogate metric, validated by extensive experiments demonstrating changes in an edited model{'}s perplexity are strongly correlated with its downstream task performances. We further conduct an in-depth study on sequential editing, a practical setting for real-world scenarios, across various editing methods and LLMs, focusing on hard cases from our previous single edit studies. The results indicate that nearly all examined editing methods result in model collapse after only few edits. To facilitate further research, we have utilized GPT-3.5 to develop a new dataset, HardEdit, based on those hard cases. This dataset aims to establish the foundation for pioneering research in reliable model editing and the mechanisms underlying editing-induced model collapse. We hope this work can draw the community{'}s attention to the potential risks inherent in model editing practices. | [
"Yang, Wanli",
"Sun, Fei",
"Ma, Xinyu",
"Liu, Xun",
"Yin, Dawei",
"Cheng, Xueqi"
]
| The Butterfly Effect of Model Editing: Few Edits Can Trigger Large Language Models Collapse | findings-acl.322 | Poster | 2402.09656v4 |
https://aclanthology.org/2024.findings-acl.323.bib | @inproceedings{li-chu-2024-continually,
title = "Can We Continually Edit Language Models? On the Knowledge Attenuation in Sequential Model Editing",
author = "Li, Qi and
Chu, Xiaowen",
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.323",
pages = "5438--5455",
abstract = "Model editing has become a promising method for precisely and effectively updating knowledge in language models. In this paper, we investigate knowledge attenuation, in which the retention of updated knowledge within the language model decreases as the number of edits increases after sequential editing. Through empirical study, we discovered that existing editing methods generally suffer from knowledge attenuation. We attribute this phenomenon to two aspects: (1) redundant parameters interference and (2) update weight disentanglement. To this end, we propose the AdaPLE method. It not only mitigates the knowledge attenuation issue but also improves the performance on existing benchmarks. To the best of our knowledge, we are the first to investigate the cause and mitigation of knowledge attenuation in sequential LLM editing.",
}
| Model editing has become a promising method for precisely and effectively updating knowledge in language models. In this paper, we investigate knowledge attenuation, in which the retention of updated knowledge within the language model decreases as the number of edits increases after sequential editing. Through empirical study, we discovered that existing editing methods generally suffer from knowledge attenuation. We attribute this phenomenon to two aspects: (1) redundant parameters interference and (2) update weight disentanglement. To this end, we propose the AdaPLE method. It not only mitigates the knowledge attenuation issue but also improves the performance on existing benchmarks. To the best of our knowledge, we are the first to investigate the cause and mitigation of knowledge attenuation in sequential LLM editing. | [
"Li, Qi",
"Chu, Xiaowen"
]
| Can We Continually Edit Language Models? On the Knowledge Attenuation in Sequential Model Editing | findings-acl.323 | Poster | 2405.16821v1 |
https://aclanthology.org/2024.findings-acl.324.bib | @inproceedings{qu-etal-2024-generation,
title = "Before Generation, Align it! A Novel and Effective Strategy for Mitigating Hallucinations in Text-to-{SQL} Generation",
author = "Qu, Ge and
Li, Jinyang and
Li, Bowen and
Qin, Bowen and
Huo, Nan and
Ma, Chenhao and
Cheng, Reynold",
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.324",
pages = "5456--5471",
abstract = "Large Language Models (LLMs) driven by In-Context Learning (ICL) have significantly improved the performance of text-to-SQL. Previous methods generally employ a two-stage reasoning framework, namely 1) schema linking and 2) logical synthesis, making the framework not only effective but also interpretable. Despite these advancements, the inherent bad nature of the generalization of LLMs often results in hallucinations, which limits the full potential of LLMs. In this work, we first identify and categorize the common types of hallucinations at each stage in text-to-SQL. We then introduce a novel strategy, Task Alignment (TA), designed to mitigate hallucinations at each stage. TA encourages LLMs to take advantage of experiences from similar tasks rather than starting the tasks from scratch. This can help LLMs reduce the burden of generalization, thereby mitigating hallucinations effectively. We further propose TA-SQL, a text-to-SQL framework based on this strategy. The experimental results and comprehensive analysis demonstrate the effectiveness and robustness of our framework. Specifically, it enhances the performance of the GPT-4 baseline by 21.23{\%} relatively on BIRD dev and it yields significant improvements across six models and four mainstream, complex text-to-SQL benchmarks.",
}
| Large Language Models (LLMs) driven by In-Context Learning (ICL) have significantly improved the performance of text-to-SQL. Previous methods generally employ a two-stage reasoning framework, namely 1) schema linking and 2) logical synthesis, making the framework not only effective but also interpretable. Despite these advancements, the inherent bad nature of the generalization of LLMs often results in hallucinations, which limits the full potential of LLMs. In this work, we first identify and categorize the common types of hallucinations at each stage in text-to-SQL. We then introduce a novel strategy, Task Alignment (TA), designed to mitigate hallucinations at each stage. TA encourages LLMs to take advantage of experiences from similar tasks rather than starting the tasks from scratch. This can help LLMs reduce the burden of generalization, thereby mitigating hallucinations effectively. We further propose TA-SQL, a text-to-SQL framework based on this strategy. The experimental results and comprehensive analysis demonstrate the effectiveness and robustness of our framework. Specifically, it enhances the performance of the GPT-4 baseline by 21.23{\%} relatively on BIRD dev and it yields significant improvements across six models and four mainstream, complex text-to-SQL benchmarks. | [
"Qu, Ge",
"Li, Jinyang",
"Li, Bowen",
"Qin, Bowen",
"Huo, Nan",
"Ma, Chenhao",
"Cheng, Reynold"
]
| Before Generation, Align it! A Novel and Effective Strategy for Mitigating Hallucinations in Text-to-{SQL} Generation | findings-acl.324 | Poster | 2406.14492v1 |
https://aclanthology.org/2024.findings-acl.325.bib | @inproceedings{lan-etal-2024-translatotron,
title = "Translatotron-{V}(ison): An End-to-End Model for In-Image Machine Translation",
author = "Lan, Zhibin and
Niu, Liqiang and
Meng, Fandong and
Zhou, Jie and
Zhang, Min and
Su, Jinsong",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.325",
pages = "5472--5485",
abstract = "In-image machine translation (IIMT) aims to translate an image containing texts in source language into an image containing translations in target language. In this regard, conventional cascaded methods suffer from issues such as error propagation, massive parameters, and difficulties in deployment and retaining visual characteristics of the input image.Thus, constructing end-to-end models has become an option, which, however, faces two main challenges: 1) the huge modeling burden, as it is required to simultaneously learn alignment across languages and preserve the visual characteristics of the input image; 2) the difficulties of directly predicting excessively lengthy pixel sequences.In this paper, we propose $\textit{Translatotron-V(ision)}$, an end-to-end IIMT model consisting of four modules. In addition to an image encoder, and an image decoder, our model contains a target text decoder and an image tokenizer. Among them, the target text decoder is used to alleviate the language alignment burden, and the image tokenizer converts long sequences of pixels into shorter sequences of visual tokens, preventing the model from focusing on low-level visual features. Besides, we present a two-stage training framework for our model to assist the model in learning alignment across modalities and languages. Finally, we propose a location-aware evaluation metric called Structure-BLEU to assess the translation quality of the generated images. Experimental results demonstrate that our model achieves competitive performance compared to cascaded models with only 70.9{\%} of parameters, and significantly outperforms the pixel-level end-to-end IIMT model.",
}
| In-image machine translation (IIMT) aims to translate an image containing texts in source language into an image containing translations in target language. In this regard, conventional cascaded methods suffer from issues such as error propagation, massive parameters, and difficulties in deployment and retaining visual characteristics of the input image.Thus, constructing end-to-end models has become an option, which, however, faces two main challenges: 1) the huge modeling burden, as it is required to simultaneously learn alignment across languages and preserve the visual characteristics of the input image; 2) the difficulties of directly predicting excessively lengthy pixel sequences.In this paper, we propose $\textit{Translatotron-V(ision)}$, an end-to-end IIMT model consisting of four modules. In addition to an image encoder, and an image decoder, our model contains a target text decoder and an image tokenizer. Among them, the target text decoder is used to alleviate the language alignment burden, and the image tokenizer converts long sequences of pixels into shorter sequences of visual tokens, preventing the model from focusing on low-level visual features. Besides, we present a two-stage training framework for our model to assist the model in learning alignment across modalities and languages. Finally, we propose a location-aware evaluation metric called Structure-BLEU to assess the translation quality of the generated images. Experimental results demonstrate that our model achieves competitive performance compared to cascaded models with only 70.9{\%} of parameters, and significantly outperforms the pixel-level end-to-end IIMT model. | [
"Lan, Zhibin",
"Niu, Liqiang",
"Meng, F",
"ong",
"Zhou, Jie",
"Zhang, Min",
"Su, Jinsong"
]
| Translatotron-{V}(ison): An End-to-End Model for In-Image Machine Translation | findings-acl.325 | Poster | 2407.02894v1 |
https://aclanthology.org/2024.findings-acl.326.bib | @inproceedings{nooralahzadeh-etal-2024-statbot,
title = "{S}tat{B}ot.{S}wiss: Bilingual Open Data Exploration in Natural Language",
author = {Nooralahzadeh, Farhad and
Zhang, Yi and
Smith, Ellery and
Maennel, Sabine and
Matthey-Doret, Cyril and
De Fondeville, Rapha{\"e}l and
Stockinger, Kurt},
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.326",
pages = "5486--5507",
abstract = "The potential for improvements brought by Large Language Models (LLMs) in Text-to-SQL systems is mostly assessed on monolingual English datasets. However, LLMs{'} performance for other languages remains vastly unexplored. In this work, we release the StatBot.Swiss dataset, the first bilingual benchmark for evaluating Text-to-SQL systems based on real-world applications. The StatBot.Swiss dataset contains 455 natural language/SQL-pairs over 35 big databases with varying level of complexity for both English and German.We evaluate the performance of state-of-the-art LLMs such as GPT-3.5-Turbo and mixtral-8x7b-instruct for the Text-to-SQL translation task using an in-context learning approach. Our experimental analysis illustrates that current LLMs struggle to generalize well in generating SQL queries on our novel bilingual dataset.",
}
| The potential for improvements brought by Large Language Models (LLMs) in Text-to-SQL systems is mostly assessed on monolingual English datasets. However, LLMs{'} performance for other languages remains vastly unexplored. In this work, we release the StatBot.Swiss dataset, the first bilingual benchmark for evaluating Text-to-SQL systems based on real-world applications. The StatBot.Swiss dataset contains 455 natural language/SQL-pairs over 35 big databases with varying level of complexity for both English and German.We evaluate the performance of state-of-the-art LLMs such as GPT-3.5-Turbo and mixtral-8x7b-instruct for the Text-to-SQL translation task using an in-context learning approach. Our experimental analysis illustrates that current LLMs struggle to generalize well in generating SQL queries on our novel bilingual dataset. | [
"Nooralahzadeh, Farhad",
"Zhang, Yi",
"Smith, Ellery",
"Maennel, Sabine",
"Matthey-Doret, Cyril",
"De Fondeville, Rapha{\\\"e}l",
"Stockinger, Kurt"
]
| {S}tat{B}ot.{S}wiss: Bilingual Open Data Exploration in Natural Language | findings-acl.326 | Poster | 1705.00894v1 |
https://aclanthology.org/2024.findings-acl.327.bib | @inproceedings{ren-etal-2024-subtle,
title = "Subtle Signatures, Strong Shields: Advancing Robust and Imperceptible Watermarking in Large Language Models",
author = "Ren, Yubing and
Guo, Ping and
Cao, Yanan and
Ma, Wei",
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.327",
pages = "5508--5519",
abstract = "The widespread adoption of Large Language Models (LLMs) has led to an increase in AI-generated text on the Internet, presenting a crucial challenge to differentiate AI-created content from human-written text. This challenge is critical to prevent issues of authenticity, trust, and potential copyright violations. Current research focuses on watermarking LLM-generated text, but traditional techniques struggle to balance robustness with text quality. We introduce a novel watermarking approach, Robust and Imperceptible Watermarking (RIW) for LLMs, which leverages token prior probabilities to improve detectability and maintain watermark imperceptibility. RIW methodically embeds watermarks by partitioning selected tokens into two distinct groups based on their prior probabilities and employing tailored strategies for each group. In the detection stage, the RIW method employs the {`}voted z-test{'} to provide a statistically robust framework to identify the presence of a watermark accurately. The effectiveness of RIW is evaluated across three key dimensions: success rate, text quality, and robustness against removal attacks. Our experimental results on various LLMs, including GPT2-XL, OPT-1.3B, and LLaMA2-7B, indicate that RIW surpasses existing models, and also exhibits increased robustness against various attacks and good imperceptibility, thus promoting the responsible use of LLMs.",
}
| The widespread adoption of Large Language Models (LLMs) has led to an increase in AI-generated text on the Internet, presenting a crucial challenge to differentiate AI-created content from human-written text. This challenge is critical to prevent issues of authenticity, trust, and potential copyright violations. Current research focuses on watermarking LLM-generated text, but traditional techniques struggle to balance robustness with text quality. We introduce a novel watermarking approach, Robust and Imperceptible Watermarking (RIW) for LLMs, which leverages token prior probabilities to improve detectability and maintain watermark imperceptibility. RIW methodically embeds watermarks by partitioning selected tokens into two distinct groups based on their prior probabilities and employing tailored strategies for each group. In the detection stage, the RIW method employs the {`}voted z-test{'} to provide a statistically robust framework to identify the presence of a watermark accurately. The effectiveness of RIW is evaluated across three key dimensions: success rate, text quality, and robustness against removal attacks. Our experimental results on various LLMs, including GPT2-XL, OPT-1.3B, and LLaMA2-7B, indicate that RIW surpasses existing models, and also exhibits increased robustness against various attacks and good imperceptibility, thus promoting the responsible use of LLMs. | [
"Ren, Yubing",
"Guo, Ping",
"Cao, Yanan",
"Ma, Wei"
]
| Subtle Signatures, Strong Shields: Advancing Robust and Imperceptible Watermarking in Large Language Models | findings-acl.327 | Poster | 2405.13517v1 |
https://aclanthology.org/2024.findings-acl.328.bib | @inproceedings{shuang-etal-2024-thinking,
title = "Thinking about how to extract: Energizing {LLM}s{'} emergence capabilities for document-level event argument extraction",
author = "Shuang, Kai and
Zhouji, Zhouji and
Qiwei, Wang and
Guo, Jinyu",
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.328",
pages = "5520--5532",
abstract = "There are two key challenges remaining for the document-level event argument extraction (D-EAE) tasks: key feature forgetting and cross-event argument confusion. The emergence capability of large language models (LLMs) holds promise for solving the above two challenges. In this paper, we propose a document-level event argument extraction method based on guided summarization and reasoning (EAESR), which leverages the emergence capabilities of LLMs to highlight key event information and to clarify the explicit and implicit association between multiple events. Specifically, we generate document summarization information that shorten the length of the event context while preserving the key event features. In addition, we generate inter-event reasoning information, which helps EAESR make sense of the correlations between events and reduces their dependence on the event context, especially to better cope with the few-shot D-EAE task. Then, we obtain named entity information to enable EAESR to learn argument boundary features to improve the sensitivity of its argument boundary recognition. Eventually, we fused the above features and sentence features to make EAESR have summarizing and reasoning capabilities simultaneously. Extensive experiments on WIKIEVENTS and RAMS have shown that EAESR achieves a new state-of-the-art that outperforms the baseline models by 1.3{\%} F1 and 1.6{\%} F1, respectively, and averages 11{\%} F1 in few-shot settings.",
}
| There are two key challenges remaining for the document-level event argument extraction (D-EAE) tasks: key feature forgetting and cross-event argument confusion. The emergence capability of large language models (LLMs) holds promise for solving the above two challenges. In this paper, we propose a document-level event argument extraction method based on guided summarization and reasoning (EAESR), which leverages the emergence capabilities of LLMs to highlight key event information and to clarify the explicit and implicit association between multiple events. Specifically, we generate document summarization information that shorten the length of the event context while preserving the key event features. In addition, we generate inter-event reasoning information, which helps EAESR make sense of the correlations between events and reduces their dependence on the event context, especially to better cope with the few-shot D-EAE task. Then, we obtain named entity information to enable EAESR to learn argument boundary features to improve the sensitivity of its argument boundary recognition. Eventually, we fused the above features and sentence features to make EAESR have summarizing and reasoning capabilities simultaneously. Extensive experiments on WIKIEVENTS and RAMS have shown that EAESR achieves a new state-of-the-art that outperforms the baseline models by 1.3{\%} F1 and 1.6{\%} F1, respectively, and averages 11{\%} F1 in few-shot settings. | [
"Shuang, Kai",
"Zhouji, Zhouji",
"Qiwei, Wang",
"Guo, Jinyu"
]
| Thinking about how to extract: Energizing {LLM}s{'} emergence capabilities for document-level event argument extraction | findings-acl.328 | Poster | 2401.13218v1 |
https://aclanthology.org/2024.findings-acl.329.bib | @inproceedings{si-etal-2024-improving,
title = "Improving the Robustness of Distantly-Supervised Named Entity Recognition via Uncertainty-Aware Teacher Learning and Student-Student Collaborative Learning",
author = "Si, Shuzheng and
Hu, Helan and
Zhao, Haozhe and
Zeng, Shuang and
An, Kaikai and
Cai, Zefan and
Chang, Baobao",
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.329",
pages = "5533--5546",
abstract = "Distantly-Supervised Named Entity Recognition (DS-NER) effectively alleviates the burden of annotation, but meanwhile suffers from the label noise. Recent works attempt to adopt the teacher-student framework to gradually refine the training labels and improve the overall robustness. However, we argue that these teacher-student methods achieve limited performance because the poor calibration of the teacher network produces incorrectly pseudo-labeled samples, leading to error propagation. Therefore, we attempt to mitigate this issue by proposing: (1) Uncertainty-Aware Teacher Learning that leverages the prediction uncertainty to reduce the number of incorrect pseudo labels in the self-training stage; (2) Student-Student Collaborative Learning that allows the transfer of reliable labels between two student networks instead of indiscriminately relying on all pseudo labels from its teacher. This approach further enables a full exploration of mislabeled samples rather than simply filtering unreliable pseudo-labeled samples. We evaluate our proposed method on five DS-NER datasets, demonstrating that our method is superior to the state-of-the-art DS-NER denoising methods.",
}
| Distantly-Supervised Named Entity Recognition (DS-NER) effectively alleviates the burden of annotation, but meanwhile suffers from the label noise. Recent works attempt to adopt the teacher-student framework to gradually refine the training labels and improve the overall robustness. However, we argue that these teacher-student methods achieve limited performance because the poor calibration of the teacher network produces incorrectly pseudo-labeled samples, leading to error propagation. Therefore, we attempt to mitigate this issue by proposing: (1) Uncertainty-Aware Teacher Learning that leverages the prediction uncertainty to reduce the number of incorrect pseudo labels in the self-training stage; (2) Student-Student Collaborative Learning that allows the transfer of reliable labels between two student networks instead of indiscriminately relying on all pseudo labels from its teacher. This approach further enables a full exploration of mislabeled samples rather than simply filtering unreliable pseudo-labeled samples. We evaluate our proposed method on five DS-NER datasets, demonstrating that our method is superior to the state-of-the-art DS-NER denoising methods. | [
"Si, Shuzheng",
"Hu, Helan",
"Zhao, Haozhe",
"Zeng, Shuang",
"An, Kaikai",
"Cai, Zefan",
"Chang, Baobao"
]
| Improving the Robustness of Distantly-Supervised Named Entity Recognition via Uncertainty-Aware Teacher Learning and Student-Student Collaborative Learning | findings-acl.329 | Poster | 2311.08010v2 |
https://aclanthology.org/2024.findings-acl.330.bib | @inproceedings{rowlands-etal-2024-predicting,
title = "Predicting Narratives of Climate Obstruction in Social Media Advertising",
author = "Rowlands, Harri and
Morio, Gaku and
Tanner, Dylan and
Manning, Christopher",
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.330",
pages = "5547--5558",
abstract = "Social media advertising offers a platform for fossil fuel value chain companies and their agents to reinforce their narratives, often emphasizing economic, labor market, and energy security benefits to promote oil and gas policy and products. Whether such narratives can be detected automatically and the extent to which the cost of human annotation can be reduced is our research question. We introduce a task of classifying narratives into seven categories, based on existing definitions and data.Experiments showed that RoBERTa-large outperforms other methods, while GPT-4 Turbo can serve as a viable annotator for the task, thereby reducing human annotation costs. Our findings and insights provide guidance to automate climate-related ad analysis and lead to more scalable ad scrutiny.",
}
| Social media advertising offers a platform for fossil fuel value chain companies and their agents to reinforce their narratives, often emphasizing economic, labor market, and energy security benefits to promote oil and gas policy and products. Whether such narratives can be detected automatically and the extent to which the cost of human annotation can be reduced is our research question. We introduce a task of classifying narratives into seven categories, based on existing definitions and data.Experiments showed that RoBERTa-large outperforms other methods, while GPT-4 Turbo can serve as a viable annotator for the task, thereby reducing human annotation costs. Our findings and insights provide guidance to automate climate-related ad analysis and lead to more scalable ad scrutiny. | [
"Rowl",
"s, Harri",
"Morio, Gaku",
"Tanner, Dylan",
"Manning, Christopher"
]
| Predicting Narratives of Climate Obstruction in Social Media Advertising | findings-acl.330 | Poster | 2305.06174v2 |
https://aclanthology.org/2024.findings-acl.331.bib | @inproceedings{wang-etal-2024-sss,
title = "{SSS}: Editing Factual Knowledge in Language Models towards Semantic Sparse Space",
author = "Wang, Huazheng and
Sun, Haifeng and
Wang, Jingyu and
Qi, Qi and
Xia, Zixuan and
Zhang, Menghao and
Liao, Jianxin",
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.331",
pages = "5559--5570",
abstract = "Language Models (LMs) acquire factual knowledge during pre-training and store it in the parameters, which can be valuable for downstream tasks. As world evolves, some facts may be incorrectly induced or become obsolete over time. Various model editing methods have been proposed to modify specific examples in LMs. However, existing training-based methods still suffer from sub-optimal locality, where irrelevant neighborhood examples can be adversely influenced. Model{'}s gradients are still struggling to identify the appropriate direction when updating the parameters. To address this issue, we find that directing the hidden state of the edit example towards spaces where semantics are sparse tends to help preserve the semantics of irrelevant neighborhood examples. Based on this hypothesis, we propose a novel metric, named SSS, to evaluate the degree of sparsity around a sentence embedding in the semantic space without any human or machine annotation. Subsequently, we incorporate SSS into the original loss function of the existing training-based methods to enhance locality. Experiments conducted on two datasets across various models demonstrate that SSS is effective in improving both locality and reasoning capability.",
}
| Language Models (LMs) acquire factual knowledge during pre-training and store it in the parameters, which can be valuable for downstream tasks. As world evolves, some facts may be incorrectly induced or become obsolete over time. Various model editing methods have been proposed to modify specific examples in LMs. However, existing training-based methods still suffer from sub-optimal locality, where irrelevant neighborhood examples can be adversely influenced. Model{'}s gradients are still struggling to identify the appropriate direction when updating the parameters. To address this issue, we find that directing the hidden state of the edit example towards spaces where semantics are sparse tends to help preserve the semantics of irrelevant neighborhood examples. Based on this hypothesis, we propose a novel metric, named SSS, to evaluate the degree of sparsity around a sentence embedding in the semantic space without any human or machine annotation. Subsequently, we incorporate SSS into the original loss function of the existing training-based methods to enhance locality. Experiments conducted on two datasets across various models demonstrate that SSS is effective in improving both locality and reasoning capability. | [
"Wang, Huazheng",
"Sun, Haifeng",
"Wang, Jingyu",
"Qi, Qi",
"Xia, Zixuan",
"Zhang, Menghao",
"Liao, Jianxin"
]
| {SSS}: Editing Factual Knowledge in Language Models towards Semantic Sparse Space | findings-acl.331 | Poster | 2406.16416v1 |
https://aclanthology.org/2024.findings-acl.332.bib | @inproceedings{cai-etal-2024-geohard,
title = "$\textit{GeoHard}$: Towards Measuring Class-wise Hardness through Modelling Class Semantics",
author = "Cai, Fengyu and
Zhao, Xinran and
Zhang, Hongming and
Gurevych, Iryna and
Koeppl, Heinz",
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.332",
pages = "5571--5597",
abstract = "Recent advances in measuring hardness-wise properties of data guide language models in sample selection within low-resource scenarios. However, class-specific properties are overlooked for task setup and learning. How will these properties influence model learning and is it generalizable across datasets? To answer this question, this work formally initiates the concept of $\textit{class-wise hardness}$. Experiments across eight natural language understanding (NLU) datasets demonstrate a consistent hardness distribution across learning paradigms, models, and human judgment. Subsequent experiments unveil a notable challenge in measuring such class-wise hardness with instance-level metrics in previous works. To address this, we propose $\textit{GeoHard}$ for class-wise hardness measurement by modeling class geometry in the semantic embedding space. $\textit{GeoHard}$ surpasses instance-level metrics by over 59 percent on $\textit{Pearson}${`}s correlation on measuring class-wise hardness. Our analysis theoretically and empirically underscores the generality of $\textit{GeoHard}$ as a fresh perspective on data diagnosis. Additionally, we showcase how understanding class-wise hardness can practically aid in improving task learning.",
}
| Recent advances in measuring hardness-wise properties of data guide language models in sample selection within low-resource scenarios. However, class-specific properties are overlooked for task setup and learning. How will these properties influence model learning and is it generalizable across datasets? To answer this question, this work formally initiates the concept of $\textit{class-wise hardness}$. Experiments across eight natural language understanding (NLU) datasets demonstrate a consistent hardness distribution across learning paradigms, models, and human judgment. Subsequent experiments unveil a notable challenge in measuring such class-wise hardness with instance-level metrics in previous works. To address this, we propose $\textit{GeoHard}$ for class-wise hardness measurement by modeling class geometry in the semantic embedding space. $\textit{GeoHard}$ surpasses instance-level metrics by over 59 percent on $\textit{Pearson}${`}s correlation on measuring class-wise hardness. Our analysis theoretically and empirically underscores the generality of $\textit{GeoHard}$ as a fresh perspective on data diagnosis. Additionally, we showcase how understanding class-wise hardness can practically aid in improving task learning. | [
"Cai, Fengyu",
"Zhao, Xinran",
"Zhang, Hongming",
"Gurevych, Iryna",
"Koeppl, Heinz"
]
| $\textit{GeoHard}$: Towards Measuring Class-wise Hardness through Modelling Class Semantics | findings-acl.332 | Poster | 2407.12512v1 |
https://aclanthology.org/2024.findings-acl.333.bib | @inproceedings{wei-etal-2024-unveiling,
title = "Unveiling Selection Biases: Exploring Order and Token Sensitivity in Large Language Models",
author = "Wei, Sheng-Lun and
Wu, Cheng-Kuang and
Huang, Hen-Hsen and
Chen, Hsin-Hsi",
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.333",
pages = "5598--5621",
abstract = "In this paper, we investigate the phenomena of {``}selection biases{''} in Large Language Models (LLMs), focusing on problems where models are tasked with choosing the optimal option from an ordered sequence. We delve into biases related to option order and token usage, which significantly impact LLMs{'} decision-making processes. We also quantify the impact of these biases through an extensive empirical analysis across multiple models and tasks. Furthermore, we propose mitigation strategies to enhance model performance. Our key contributions are threefold: 1) Precisely quantifying the influence of option order and token on LLMs, 2) Developing strategies to mitigate the impact of token and order sensitivity to enhance robustness, and 3) Offering a detailed analysis of sensitivity across models and tasks, which informs the creation of more stable and reliable LLM applications for selection problems.",
}
| In this paper, we investigate the phenomena of {``}selection biases{''} in Large Language Models (LLMs), focusing on problems where models are tasked with choosing the optimal option from an ordered sequence. We delve into biases related to option order and token usage, which significantly impact LLMs{'} decision-making processes. We also quantify the impact of these biases through an extensive empirical analysis across multiple models and tasks. Furthermore, we propose mitigation strategies to enhance model performance. Our key contributions are threefold: 1) Precisely quantifying the influence of option order and token on LLMs, 2) Developing strategies to mitigate the impact of token and order sensitivity to enhance robustness, and 3) Offering a detailed analysis of sensitivity across models and tasks, which informs the creation of more stable and reliable LLM applications for selection problems. | [
"Wei, Sheng-Lun",
"Wu, Cheng-Kuang",
"Huang, Hen-Hsen",
"Chen, Hsin-Hsi"
]
| Unveiling Selection Biases: Exploring Order and Token Sensitivity in Large Language Models | findings-acl.333 | Poster | 2406.03009v1 |
https://aclanthology.org/2024.findings-acl.334.bib | @inproceedings{koto-etal-2024-arabicmmlu,
title = "{A}rabic{MMLU}: Assessing Massive Multitask Language Understanding in {A}rabic",
author = "Koto, Fajri and
Li, Haonan and
Shatnawi, Sara and
Doughman, Jad and
Sadallah, Abdelrahman and
Alraeesi, Aisha and
Almubarak, Khalid and
Alyafeai, Zaid and
Sengupta, Neha and
Shehata, Shady and
Habash, Nizar and
Nakov, Preslav and
Baldwin, Timothy",
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.334",
pages = "5622--5640",
abstract = "The focus of language model evaluation has transitioned towards reasoning and knowledge-intensive tasks, driven by advancements in pretraining large models. While state-of-the-art models are partially trained on large Arabic texts, evaluating their performance in Arabic remains challenging due to the limited availability of relevant datasets. To bridge this gap, we present ArabicMMLU, the first multi-task language understanding benchmark for the Arabic language, sourced from school exams across diverse educational levels in different countries spanning North Africa, the Levant, and the Gulf regions. Our data comprises 40 tasks and 14,575 multiple-choice questions in Modern Standard Arabic (MSA) and is carefully constructed by collaborating with native speakers in the region. Our comprehensive evaluations of 35 models reveal substantial room for improvement, particularly among the best open-source models. Notably, BLOOMZ, mT0, LLama2, and Falcon struggle to achieve a score of 50{\%}, while even the top-performing Arabic-centric model only achieves a score of 62.3{\%}.",
}
| The focus of language model evaluation has transitioned towards reasoning and knowledge-intensive tasks, driven by advancements in pretraining large models. While state-of-the-art models are partially trained on large Arabic texts, evaluating their performance in Arabic remains challenging due to the limited availability of relevant datasets. To bridge this gap, we present ArabicMMLU, the first multi-task language understanding benchmark for the Arabic language, sourced from school exams across diverse educational levels in different countries spanning North Africa, the Levant, and the Gulf regions. Our data comprises 40 tasks and 14,575 multiple-choice questions in Modern Standard Arabic (MSA) and is carefully constructed by collaborating with native speakers in the region. Our comprehensive evaluations of 35 models reveal substantial room for improvement, particularly among the best open-source models. Notably, BLOOMZ, mT0, LLama2, and Falcon struggle to achieve a score of 50{\%}, while even the top-performing Arabic-centric model only achieves a score of 62.3{\%}. | [
"Koto, Fajri",
"Li, Haonan",
"Shatnawi, Sara",
"Doughman, Jad",
"Sadallah, Abdelrahman",
"Alraeesi, Aisha",
"Almubarak, Khalid",
"Alyafeai, Zaid",
"Sengupta, Neha",
"Shehata, Shady",
"Habash, Nizar",
"Nakov, Preslav",
"Baldwin, Timothy"
]
| {A}rabic{MMLU}: Assessing Massive Multitask Language Understanding in {A}rabic | findings-acl.334 | Poster | 2402.12840v2 |
https://aclanthology.org/2024.findings-acl.335.bib | @inproceedings{muskardin-etal-2024-relationship,
title = "On the Relationship Between {RNN} Hidden-State Vectors and Semantic Structures",
author = "Muskardin, Edi and
Tappler, Martin and
Pill, Ingo and
Aichernig, Bernhard and
Pock, Thomas",
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.335",
pages = "5641--5658",
abstract = "We examine the assumption that hidden-state vectors of recurrent neural networks (RNNs) tend to form clusters of semantically similar vectors, which we dub the clustering hypothesis. While this hypothesis has been assumed in RNN analyses in recent years, its validity has not been studied thoroughly on modern RNN architectures. We first consider RNNs that were trained to recognize regular languages. This enables us to draw on perfect ground-truth automata in our evaluation, against which we can compare the RNN{'}s accuracy and the distribution of the hidden-state vectors. Then, we consider context-free languages to examine if RNN states form clusters for more expressive languages.For our analysis, we fit (generalized) linear models to classify RNN states into automata states and we apply different unsupervised clustering techniques. With a new ambiguity score, derived from information entropy, we measure how well an abstraction function maps the hidden state vectors to abstract clusters. Our evaluation supports the validity of the clustering hypothesis for regular languages, especially if RNNs are well-trained, i.e., clustering techniques succeed in finding clusters of similar state vectors. However, the clustering accuracy decreases substantially for context-free languages. This suggests that clustering is not a reliable abstraction technique for RNNs used in tasks like natural language processing.",
}
| We examine the assumption that hidden-state vectors of recurrent neural networks (RNNs) tend to form clusters of semantically similar vectors, which we dub the clustering hypothesis. While this hypothesis has been assumed in RNN analyses in recent years, its validity has not been studied thoroughly on modern RNN architectures. We first consider RNNs that were trained to recognize regular languages. This enables us to draw on perfect ground-truth automata in our evaluation, against which we can compare the RNN{'}s accuracy and the distribution of the hidden-state vectors. Then, we consider context-free languages to examine if RNN states form clusters for more expressive languages.For our analysis, we fit (generalized) linear models to classify RNN states into automata states and we apply different unsupervised clustering techniques. With a new ambiguity score, derived from information entropy, we measure how well an abstraction function maps the hidden state vectors to abstract clusters. Our evaluation supports the validity of the clustering hypothesis for regular languages, especially if RNNs are well-trained, i.e., clustering techniques succeed in finding clusters of similar state vectors. However, the clustering accuracy decreases substantially for context-free languages. This suggests that clustering is not a reliable abstraction technique for RNNs used in tasks like natural language processing. | [
"Muskardin, Edi",
"Tappler, Martin",
"Pill, Ingo",
"Aichernig, Bernhard",
"Pock, Thomas"
]
| On the Relationship Between {RNN} Hidden-State Vectors and Semantic Structures | findings-acl.335 | Poster | 2306.16854v1 |
https://aclanthology.org/2024.findings-acl.336.bib | @inproceedings{liu-etal-2024-xmc,
title = "{XMC}-Agent : Dynamic Navigation over Scalable Hierarchical Index for Incremental Extreme Multi-label Classification",
author = "Liu, Yanjiang and
Zhong, Tianyun and
Lu, Yaojie and
Lin, Hongyu and
He, Ben 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.336",
pages = "5659--5672",
abstract = "The eXtreme Multi-label Classification (XMC) aims at accurately assigning large-scale labels to instances, and is challenging for learning, managing, and predicting over the large-scale and rapidly growing set of labels. Traditional XMC methods, like one-vs-all and tree-based methods struggle with the growing set of labels due to their static label assumptions, and embedding-based methods struggle with the complex mapping relationships due to their late-interaction paradigm. In this paper, we propose a large language model (LLM) powered agent framework for extreme multi-label classification {--} XMC-Agent, which can effectively learn, manage and predict the extremely large and dynamically increasing set of labels. Specifically, XMC-Agent models the extreme multi-label classification task as a dynamic navigation problem, employing a scalable hierarchical label index to effectively manage the unified label space. Additionally, we propose two algorithms to enhance the dynamic navigation capabilities of XMC-Agent: a self-construction algorithm for building the scalable hierarchical index, and an iterative feedback learning algorithm for adjusting the agent to specific tasks. Experiments show that XMC-Agentachieves the state-of-the-art performance on three standard datasets.",
}
| The eXtreme Multi-label Classification (XMC) aims at accurately assigning large-scale labels to instances, and is challenging for learning, managing, and predicting over the large-scale and rapidly growing set of labels. Traditional XMC methods, like one-vs-all and tree-based methods struggle with the growing set of labels due to their static label assumptions, and embedding-based methods struggle with the complex mapping relationships due to their late-interaction paradigm. In this paper, we propose a large language model (LLM) powered agent framework for extreme multi-label classification {--} XMC-Agent, which can effectively learn, manage and predict the extremely large and dynamically increasing set of labels. Specifically, XMC-Agent models the extreme multi-label classification task as a dynamic navigation problem, employing a scalable hierarchical label index to effectively manage the unified label space. Additionally, we propose two algorithms to enhance the dynamic navigation capabilities of XMC-Agent: a self-construction algorithm for building the scalable hierarchical index, and an iterative feedback learning algorithm for adjusting the agent to specific tasks. Experiments show that XMC-Agentachieves the state-of-the-art performance on three standard datasets. | [
"Liu, Yanjiang",
"Zhong, Tianyun",
"Lu, Yaojie",
"Lin, Hongyu",
"He, Ben",
"Zhou, Shuheng",
"Zhu, Huijia",
"Wang, Weiqiang",
"Liu, Zhongyi",
"Han, Xianpei",
"Sun, Le"
]
| {XMC}-Agent : Dynamic Navigation over Scalable Hierarchical Index for Incremental Extreme Multi-label Classification | findings-acl.336 | Poster | 2202.13018v1 |
https://aclanthology.org/2024.findings-acl.337.bib | @inproceedings{zhu-etal-2024-benchmarking,
title = "Benchmarking Large Language Models on {CFLUE} - A {C}hinese Financial Language Understanding Evaluation Dataset",
author = "Zhu, Jie and
Li, Junhui and
Wen, Yalong and
Guo, Lifan",
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.337",
pages = "5673--5693",
abstract = "In light of recent breakthroughs in large language models (LLMs) that have revolutionized natural language processing (NLP), there is an urgent need for new benchmarks to keep pace with the fast development of LLMs. In this paper, we propose CFLUE, the Chinese Financial Language Understanding Evaluation benchmark, designed to assess the capability of LLMs across various dimensions. Specifically, CFLUE provides datasets tailored for both knowledge assessment and application assessment. In knowledge assessment, it consists of 38K+ multiple-choice questions with associated solution explanations. These questions serve dual purposes: answer prediction and question reasoning. In application assessment, CFLUE features 16K+ test instances across distinct groups of NLP tasks such as text classification, machine translation, relation extraction, reading comprehension, and text generation. Upon CFLUE, we conduct a thorough evaluation of representative LLMs. The results reveal that only Qwen-72B, GPT-4, and GPT-4-turbo achieve an accuracy exceeding 60{\%} in answer prediction for knowledge assessment, suggesting that there is still substantial room for improvement in current LLMs. In application assessment, while GPT-4 and GPT-4-turbo rank as the top two performers on average, their significant advantage over open-source LLMs is noticeably diminished, given that Qwen-72B achieves the best performance in 2 out of 5 tasks. The datasets and scripts associated with CFLUE are openly accessible at \url{https://github.com/aliyun/cflue}.",
}
| In light of recent breakthroughs in large language models (LLMs) that have revolutionized natural language processing (NLP), there is an urgent need for new benchmarks to keep pace with the fast development of LLMs. In this paper, we propose CFLUE, the Chinese Financial Language Understanding Evaluation benchmark, designed to assess the capability of LLMs across various dimensions. Specifically, CFLUE provides datasets tailored for both knowledge assessment and application assessment. In knowledge assessment, it consists of 38K+ multiple-choice questions with associated solution explanations. These questions serve dual purposes: answer prediction and question reasoning. In application assessment, CFLUE features 16K+ test instances across distinct groups of NLP tasks such as text classification, machine translation, relation extraction, reading comprehension, and text generation. Upon CFLUE, we conduct a thorough evaluation of representative LLMs. The results reveal that only Qwen-72B, GPT-4, and GPT-4-turbo achieve an accuracy exceeding 60{\%} in answer prediction for knowledge assessment, suggesting that there is still substantial room for improvement in current LLMs. In application assessment, while GPT-4 and GPT-4-turbo rank as the top two performers on average, their significant advantage over open-source LLMs is noticeably diminished, given that Qwen-72B achieves the best performance in 2 out of 5 tasks. The datasets and scripts associated with CFLUE are openly accessible at \url{https://github.com/aliyun/cflue}. | [
"Zhu, Jie",
"Li, Junhui",
"Wen, Yalong",
"Guo, Lifan"
]
| Benchmarking Large Language Models on {CFLUE} - A {C}hinese Financial Language Understanding Evaluation Dataset | findings-acl.337 | Poster | 2307.03972v1 |
https://aclanthology.org/2024.findings-acl.338.bib | @inproceedings{chen-etal-2024-improving,
title = "Improving Large Language Models via Fine-grained Reinforcement Learning with Minimum Editing Constraint",
author = "Chen, Zhipeng and
Zhou, Kun and
Zhao, Xin and
Wan, Junchen and
Zhang, Fuzheng and
Zhang, Di 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.338",
pages = "5694--5711",
abstract = "Reinforcement learning (RL) has been widely used in training large language models (LLMs) for preventing unexpected outputs, e.g., reducing harmfulness and errors. However, existing RL methods mainly adopt instance-level reward, which cannot provide fine-grained supervision for complex reasoning tasks. As a result, the RL training cannot be fully aware of the specific part or step that actually leads to the incorrectness in model response. To address it, we propose a new RL method named RLMEC that incorporates a generative model as the reward model, which is trained by the erroneous solution rewriting task under the minimum editing constraint, which can produce token-level supervision for RL training. Based 0on the generative reward model, we design the token-level RL objective for training and an imitation-based regularization for stabilizing RL process. And these two objectives focus on the revision of the key tokens for the erroneous solution, reducing the effect of other unimportant tokens. Experiment results on 8 tasks have demonstrated the effectiveness of our approach. Our code and data will be publicly released.",
}
| Reinforcement learning (RL) has been widely used in training large language models (LLMs) for preventing unexpected outputs, e.g., reducing harmfulness and errors. However, existing RL methods mainly adopt instance-level reward, which cannot provide fine-grained supervision for complex reasoning tasks. As a result, the RL training cannot be fully aware of the specific part or step that actually leads to the incorrectness in model response. To address it, we propose a new RL method named RLMEC that incorporates a generative model as the reward model, which is trained by the erroneous solution rewriting task under the minimum editing constraint, which can produce token-level supervision for RL training. Based 0on the generative reward model, we design the token-level RL objective for training and an imitation-based regularization for stabilizing RL process. And these two objectives focus on the revision of the key tokens for the erroneous solution, reducing the effect of other unimportant tokens. Experiment results on 8 tasks have demonstrated the effectiveness of our approach. Our code and data will be publicly released. | [
"Chen, Zhipeng",
"Zhou, Kun",
"Zhao, Xin",
"Wan, Junchen",
"Zhang, Fuzheng",
"Zhang, Di",
"Wen, Ji-Rong"
]
| Improving Large Language Models via Fine-grained Reinforcement Learning with Minimum Editing Constraint | findings-acl.338 | Poster | 2401.06081v2 |
https://aclanthology.org/2024.findings-acl.339.bib | @inproceedings{fedorova-etal-2024-definition,
title = "Definition generation for lexical semantic change detection",
author = "Fedorova, Mariia and
Kutuzov, Andrey and
Scherrer, Yves",
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.339",
pages = "5712--5724",
abstract = "We use contextualized word definitions generated by large language models as semantic representations in the task of diachronic lexical semantic change detection (LSCD). In short, generated definitions are used as {`}senses{'}, and the change score of a target word is retrieved by comparing their distributions in two time periods under comparison. On the material of five datasets and three languages, we show that generated definitions are indeed specific and general enough to convey a signal sufficient to rank sets of words by the degree of their semantic change over time. Our approach is on par with or outperforms prior non-supervised sense-based LSCD methods. At the same time, it preserves interpretability and allows to inspect the reasons behind a specific shift in terms of discrete definitions-as-senses. This is another step in the direction of explainable semantic change modeling.",
}
| We use contextualized word definitions generated by large language models as semantic representations in the task of diachronic lexical semantic change detection (LSCD). In short, generated definitions are used as {`}senses{'}, and the change score of a target word is retrieved by comparing their distributions in two time periods under comparison. On the material of five datasets and three languages, we show that generated definitions are indeed specific and general enough to convey a signal sufficient to rank sets of words by the degree of their semantic change over time. Our approach is on par with or outperforms prior non-supervised sense-based LSCD methods. At the same time, it preserves interpretability and allows to inspect the reasons behind a specific shift in terms of discrete definitions-as-senses. This is another step in the direction of explainable semantic change modeling. | [
"Fedorova, Mariia",
"Kutuzov, Andrey",
"Scherrer, Yves"
]
| Definition generation for lexical semantic change detection | findings-acl.339 | Poster | 2406.14167v2 |
https://aclanthology.org/2024.findings-acl.340.bib | @inproceedings{costa-jussa-etal-2024-mutox,
title = "{M}u{T}ox: Universal {MU}ltilingual Audio-based {TOX}icity Dataset and Zero-shot Detector",
author = "Costa-juss{\`a}, Marta and
Meglioli, Mariano and
Andrews, Pierre and
Dale, David and
Hansanti, Prangthip and
Kalbassi, Elahe and
Mourachko, Alexandre and
Ropers, Christophe and
Wood, Carleigh",
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.340",
pages = "5725--5734",
abstract = "Research in toxicity detection in natural language processing for the speech modality (audio-based) is quite limited, particularly for languages other than English. To address these limitations and lay the groundwork for truly multilingual audio-based toxicity detection, we introduce MuTox, the first highly multilingual audio-based dataset with toxicity labels which covers 14 different linguistic families. The dataset comprises 20,000 audio utterances for English and Spanish, and 4,000 for the other 28 languages. To demonstrate the quality of this dataset, we trained the MuTox audio-based toxicity classifier, which enables zero-shot toxicity detection across a wide range of languages. This classifier performs on par with existing text-based trainable classifiers, while expanding the language coverage more than tenfold. When compared to a wordlist-based classifier that covers a similar number of languages, MuTox improves F1-Score by an average of 100{\%}. This significant improvement underscores the potential of MuTox in advancing the field of audio-based toxicity detection.",
}
| Research in toxicity detection in natural language processing for the speech modality (audio-based) is quite limited, particularly for languages other than English. To address these limitations and lay the groundwork for truly multilingual audio-based toxicity detection, we introduce MuTox, the first highly multilingual audio-based dataset with toxicity labels which covers 14 different linguistic families. The dataset comprises 20,000 audio utterances for English and Spanish, and 4,000 for the other 28 languages. To demonstrate the quality of this dataset, we trained the MuTox audio-based toxicity classifier, which enables zero-shot toxicity detection across a wide range of languages. This classifier performs on par with existing text-based trainable classifiers, while expanding the language coverage more than tenfold. When compared to a wordlist-based classifier that covers a similar number of languages, MuTox improves F1-Score by an average of 100{\%}. This significant improvement underscores the potential of MuTox in advancing the field of audio-based toxicity detection. | [
"Costa-juss{\\`a}, Marta",
"Meglioli, Mariano",
"Andrews, Pierre",
"Dale, David",
"Hansanti, Prangthip",
"Kalbassi, Elahe",
"Mourachko, Alex",
"re",
"Ropers, Christophe",
"Wood, Carleigh"
]
| {M}u{T}ox: Universal {MU}ltilingual Audio-based {TOX}icity Dataset and Zero-shot Detector | findings-acl.340 | Poster | 2401.05060v2 |
https://aclanthology.org/2024.findings-acl.341.bib | @inproceedings{pang-etal-2024-phased,
title = "Phased Instruction Fine-Tuning for Large Language Models",
author = "Pang, Wei and
Zhou, Chuan and
Zhou, Xiao-Hua and
Wang, Xiaojie",
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.341",
pages = "5735--5748",
abstract = "Instruction Fine-Tuning, a method enhancing pre-trained language models{'} capabilities from mere next-word prediction to complex instruction following, often employs a one-off training approach on diverse instruction dataset. However, this method may not effectively enhance models{'} adherence to instructions due to the simultaneous handling of varying instruction complexities. To address this, we propose a novel phased instruction fine-tuning (Phased IFT) method, grounded in the hypothesis of progressive alignment, which posits that the transition of a pre-trained language model from simple next-word prediction to sophisticated instruction following is a gradual learning process. Specifically, we obtain the score of difficulty for each instruction via GPT-4, stratify the instruction data into subsets of increasing difficulty, and sequentially uptrain on these subsets using the standard supervised loss. Through extensive experiments on the pre-trained models Llama-2 7B/13B, and Mistral-7B using the 52K Alpaca instruction data, we demonstrate that Phased IFT significantly surpasses traditional one-off instruction fine-tuning (One-off IFT) method in win rate, empirically validating the progressive alignment hypothesis. Our findings suggest that Phased IFT offers a simple yet effective pathway for elevating the instruction-following capabilities of pre-trained language models.",
}
| Instruction Fine-Tuning, a method enhancing pre-trained language models{'} capabilities from mere next-word prediction to complex instruction following, often employs a one-off training approach on diverse instruction dataset. However, this method may not effectively enhance models{'} adherence to instructions due to the simultaneous handling of varying instruction complexities. To address this, we propose a novel phased instruction fine-tuning (Phased IFT) method, grounded in the hypothesis of progressive alignment, which posits that the transition of a pre-trained language model from simple next-word prediction to sophisticated instruction following is a gradual learning process. Specifically, we obtain the score of difficulty for each instruction via GPT-4, stratify the instruction data into subsets of increasing difficulty, and sequentially uptrain on these subsets using the standard supervised loss. Through extensive experiments on the pre-trained models Llama-2 7B/13B, and Mistral-7B using the 52K Alpaca instruction data, we demonstrate that Phased IFT significantly surpasses traditional one-off instruction fine-tuning (One-off IFT) method in win rate, empirically validating the progressive alignment hypothesis. Our findings suggest that Phased IFT offers a simple yet effective pathway for elevating the instruction-following capabilities of pre-trained language models. | [
"Pang, Wei",
"Zhou, Chuan",
"Zhou, Xiao-Hua",
"Wang, Xiaojie"
]
| Phased Instruction Fine-Tuning for Large Language Models | findings-acl.341 | Poster | 2406.16346v1 |
https://aclanthology.org/2024.findings-acl.342.bib | @inproceedings{zhuang-etal-2024-toree,
title = "{TOREE}: Evaluating Topic Relevance of Student Essays for {C}hinese Primary and Middle School Education",
author = "Zhuang, Xinlin and
Wu, Hongyi and
Shen, Xinshu and
Yu, Peimin and
Yi, Gaowei and
Chen, Xinhao and
Hu, Tu and
Chen, Yang and
Ren, Yupei and
Zhang, Yadong and
Song, Youqi and
Liu, Binxuan and
Lan, Man",
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.342",
pages = "5749--5765",
abstract = "Topic relevance of an essay demands that the composition adheres to a clear theme and aligns well with the essay prompt requirements, a critical aspect of essay quality evaluation. However, existing research of Automatic Essay Scoring (AES) for Chinese essays has overlooked topic relevance and lacks detailed feedback, while Automatic Essay Comment Generation (AECG) faces much complexity and difficulty. Additionally, current Large Language Models, including GPT-4, often make incorrect judgments and provide overly impractical feedback when evaluating topic relevance. This paper introduces \textbf{TOREE} (\textbf{To}pic \textbf{Re}levance \textbf{E}valuation), a comprehensive dataset developed to assess topic relevance in Chinese primary and middle school students{'} essays, which is beneficial for AES, AECG and other applications. Moreover, our proposed two-step method utilizes TOREE through a combination of Supervised Fine-tuning and Preference Learning. Experimental results demonstrate that TOREE is of high quality, and our method significantly enhances models{'} performance on two designed tasks for topic relevance evaluation, improving both automatic and human evaluations across four diverse LLMs.",
}
| Topic relevance of an essay demands that the composition adheres to a clear theme and aligns well with the essay prompt requirements, a critical aspect of essay quality evaluation. However, existing research of Automatic Essay Scoring (AES) for Chinese essays has overlooked topic relevance and lacks detailed feedback, while Automatic Essay Comment Generation (AECG) faces much complexity and difficulty. Additionally, current Large Language Models, including GPT-4, often make incorrect judgments and provide overly impractical feedback when evaluating topic relevance. This paper introduces \textbf{TOREE} (\textbf{To}pic \textbf{Re}levance \textbf{E}valuation), a comprehensive dataset developed to assess topic relevance in Chinese primary and middle school students{'} essays, which is beneficial for AES, AECG and other applications. Moreover, our proposed two-step method utilizes TOREE through a combination of Supervised Fine-tuning and Preference Learning. Experimental results demonstrate that TOREE is of high quality, and our method significantly enhances models{'} performance on two designed tasks for topic relevance evaluation, improving both automatic and human evaluations across four diverse LLMs. | [
"Zhuang, Xinlin",
"Wu, Hongyi",
"Shen, Xinshu",
"Yu, Peimin",
"Yi, Gaowei",
"Chen, Xinhao",
"Hu, Tu",
"Chen, Yang",
"Ren, Yupei",
"Zhang, Yadong",
"Song, Youqi",
"Liu, Binxuan",
"Lan, Man"
]
| {TOREE}: Evaluating Topic Relevance of Student Essays for {C}hinese Primary and Middle School Education | findings-acl.342 | Poster | 2306.17342v2 |
https://aclanthology.org/2024.findings-acl.343.bib | @inproceedings{cai-etal-2024-predicting,
title = "Predicting the Unpredictable: Uncertainty-Aware Reasoning over Temporal Knowledge Graphs via Diffusion Process",
author = "Cai, Yuxiang and
Liu, Qiao and
Gan, Yanglei and
Li, Changlin and
Liu, Xueyi and
Lin, Run and
Luo, Da and
JiayeYang, JiayeYang",
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.343",
pages = "5766--5778",
abstract = "Temporal Knowledge Graph (TKG) reasoning seeks to predict future incomplete facts leveraging historical data. While existing approaches have shown effectiveness in addressing the task through various perspectives, such as graph learning and logic rules, they are limited in capturing the indeterminacy in future events, particularly in the case of rare/unseen facts. To tackle the highlighted issues, we introduce a novel approach by conceptualizing TKG reasoning as a sequence denoising process for future facts, namely DiffuTKG. Concretely, we first encodes the historical events as the conditional sequence. Then we gradually introduce Gaussian noise to corrupt target facts during the forward process and then employ a transformer-based conditional denoiser to restore them in the reverse phase. Moreover, we introduce an uncertainty regularization loss to mitigate the risk of prediction biases by favoring frequent scenarios over rare/unseen facts. Empirical results on four real-world datasets show that DiffuTKG outperforms state-of-the-art methods across multiple evaluation metrics.",
}
| Temporal Knowledge Graph (TKG) reasoning seeks to predict future incomplete facts leveraging historical data. While existing approaches have shown effectiveness in addressing the task through various perspectives, such as graph learning and logic rules, they are limited in capturing the indeterminacy in future events, particularly in the case of rare/unseen facts. To tackle the highlighted issues, we introduce a novel approach by conceptualizing TKG reasoning as a sequence denoising process for future facts, namely DiffuTKG. Concretely, we first encodes the historical events as the conditional sequence. Then we gradually introduce Gaussian noise to corrupt target facts during the forward process and then employ a transformer-based conditional denoiser to restore them in the reverse phase. Moreover, we introduce an uncertainty regularization loss to mitigate the risk of prediction biases by favoring frequent scenarios over rare/unseen facts. Empirical results on four real-world datasets show that DiffuTKG outperforms state-of-the-art methods across multiple evaluation metrics. | [
"Cai, Yuxiang",
"Liu, Qiao",
"Gan, Yanglei",
"Li, Changlin",
"Liu, Xueyi",
"Lin, Run",
"Luo, Da",
"JiayeYang, JiayeYang"
]
| Predicting the Unpredictable: Uncertainty-Aware Reasoning over Temporal Knowledge Graphs via Diffusion Process | findings-acl.343 | Poster | 2405.14170v1 |
https://aclanthology.org/2024.findings-acl.344.bib | @inproceedings{shahgir-etal-2024-asymmetric,
title = "Asymmetric Bias in Text-to-Image Generation with Adversarial Attacks",
author = "Shahgir, Haz and
Kong, Xianghao and
Ver Steeg, Greg and
Dong, 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.344",
pages = "5779--5796",
abstract = "The widespread use of Text-to-Image (T2I) models in content generation requires careful examination of their safety, including their robustness to adversarial attacks. Despite extensive research on adversarial attacks, the reasons for their effectiveness remain underexplored. This paper presents an empirical study on adversarial attacks against T2I models, focusing on analyzing factors associated with attack success rates (ASR). We introduce a new attack objective - entity swapping using adversarial suffixes and two gradient-based attack algorithms. Human and automatic evaluations reveal the asymmetric nature of ASRs on entity swap: for example, it is easier to replace {``}human{''} with {``}robot{''} in the prompt {``}a human dancing in the rain.{''} with an adversarial suffix, but the reverse replacement is significantly harder. We further propose probing metrics to establish indicative signals from the model{'}s beliefs to the adversarial ASR. We identify conditions that result in a success probability of 60{\%} for adversarial attacks and others where this likelihood drops below 5{\%}. The code and data are available at https://github.com/Patchwork53/AsymmetricAttack",
}
| The widespread use of Text-to-Image (T2I) models in content generation requires careful examination of their safety, including their robustness to adversarial attacks. Despite extensive research on adversarial attacks, the reasons for their effectiveness remain underexplored. This paper presents an empirical study on adversarial attacks against T2I models, focusing on analyzing factors associated with attack success rates (ASR). We introduce a new attack objective - entity swapping using adversarial suffixes and two gradient-based attack algorithms. Human and automatic evaluations reveal the asymmetric nature of ASRs on entity swap: for example, it is easier to replace {``}human{''} with {``}robot{''} in the prompt {``}a human dancing in the rain.{''} with an adversarial suffix, but the reverse replacement is significantly harder. We further propose probing metrics to establish indicative signals from the model{'}s beliefs to the adversarial ASR. We identify conditions that result in a success probability of 60{\%} for adversarial attacks and others where this likelihood drops below 5{\%}. The code and data are available at https://github.com/Patchwork53/AsymmetricAttack | [
"Shahgir, Haz",
"Kong, Xianghao",
"Ver Steeg, Greg",
"Dong, Yue"
]
| Asymmetric Bias in Text-to-Image Generation with Adversarial Attacks | findings-acl.344 | Poster | 2312.14440v3 |
https://aclanthology.org/2024.findings-acl.345.bib | @inproceedings{liang-etal-2024-controlled,
title = "Controlled Text Generation for Large Language Model with Dynamic Attribute Graphs",
author = "Liang, Xun and
Wang, Hanyu and
Song, Shichao and
Hu, Mengting and
Wang, Xunzhi and
Li, Zhiyu and
Xiong, Feiyu and
Tang, 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.345",
pages = "5797--5814",
abstract = "Controlled Text Generation (CTG) aims to produce texts that exhibit specific desired attributes. In this study, we introduce a pluggable CTG framework for Large Language Models (LLMs) named Dynamic Attribute Graphs-based controlled text generation (DATG). This framework utilizes an attribute scorer to evaluate the attributes of sentences generated by LLMs and constructs dynamic attribute graphs. DATG modulates the occurrence of key attribute words and key anti-attribute words, achieving effective attribute control without compromising the original capabilities of the model. We conduct experiments across four datasets in two tasks: toxicity mitigation and sentiment transformation, employing five LLMs as foundational models. Our findings highlight a remarkable enhancement in control accuracy, achieving a peak improvement of 19.29{\%} over baseline methods in the most favorable task across four datasets. Additionally, we observe a significant decrease in perplexity, markedly improving text fluency.",
}
| Controlled Text Generation (CTG) aims to produce texts that exhibit specific desired attributes. In this study, we introduce a pluggable CTG framework for Large Language Models (LLMs) named Dynamic Attribute Graphs-based controlled text generation (DATG). This framework utilizes an attribute scorer to evaluate the attributes of sentences generated by LLMs and constructs dynamic attribute graphs. DATG modulates the occurrence of key attribute words and key anti-attribute words, achieving effective attribute control without compromising the original capabilities of the model. We conduct experiments across four datasets in two tasks: toxicity mitigation and sentiment transformation, employing five LLMs as foundational models. Our findings highlight a remarkable enhancement in control accuracy, achieving a peak improvement of 19.29{\%} over baseline methods in the most favorable task across four datasets. Additionally, we observe a significant decrease in perplexity, markedly improving text fluency. | [
"Liang, Xun",
"Wang, Hanyu",
"Song, Shichao",
"Hu, Mengting",
"Wang, Xunzhi",
"Li, Zhiyu",
"Xiong, Feiyu",
"Tang, Bo"
]
| Controlled Text Generation for Large Language Model with Dynamic Attribute Graphs | findings-acl.345 | Poster | 2402.11218v2 |
https://aclanthology.org/2024.findings-acl.346.bib | @inproceedings{park-etal-2024-coconut,
title = "Coconut: Contextualized Commonsense Unified Transformers for Graph-Based Commonsense Augmentation of Language Models",
author = "Park, Jun-Hyung and
Lee, Mingyu and
Kim, Junho and
Lee, SangKeun",
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.346",
pages = "5815--5830",
abstract = "In this paper, we introduce COCONUT to effectively guide the contextualization of structured commonsense knowledge based on largelanguage models. COCONUT employs a contextualized knowledge prompting scheme to gather high-quality contextualization examplesfrom a large language model. These examples are subsequently distilled into small language models to enhance their contextualization capability. Extensive evaluations show that COCONUT considerably improves commonsense reasoning performance across diverse benchmarks, models, and settings, exhibiting its flexibility and universality in generating contextualized commonsense knowledge. Notably,COCONUT consistently outperforms the state-of-the-art technique by an average of 5.8{\%}.",
}
| In this paper, we introduce COCONUT to effectively guide the contextualization of structured commonsense knowledge based on largelanguage models. COCONUT employs a contextualized knowledge prompting scheme to gather high-quality contextualization examplesfrom a large language model. These examples are subsequently distilled into small language models to enhance their contextualization capability. Extensive evaluations show that COCONUT considerably improves commonsense reasoning performance across diverse benchmarks, models, and settings, exhibiting its flexibility and universality in generating contextualized commonsense knowledge. Notably,COCONUT consistently outperforms the state-of-the-art technique by an average of 5.8{\%}. | [
"Park, Jun-Hyung",
"Lee, Mingyu",
"Kim, Junho",
"Lee, SangKeun"
]
| Coconut: Contextualized Commonsense Unified Transformers for Graph-Based Commonsense Augmentation of Language Models | findings-acl.346 | Poster | 2210.12887v1 |
https://aclanthology.org/2024.findings-acl.347.bib | @inproceedings{mela-etal-2024-mass,
title = "Mass-Editing Memory with Attention in Transformers: A cross-lingual exploration of knowledge",
author = "Mela, Daniel and
Gonzalez-Agirre, Aitor and
Hernando, Javier and
Villegas, Marta",
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.347",
pages = "5831--5847",
abstract = "Recent research has explored methods for updating and modifying factual knowledge in large language models, often focusing on specific multi-layer perceptron blocks. This study expands on this work by examining the effectiveness of existing knowledge editing methods across languages and delving into the role of attention mechanisms in this process. Drawing from the insights gained, we propose Mass-Editing Memory with Attention in Transformers (MEMAT), a method that achieves significant improvements in all metrics while requiring minimal parameter modifications. MEMAT delivers a remarkable 10{\%} increase in magnitude metrics, benefits languages not included in the training data and also demonstrates a high degree of portability. Our code and data are at https://github.com/dtamayo-nlp/MEMAT.",
}
| Recent research has explored methods for updating and modifying factual knowledge in large language models, often focusing on specific multi-layer perceptron blocks. This study expands on this work by examining the effectiveness of existing knowledge editing methods across languages and delving into the role of attention mechanisms in this process. Drawing from the insights gained, we propose Mass-Editing Memory with Attention in Transformers (MEMAT), a method that achieves significant improvements in all metrics while requiring minimal parameter modifications. MEMAT delivers a remarkable 10{\%} increase in magnitude metrics, benefits languages not included in the training data and also demonstrates a high degree of portability. Our code and data are at https://github.com/dtamayo-nlp/MEMAT. | [
"Mela, Daniel",
"Gonzalez-Agirre, Aitor",
"Hern",
"o, Javier",
"Villegas, Marta"
]
| Mass-Editing Memory with Attention in Transformers: A cross-lingual exploration of knowledge | findings-acl.347 | Poster | 1712.01813v1 |
https://aclanthology.org/2024.findings-acl.348.bib | @inproceedings{labrak-etal-2024-biomistral,
title = "{B}io{M}istral: A Collection of Open-Source Pretrained Large Language Models for Medical Domains",
author = "Labrak, Yanis and
Bazoge, Adrien and
Morin, Emmanuel and
Gourraud, Pierre-Antoine and
Rouvier, Mickael and
Dufour, Richard",
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.348",
pages = "5848--5864",
abstract = "Large Language Models (LLMs) have demonstrated remarkable versatility in recent years, offering potential applications across specialized domains such as healthcare and medicine. Despite the availability of various open-source LLMs tailored for health contexts, adapting general-purpose LLMs to the medical domain presents significant challenges.In this paper, we introduce BioMistral, an open-source LLM tailored for the biomedical domain, utilizing Mistral as its foundation model and further pre-trained on PubMed Central. We conduct a comprehensive evaluation of BioMistral on a benchmark comprising 10 established medical question-answering (QA) tasks in English. We also explore lightweight models obtained through quantization and model merging approaches. Our results demonstrate BioMistral{'}s superior performance compared to existing open-source medical models and its competitive edge against proprietary counterparts. Finally, to address the limited availability of data beyond English and to assess the multilingual generalization of medical LLMs, we automatically translated and evaluated this benchmark into 7 other languages. This marks the first large-scale multilingual evaluation of LLMs in the medical domain. Datasets, multilingual evaluation benchmarks, scripts, and all the models obtained during our experiments are freely released.",
}
| Large Language Models (LLMs) have demonstrated remarkable versatility in recent years, offering potential applications across specialized domains such as healthcare and medicine. Despite the availability of various open-source LLMs tailored for health contexts, adapting general-purpose LLMs to the medical domain presents significant challenges.In this paper, we introduce BioMistral, an open-source LLM tailored for the biomedical domain, utilizing Mistral as its foundation model and further pre-trained on PubMed Central. We conduct a comprehensive evaluation of BioMistral on a benchmark comprising 10 established medical question-answering (QA) tasks in English. We also explore lightweight models obtained through quantization and model merging approaches. Our results demonstrate BioMistral{'}s superior performance compared to existing open-source medical models and its competitive edge against proprietary counterparts. Finally, to address the limited availability of data beyond English and to assess the multilingual generalization of medical LLMs, we automatically translated and evaluated this benchmark into 7 other languages. This marks the first large-scale multilingual evaluation of LLMs in the medical domain. Datasets, multilingual evaluation benchmarks, scripts, and all the models obtained during our experiments are freely released. | [
"Labrak, Yanis",
"Bazoge, Adrien",
"Morin, Emmanuel",
"Gourraud, Pierre-Antoine",
"Rouvier, Mickael",
"Dufour, Richard"
]
| {B}io{M}istral: A Collection of Open-Source Pretrained Large Language Models for Medical Domains | findings-acl.348 | Poster | 2303.17636v1 |
https://aclanthology.org/2024.findings-acl.349.bib | @inproceedings{wang-etal-2024-languages,
title = "All Languages Matter: On the Multilingual Safety of {LLM}s",
author = "Wang, Wenxuan and
Tu, Zhaopeng and
Chen, Chang and
Yuan, Youliang and
Huang, Jen-tse and
Jiao, Wenxiang and
Lyu, Michael",
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.349",
pages = "5865--5877",
abstract = "Safety lies at the core of developing and deploying large language models (LLMs). However, previous safety benchmarks only concern the safety in one language, e.g. the majority language in the pretraining data such as English. In this work, we build the first multilingual safety benchmark for LLMs, XSafety, in response to the global deployment of LLMs in practice. XSafety covers 14 kinds of commonly used safety issues across 10 languages that span several language families. We utilize XSafety to empirically study the multilingual safety for 4 widely-used LLMs, including both close-API and open-source models. Experimental results show that all LLMs produce significantly more unsafe responses for non-English queries than English ones, indicating the necessity of developing safety alignment for non-English languages. In addition, we propose a simple and effective prompting method to improve the multilingual safety of ChatGPT by enhancing cross-lingual generalization of safety alignment. Our prompting method can significantly reduce the ratio of unsafe responses by 42{\%} for non-English queries. We will release all the data and results to facilitate future research on LLMs{'} safety.",
}
| Safety lies at the core of developing and deploying large language models (LLMs). However, previous safety benchmarks only concern the safety in one language, e.g. the majority language in the pretraining data such as English. In this work, we build the first multilingual safety benchmark for LLMs, XSafety, in response to the global deployment of LLMs in practice. XSafety covers 14 kinds of commonly used safety issues across 10 languages that span several language families. We utilize XSafety to empirically study the multilingual safety for 4 widely-used LLMs, including both close-API and open-source models. Experimental results show that all LLMs produce significantly more unsafe responses for non-English queries than English ones, indicating the necessity of developing safety alignment for non-English languages. In addition, we propose a simple and effective prompting method to improve the multilingual safety of ChatGPT by enhancing cross-lingual generalization of safety alignment. Our prompting method can significantly reduce the ratio of unsafe responses by 42{\%} for non-English queries. We will release all the data and results to facilitate future research on LLMs{'} safety. | [
"Wang, Wenxuan",
"Tu, Zhaopeng",
"Chen, Chang",
"Yuan, Youliang",
"Huang, Jen-tse",
"Jiao, Wenxiang",
"Lyu, Michael"
]
| All Languages Matter: On the Multilingual Safety of {LLM}s | findings-acl.349 | Poster | 2310.00905v2 |
https://aclanthology.org/2024.findings-acl.350.bib | @inproceedings{zhang-etal-2024-ljpcheck,
title = "{LJPC}heck: Functional Tests for Legal Judgment Prediction",
author = "Zhang, Yuan and
Huang, Wanhong and
Feng, Yi and
Li, Chuanyi and
Fei, Zhiwei and
Ge, Jidong and
Luo, Bin and
Ng, Vincent",
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.350",
pages = "5878--5894",
abstract = "Legal Judgment Prediction (LJP) refers to the task of automatically predicting judgment results (e.g., charges, law articles and term of penalty) given the fact description of cases. While SOTA models have achieved high accuracy and F1 scores on public datasets, existing datasets fail to evaluate specific aspects of these models (e.g., legal fairness, which significantly impact their applications in real scenarios). Inspired by functional testing in software engineering, we introduce LJPCHECK, a suite of functional tests for LJP models, to comprehend LJP models{'} behaviors and offer diagnostic insights. We illustrate the utility of LJPCHECK on five SOTA LJP models. Extensive experiments reveal vulnerabilities in these models, prompting an in-depth discussion into the underlying reasons of their shortcomings.",
}
| Legal Judgment Prediction (LJP) refers to the task of automatically predicting judgment results (e.g., charges, law articles and term of penalty) given the fact description of cases. While SOTA models have achieved high accuracy and F1 scores on public datasets, existing datasets fail to evaluate specific aspects of these models (e.g., legal fairness, which significantly impact their applications in real scenarios). Inspired by functional testing in software engineering, we introduce LJPCHECK, a suite of functional tests for LJP models, to comprehend LJP models{'} behaviors and offer diagnostic insights. We illustrate the utility of LJPCHECK on five SOTA LJP models. Extensive experiments reveal vulnerabilities in these models, prompting an in-depth discussion into the underlying reasons of their shortcomings. | [
"Zhang, Yuan",
"Huang, Wanhong",
"Feng, Yi",
"Li, Chuanyi",
"Fei, Zhiwei",
"Ge, Jidong",
"Luo, Bin",
"Ng, Vincent"
]
| {LJPC}heck: Functional Tests for Legal Judgment Prediction | findings-acl.350 | Poster | 2312.09591v2 |
https://aclanthology.org/2024.findings-acl.351.bib | @inproceedings{huang-etal-2024-cmdl,
title = "{CMDL}: A Large-Scale {C}hinese Multi-Defendant Legal Judgment Prediction Dataset",
author = "Huang, Wanhong and
Feng, Yi and
Li, Chuanyi and
Wu, Honghan and
Ge, Jidong and
Ng, Vincent",
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.351",
pages = "5895--5906",
abstract = "Legal Judgment Prediction (LJP) has attracted significant attention in recent years. However, previous studies have primarily focused on cases involving only a single defendant, skipping multi-defendant cases due to complexity and difficulty. To advance research, we introduce CMDL, a large-scale real-world Chinese Multi-Defendant LJP dataset, which consists of over 393,945 cases with nearly 1.2 million defendants in total. For performance evaluation, we propose case-level evaluation metrics dedicated for the multi-defendant scenario. Experimental results on CMDL show existing SOTA approaches demonstrate weakness when applied to cases involving multiple defendants. We highlight several challenges that require attention and resolution.",
}
| Legal Judgment Prediction (LJP) has attracted significant attention in recent years. However, previous studies have primarily focused on cases involving only a single defendant, skipping multi-defendant cases due to complexity and difficulty. To advance research, we introduce CMDL, a large-scale real-world Chinese Multi-Defendant LJP dataset, which consists of over 393,945 cases with nearly 1.2 million defendants in total. For performance evaluation, we propose case-level evaluation metrics dedicated for the multi-defendant scenario. Experimental results on CMDL show existing SOTA approaches demonstrate weakness when applied to cases involving multiple defendants. We highlight several challenges that require attention and resolution. | [
"Huang, Wanhong",
"Feng, Yi",
"Li, Chuanyi",
"Wu, Honghan",
"Ge, Jidong",
"Ng, Vincent"
]
| {CMDL}: A Large-Scale {C}hinese Multi-Defendant Legal Judgment Prediction Dataset | findings-acl.351 | Poster | 2312.09591v2 |
https://aclanthology.org/2024.findings-acl.352.bib | @inproceedings{gangadhar-stratos-2024-model,
title = "Model Editing by Standard Fine-Tuning",
author = "Gangadhar, Govind Krishnan and
Stratos, Karl",
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.352",
pages = "5907--5913",
abstract = "Standard fine-tuning is considered not as effective as specialized methods for model editing due to its comparatively poor performance. However, it is simple, agnostic to the architectural details of the model being edited, and able to leverage advances in standard training techniques with no additional work (e.g., black-box PEFT for computational efficiency), making it an appealing choice for a model editor. In this work, we show that standard fine-tuning alone can yield competitive model editing performance with two minor modifications. First, we optimize the conditional likelihood rather than the full likelihood. Second, in addition to the typical practice of training on randomly paraphrased edit prompts to encourage generalization, we also train on random or similar unedited facts to encourage locality. Our experiments on the ZsRE and CounterFact datasets demonstrate that these simple modifications allow standard fine-tuning to match or outperform highly specialized editors in terms of edit score.",
}
| Standard fine-tuning is considered not as effective as specialized methods for model editing due to its comparatively poor performance. However, it is simple, agnostic to the architectural details of the model being edited, and able to leverage advances in standard training techniques with no additional work (e.g., black-box PEFT for computational efficiency), making it an appealing choice for a model editor. In this work, we show that standard fine-tuning alone can yield competitive model editing performance with two minor modifications. First, we optimize the conditional likelihood rather than the full likelihood. Second, in addition to the typical practice of training on randomly paraphrased edit prompts to encourage generalization, we also train on random or similar unedited facts to encourage locality. Our experiments on the ZsRE and CounterFact datasets demonstrate that these simple modifications allow standard fine-tuning to match or outperform highly specialized editors in terms of edit score. | [
"Gangadhar, Govind Krishnan",
"Stratos, Karl"
]
| Model Editing by Standard Fine-Tuning | findings-acl.352 | Poster | 2207.10858v1 |
https://aclanthology.org/2024.findings-acl.353.bib | @inproceedings{bao-etal-2024-abstract,
title = "{A}bstract {M}eaning {R}epresentation-Based Logic-Driven Data Augmentation for Logical Reasoning",
author = "Bao, Qiming and
Peng, Alex and
Deng, Zhenyun and
Zhong, Wanjun and
Gendron, Gael and
Pistotti, Timothy and
Tan, Neset and
Young, Nathan and
Chen, Yang and
Zhu, Yonghua and
Denny, Paul and
Witbrock, Michael and
Liu, Jiamou",
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.353",
pages = "5914--5934",
abstract = "Combining large language models with logical reasoning enhances their capacity to address problems in a robust and reliable manner. Nevertheless, the intricate nature of logical reasoning poses challenges when gathering reliable data from the web to build comprehensive training datasets, subsequently affecting performance on downstream tasks. To address this, we introduce a novel logic-driven data augmentation approach, AMR-LDA. AMR-LDA converts the original text into an Abstract Meaning Representation (AMR) graph, a structured semantic representation that encapsulates the logical structure of the sentence, upon which operations are performed to generate logically modified AMR graphs. The modified AMR graphs are subsequently converted back into text to create augmented data. Notably, our methodology is architecture-agnostic and enhances both generative large language models, such as GPT-3.5 and GPT-4, through prompt augmentation, and discriminative large language models through contrastive learning with logic-driven data augmentation. Empirical evidence underscores the efficacy of our proposed method with improvement in performance across seven downstream tasks, such as reading comprehension requiring logical reasoning, textual entailment, and natural language inference. Furthermore, our method leads on the ReClor leaderboard. The source code and data are publicly available",
}
| Combining large language models with logical reasoning enhances their capacity to address problems in a robust and reliable manner. Nevertheless, the intricate nature of logical reasoning poses challenges when gathering reliable data from the web to build comprehensive training datasets, subsequently affecting performance on downstream tasks. To address this, we introduce a novel logic-driven data augmentation approach, AMR-LDA. AMR-LDA converts the original text into an Abstract Meaning Representation (AMR) graph, a structured semantic representation that encapsulates the logical structure of the sentence, upon which operations are performed to generate logically modified AMR graphs. The modified AMR graphs are subsequently converted back into text to create augmented data. Notably, our methodology is architecture-agnostic and enhances both generative large language models, such as GPT-3.5 and GPT-4, through prompt augmentation, and discriminative large language models through contrastive learning with logic-driven data augmentation. Empirical evidence underscores the efficacy of our proposed method with improvement in performance across seven downstream tasks, such as reading comprehension requiring logical reasoning, textual entailment, and natural language inference. Furthermore, our method leads on the ReClor leaderboard. The source code and data are publicly available | [
"Bao, Qiming",
"Peng, Alex",
"Deng, Zhenyun",
"Zhong, Wanjun",
"Gendron, Gael",
"Pistotti, Timothy",
"Tan, Neset",
"Young, Nathan",
"Chen, Yang",
"Zhu, Yonghua",
"Denny, Paul",
"Witbrock, Michael",
"Liu, Jiamou"
]
| {A}bstract {M}eaning {R}epresentation-Based Logic-Driven Data Augmentation for Logical Reasoning | findings-acl.353 | Poster | 2305.12599v6 |
https://aclanthology.org/2024.findings-acl.354.bib | @inproceedings{beau-crabbe-2024-codeinsight,
title = "{C}ode{I}nsight: A Curated Dataset of Practical Coding Solutions from {S}tack {O}verflow",
author = {Beau, Nathana{\"e}l and
Crabb{\'e}, Benoit},
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.354",
pages = "5935--5947",
abstract = "We introduce a novel dataset tailored for code generation, aimed at aiding developers in common tasks. Our dataset provides examples that include a clarified intent, code snippets associated, and an average of three related unit tests. It encompasses a range of libraries such as Pandas, Numpy, and Regex, along with more than 70 standard libraries in Python code derived from Stack Overflow. Comprising 3,402 crafted examples by Python experts, our dataset is designed for both model finetuning and standalone evaluation. To complete unit tests evaluation, we categorize examples in order to get more fine grained analysis, enhancing the understanding of models{'} strengths and weaknesses in specific coding tasks. The examples have been refined to reduce data contamination, a process confirmed by the performance of three leading models: Mistral 7B, CodeLLAMA 13B, and Starcoder 15B. We further investigate data-contamination testing GPT-4 performance on a part of our dataset. The benchmark can be accessed at anonymized address.",
}
| We introduce a novel dataset tailored for code generation, aimed at aiding developers in common tasks. Our dataset provides examples that include a clarified intent, code snippets associated, and an average of three related unit tests. It encompasses a range of libraries such as Pandas, Numpy, and Regex, along with more than 70 standard libraries in Python code derived from Stack Overflow. Comprising 3,402 crafted examples by Python experts, our dataset is designed for both model finetuning and standalone evaluation. To complete unit tests evaluation, we categorize examples in order to get more fine grained analysis, enhancing the understanding of models{'} strengths and weaknesses in specific coding tasks. The examples have been refined to reduce data contamination, a process confirmed by the performance of three leading models: Mistral 7B, CodeLLAMA 13B, and Starcoder 15B. We further investigate data-contamination testing GPT-4 performance on a part of our dataset. The benchmark can be accessed at anonymized address. | [
"Beau, Nathana{\\\"e}l",
"Crabb{\\'e}, Benoit"
]
| {C}ode{I}nsight: A Curated Dataset of Practical Coding Solutions from {S}tack {O}verflow | findings-acl.354 | Poster | 1806.07659v2 |
https://aclanthology.org/2024.findings-acl.355.bib | @inproceedings{thanh-nguyen-2024-vihatet5,
title = "{V}i{H}ate{T}5: Enhancing Hate Speech Detection in {V}ietnamese With a Unified Text-to-Text Transformer Model",
author = "Thanh Nguyen, Luan",
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.355",
pages = "5948--5961",
abstract = "Recent advancements in hate speech detection (HSD) in Vietnamese have made significant progress, primarily attributed to the emergence of transformer-based pre-trained language models, particularly those built on the BERT architecture. However, the necessity for specialized fine-tuned models has resulted in the complexity and fragmentation of developing a multitasking HSD system. Moreover, most current methodologies focus on fine-tuning general pre-trained models, primarily trained on formal textual datasets like Wikipedia, which may not accurately capture human behavior on online platforms. In this research, we introduce ViHateT5, a T5-based model pre-trained on our proposed large-scale domain-specific dataset named VOZ-HSD. By harnessing the power of a text-to-text architecture, ViHateT5 can tackle multiple tasks using a unified model and achieve state-of-the-art performance across all standard HSD benchmarks in Vietnamese. Our experiments also underscore the significance of label distribution in pre-training data on model efficacy. We provide our experimental materials for research purposes, including the VOZ-HSD dataset, pre-trained checkpoint, the unified HSD-multitask ViHateT5 model, and related source code on GitHub publicly.",
}
| Recent advancements in hate speech detection (HSD) in Vietnamese have made significant progress, primarily attributed to the emergence of transformer-based pre-trained language models, particularly those built on the BERT architecture. However, the necessity for specialized fine-tuned models has resulted in the complexity and fragmentation of developing a multitasking HSD system. Moreover, most current methodologies focus on fine-tuning general pre-trained models, primarily trained on formal textual datasets like Wikipedia, which may not accurately capture human behavior on online platforms. In this research, we introduce ViHateT5, a T5-based model pre-trained on our proposed large-scale domain-specific dataset named VOZ-HSD. By harnessing the power of a text-to-text architecture, ViHateT5 can tackle multiple tasks using a unified model and achieve state-of-the-art performance across all standard HSD benchmarks in Vietnamese. Our experiments also underscore the significance of label distribution in pre-training data on model efficacy. We provide our experimental materials for research purposes, including the VOZ-HSD dataset, pre-trained checkpoint, the unified HSD-multitask ViHateT5 model, and related source code on GitHub publicly. | [
"Thanh Nguyen, Luan"
]
| {V}i{H}ate{T}5: Enhancing Hate Speech Detection in {V}ietnamese With a Unified Text-to-Text Transformer Model | findings-acl.355 | Poster | 2310.11166v2 |
https://aclanthology.org/2024.findings-acl.356.bib | @inproceedings{steen-markert-2024-bias,
title = "Bias in News Summarization: Measures, Pitfalls and Corpora",
author = "Steen, Julius and
Markert, Katja",
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.356",
pages = "5962--5983",
abstract = "Summarization is an important application of large language models (LLMs). Most previous evaluation of summarization models has focused on their content selection, faithfulness, grammaticality and coherence. However, it is well known that LLMs can reproduce and reinforce harmful social biases. This raises the question: Do biases affect model outputs in a constrained setting like summarization?To help answer this question, we first motivate and introduce a number of definitions for biased behaviours in summarization models, along with practical operationalizations. Since we find that biases inherent to input documents can confound bias analysis in summaries, we propose a method to generate input documents with carefully controlled demographic attributes. This allows us to study summarizer behavior in a controlled setting, while still working with realistic input documents.We measure gender bias in English summaries generated by both purpose-built summarization models and general purpose chat models as a case study. We find content selection in single document summarization to be largely unaffected by gender bias, while hallucinations exhibit evidence of bias.To demonstrate the generality of our approach, we additionally investigate racial bias, including intersectional settings.",
}
| Summarization is an important application of large language models (LLMs). Most previous evaluation of summarization models has focused on their content selection, faithfulness, grammaticality and coherence. However, it is well known that LLMs can reproduce and reinforce harmful social biases. This raises the question: Do biases affect model outputs in a constrained setting like summarization?To help answer this question, we first motivate and introduce a number of definitions for biased behaviours in summarization models, along with practical operationalizations. Since we find that biases inherent to input documents can confound bias analysis in summaries, we propose a method to generate input documents with carefully controlled demographic attributes. This allows us to study summarizer behavior in a controlled setting, while still working with realistic input documents.We measure gender bias in English summaries generated by both purpose-built summarization models and general purpose chat models as a case study. We find content selection in single document summarization to be largely unaffected by gender bias, while hallucinations exhibit evidence of bias.To demonstrate the generality of our approach, we additionally investigate racial bias, including intersectional settings. | [
"Steen, Julius",
"Markert, Katja"
]
| Bias in News Summarization: Measures, Pitfalls and Corpora | findings-acl.356 | Poster | 2309.08047v3 |
https://aclanthology.org/2024.findings-acl.357.bib | @inproceedings{tao-etal-2024-trust,
title = "When to Trust {LLM}s: Aligning Confidence with Response Quality",
author = "Tao, Shuchang and
Yao, Liuyi and
Ding, Hanxing and
Xie, Yuexiang and
Cao, Qi and
Sun, Fei and
Gao, Jinyang and
Shen, Huawei and
Ding, Bolin",
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.357",
pages = "5984--5996",
abstract = "Despite the success of large language models (LLMs) in natural language generation, much evidence shows that LLMs may produce incorrect or nonsensical text. This limitation highlights the importance of discerning when to trust LLMs, especially in safety-critical domains. Existing methods often express reliability by confidence level, however, their effectiveness is limited by the lack of objective guidance. To address this, we propose CONfidence-Quality-ORDer-preserving alignment approach (CONQORD), which leverages reinforcement learning guided by a tailored dual-component reward function. This function integrates quality reward and order-preserving alignment reward functions. Specifically, the order-preserving reward incentivizes the model to verbalize greater confidence for responses of higher quality to align the order of confidence and quality. Experiments demonstrate that CONQORD significantly improves the alignment performance between confidence and response accuracy, without causing over-cautious. Furthermore, the aligned confidence provided by CONQORD informs when to trust LLMs, and acts as a determinant for initiating the retrieval process of external knowledge. Aligning confidence with response quality ensures more transparent and reliable responses, providing better trustworthiness.",
}
| Despite the success of large language models (LLMs) in natural language generation, much evidence shows that LLMs may produce incorrect or nonsensical text. This limitation highlights the importance of discerning when to trust LLMs, especially in safety-critical domains. Existing methods often express reliability by confidence level, however, their effectiveness is limited by the lack of objective guidance. To address this, we propose CONfidence-Quality-ORDer-preserving alignment approach (CONQORD), which leverages reinforcement learning guided by a tailored dual-component reward function. This function integrates quality reward and order-preserving alignment reward functions. Specifically, the order-preserving reward incentivizes the model to verbalize greater confidence for responses of higher quality to align the order of confidence and quality. Experiments demonstrate that CONQORD significantly improves the alignment performance between confidence and response accuracy, without causing over-cautious. Furthermore, the aligned confidence provided by CONQORD informs when to trust LLMs, and acts as a determinant for initiating the retrieval process of external knowledge. Aligning confidence with response quality ensures more transparent and reliable responses, providing better trustworthiness. | [
"Tao, Shuchang",
"Yao, Liuyi",
"Ding, Hanxing",
"Xie, Yuexiang",
"Cao, Qi",
"Sun, Fei",
"Gao, Jinyang",
"Shen, Huawei",
"Ding, Bolin"
]
| When to Trust {LLM}s: Aligning Confidence with Response Quality | findings-acl.357 | Poster | 2404.17287v2 |
https://aclanthology.org/2024.findings-acl.358.bib | @inproceedings{ai-huang-2024-zero,
title = "Zero-shot Cross-lingual Alignment for Embedding Initialization",
author = "Ai, Xi and
Huang, Zhiyong",
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.358",
pages = "5997--6007",
abstract = "For multilingual training, we present CrossInit, an initialization method that initializes embeddings into similar geometrical structures across languages in an unsupervised manner. CrossInit leverages a common cognitive linguistic mechanism, Zipf{'}s law, which indicates that similar concepts across languages have similar word ranks or frequencies in their monolingual corpora. Instead of considering point-to-point alignments based on ranks, CrossInit considers the same span of consecutive ranks in each language as the Positive pairs for alignment, while others out of the span are used as Negative pairs. CrossInit then employs Contrastive Learning to iteratively refine randomly initialized embeddings for similar geometrical structures across languages. Our experiments on Unsupervised NMT, XNLI, and MLQA showed significant gains in low-resource and dissimilar languages after applying CrossInit.",
}
| For multilingual training, we present CrossInit, an initialization method that initializes embeddings into similar geometrical structures across languages in an unsupervised manner. CrossInit leverages a common cognitive linguistic mechanism, Zipf{'}s law, which indicates that similar concepts across languages have similar word ranks or frequencies in their monolingual corpora. Instead of considering point-to-point alignments based on ranks, CrossInit considers the same span of consecutive ranks in each language as the Positive pairs for alignment, while others out of the span are used as Negative pairs. CrossInit then employs Contrastive Learning to iteratively refine randomly initialized embeddings for similar geometrical structures across languages. Our experiments on Unsupervised NMT, XNLI, and MLQA showed significant gains in low-resource and dissimilar languages after applying CrossInit. | [
"Ai, Xi",
"Huang, Zhiyong"
]
| Zero-shot Cross-lingual Alignment for Embedding Initialization | findings-acl.358 | Poster | 1712.01813v1 |
https://aclanthology.org/2024.findings-acl.359.bib | @inproceedings{manevich-tsarfaty-2024-mitigating,
title = "Mitigating Hallucinations in Large Vision-Language Models ({LVLM}s) via Language-Contrastive Decoding ({LCD})",
author = "Manevich, Avshalom 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.359",
pages = "6008--6022",
abstract = "Large Vision-Language Models (LVLMs) are an extension of Large Language Models (LLMs) that facilitate processing both image and text inputs, expanding AI capabilities. However, LVLMs struggle with object hallucinations due to their reliance on text cues and learned object co-occurrence biases. While most research quantifies these hallucinations, mitigation strategies are still lacking. Our study introduces a Language Contrastive Decoding (LCD) algorithm that adjusts LVLM outputs based on LLM distribution confidence levels, effectively reducing object hallucinations. We demonstrate the advantages of LCD in leading LVLMs, showing up to {\%}4 improvement in POPE F1 scores and up to {\%}36 reduction in CHAIR scores on the COCO validation set, while also improving captioning quality scores. Our method effectively improves LVLMs without needing complex post-processing or retraining, and is easily applicable to different models. Our findings highlight the potential of further exploration of LVLM-specific decoding algorithms.",
}
| Large Vision-Language Models (LVLMs) are an extension of Large Language Models (LLMs) that facilitate processing both image and text inputs, expanding AI capabilities. However, LVLMs struggle with object hallucinations due to their reliance on text cues and learned object co-occurrence biases. While most research quantifies these hallucinations, mitigation strategies are still lacking. Our study introduces a Language Contrastive Decoding (LCD) algorithm that adjusts LVLM outputs based on LLM distribution confidence levels, effectively reducing object hallucinations. We demonstrate the advantages of LCD in leading LVLMs, showing up to {\%}4 improvement in POPE F1 scores and up to {\%}36 reduction in CHAIR scores on the COCO validation set, while also improving captioning quality scores. Our method effectively improves LVLMs without needing complex post-processing or retraining, and is easily applicable to different models. Our findings highlight the potential of further exploration of LVLM-specific decoding algorithms. | [
"Manevich, Avshalom",
"Tsarfaty, Reut"
]
| Mitigating Hallucinations in Large Vision-Language Models ({LVLM}s) via Language-Contrastive Decoding ({LCD}) | findings-acl.359 | Poster | 2408.04664v1 |
https://aclanthology.org/2024.findings-acl.360.bib | @inproceedings{dinu-etal-2024-takes,
title = "It takes two to borrow: a donor and a recipient. Who{'}s who?",
author = "Dinu, Liviu and
Uban, Ana and
Dinu, Anca and
Iordache, Ioan-Bogdan and
Georgescu, Simona and
Zoicas, Laurentiu",
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.360",
pages = "6023--6035",
abstract = "We address the open problem of automatically identifying the direction of lexical borrowing, given word pairs in the donor and recipient languages. We propose strong benchmarks for this task, by applying a set of machine learning models. We extract and publicly release a comprehensive borrowings dataset from the recent RoBoCoP cognates and borrowings database for five Romance languages. We experiment on this dataset with both graphic and phonetic representations and with different features, models and architectures. We interpret the results, in terms of F1 score, commenting on the influence of features and model choice, of the imbalanced data and of the inherent difficulty of the task for particular language pairs. We show that automatically determining the direction of borrowing is a feasible task, and propose additional directions for future work.",
}
| We address the open problem of automatically identifying the direction of lexical borrowing, given word pairs in the donor and recipient languages. We propose strong benchmarks for this task, by applying a set of machine learning models. We extract and publicly release a comprehensive borrowings dataset from the recent RoBoCoP cognates and borrowings database for five Romance languages. We experiment on this dataset with both graphic and phonetic representations and with different features, models and architectures. We interpret the results, in terms of F1 score, commenting on the influence of features and model choice, of the imbalanced data and of the inherent difficulty of the task for particular language pairs. We show that automatically determining the direction of borrowing is a feasible task, and propose additional directions for future work. | [
"Dinu, Liviu",
"Uban, Ana",
"Dinu, Anca",
"Iordache, Ioan-Bogdan",
"Georgescu, Simona",
"Zoicas, Laurentiu"
]
| It takes two to borrow: a donor and a recipient. Who{'}s who? | findings-acl.360 | Poster | 2101.11769v2 |
https://aclanthology.org/2024.findings-acl.361.bib | @inproceedings{guan-greene-2024-advancing,
title = "Advancing Post-{OCR} Correction: A Comparative Study of Synthetic Data",
author = "Guan, Shuhao and
Greene, Derek",
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.361",
pages = "6036--6047",
abstract = "This paper explores the application of synthetic data in the post-OCR domain on multiple fronts by conducting experiments to assess the impact of data volume, augmentation, and synthetic data generation methods on model performance. Furthermore, we introduce a novel algorithm that leverages computer vision feature detection algorithms to calculate glyph similarity for constructing post-OCR synthetic data. Through experiments conducted across a variety of languages, including several low-resource ones, we demonstrate that models like ByT5 can significantly reduce Character Error Rates (CER) without the need for manually annotated data, and our proposed synthetic data generation method shows advantages over traditional methods, particularly in low-resource languages.",
}
| This paper explores the application of synthetic data in the post-OCR domain on multiple fronts by conducting experiments to assess the impact of data volume, augmentation, and synthetic data generation methods on model performance. Furthermore, we introduce a novel algorithm that leverages computer vision feature detection algorithms to calculate glyph similarity for constructing post-OCR synthetic data. Through experiments conducted across a variety of languages, including several low-resource ones, we demonstrate that models like ByT5 can significantly reduce Character Error Rates (CER) without the need for manually annotated data, and our proposed synthetic data generation method shows advantages over traditional methods, particularly in low-resource languages. | [
"Guan, Shuhao",
"Greene, Derek"
]
| Advancing Post-{OCR} Correction: A Comparative Study of Synthetic Data | findings-acl.361 | Poster | 2408.02253v2 |
https://aclanthology.org/2024.findings-acl.362.bib | @inproceedings{huang-etal-2024-geoagent,
title = "{G}eo{A}gent: To Empower {LLM}s using Geospatial Tools for Address Standardization",
author = "Huang, Chenghua and
Chen, Shisong and
Li, Zhixu and
Qu, Jianfeng and
Xiao, Yanghua and
Liu, Jiaxin and
Chen, Zhigang",
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.362",
pages = "6048--6063",
abstract = "This paper presents a novel solution to tackle the challenges that posed by the abundance of non-standard addresses, which input by users in modern applications such as navigation maps, ride-hailing apps, food delivery platforms, and logistics services. These manually entered addresses often contain irregularities, such as missing information, spelling errors, colloquial descriptions, and directional offsets, which hinder address-related tasks like address matching and linking. To tackle these challenges, we propose GeoAgent, a new framework comprising two main components: a large language model (LLM) and a suite of geographical tools. By harnessing the semantic understanding capabilities of the LLM and integrating specific geospatial tools, GeoAgent incorporates spatial knowledge into address texts and achieves efficient address standardization. Further, to verify the effectiveness and practicality of our approach, we construct a comprehensive dataset of complex non-standard addresses, which fills the gaps in existing datasets and proves invaluable for training and evaluating the performance of address standardization models in this community. Experimental results demonstrate the efficacy of GeoAgent, showcasing substantial improvements in the performance of address-related models across various downstream tasks.",
}
| This paper presents a novel solution to tackle the challenges that posed by the abundance of non-standard addresses, which input by users in modern applications such as navigation maps, ride-hailing apps, food delivery platforms, and logistics services. These manually entered addresses often contain irregularities, such as missing information, spelling errors, colloquial descriptions, and directional offsets, which hinder address-related tasks like address matching and linking. To tackle these challenges, we propose GeoAgent, a new framework comprising two main components: a large language model (LLM) and a suite of geographical tools. By harnessing the semantic understanding capabilities of the LLM and integrating specific geospatial tools, GeoAgent incorporates spatial knowledge into address texts and achieves efficient address standardization. Further, to verify the effectiveness and practicality of our approach, we construct a comprehensive dataset of complex non-standard addresses, which fills the gaps in existing datasets and proves invaluable for training and evaluating the performance of address standardization models in this community. Experimental results demonstrate the efficacy of GeoAgent, showcasing substantial improvements in the performance of address-related models across various downstream tasks. | [
"Huang, Chenghua",
"Chen, Shisong",
"Li, Zhixu",
"Qu, Jianfeng",
"Xiao, Yanghua",
"Liu, Jiaxin",
"Chen, Zhigang"
]
| {G}eo{A}gent: To Empower {LLM}s using Geospatial Tools for Address Standardization | findings-acl.362 | Poster | 2405.07664v1 |
https://aclanthology.org/2024.findings-acl.363.bib | @inproceedings{maarouf-etal-2024-hqp,
title = "{HQP}: A Human-Annotated Dataset for Detecting Online Propaganda",
author = {Maarouf, Abdurahman and
B{\"a}r, Dominik and
Geissler, Dominique and
Feuerriegel, Stefan},
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.363",
pages = "6064--6089",
abstract = "Online propaganda poses a severe threat to the integrity of societies. However, existing datasets for detecting online propaganda have a key limitation: they were annotated using weak labels that can be noisy and even incorrect. To address this limitation, our work makes the following contributions: (1) We present HQP: a novel dataset (N=30000) for detecting online propaganda with high-quality labels. To the best of our knowledge, HQP is the first large-scale dataset for detecting online propaganda that was created through human annotation. (2) We show empirically that state-of-the-art language models fail in detecting online propaganda when trained with weak labels (AUC: 64.03). In contrast, state-of-the-art language models can accurately detect online propaganda when trained with our high-quality labels (AUC: 92.25), which is an improvement of 44{\%}. (3) We show that prompt-based learning using a small sample of high-quality labels can still achieve a reasonable performance (AUC: 80.27) while significantly reducing the cost of labeling. (4) We extend HQP to HQP+ to test how well propaganda across different contexts can be detected. Crucially, our work highlights the importance of high-quality labels for sensitive NLP tasks such as propaganda detection.",
}
| Online propaganda poses a severe threat to the integrity of societies. However, existing datasets for detecting online propaganda have a key limitation: they were annotated using weak labels that can be noisy and even incorrect. To address this limitation, our work makes the following contributions: (1) We present HQP: a novel dataset (N=30000) for detecting online propaganda with high-quality labels. To the best of our knowledge, HQP is the first large-scale dataset for detecting online propaganda that was created through human annotation. (2) We show empirically that state-of-the-art language models fail in detecting online propaganda when trained with weak labels (AUC: 64.03). In contrast, state-of-the-art language models can accurately detect online propaganda when trained with our high-quality labels (AUC: 92.25), which is an improvement of 44{\%}. (3) We show that prompt-based learning using a small sample of high-quality labels can still achieve a reasonable performance (AUC: 80.27) while significantly reducing the cost of labeling. (4) We extend HQP to HQP+ to test how well propaganda across different contexts can be detected. Crucially, our work highlights the importance of high-quality labels for sensitive NLP tasks such as propaganda detection. | [
"Maarouf, Abdurahman",
"B{\\\"a}r, Dominik",
"Geissler, Dominique",
"Feuerriegel, Stefan"
]
| {HQP}: A Human-Annotated Dataset for Detecting Online Propaganda | findings-acl.363 | Poster | 2304.14931v2 |
https://aclanthology.org/2024.findings-acl.364.bib | @inproceedings{hu-etal-2024-teaching,
title = "Teaching Language Models to Self-Improve by Learning from Language Feedback",
author = "Hu, Chi and
Hu, Yimin and
Cao, Hang 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.364",
pages = "6090--6101",
abstract = "Aligning Large Language Models (LLMs) with human intentions and values is crucial yet challenging. Current methods primarily rely on human preferences, which are costly and insufficient in capturing nuanced feedback expressed in natural language. In this paper, we present Self-Refinement Tuning (SRT), a method that leverages model feedback for alignment, thereby reducing reliance on human annotations. SRT uses a base language model (e.g., Tulu2) to generate initial responses, which are critiqued and refined by a more advanced model (e.g., GPT-4-Turbo). This process enables the base model to self-evaluate and improve its outputs, facilitating continuous learning. SRT further optimizes the model by learning from its self-generated feedback and refinements, creating a feedback loop that promotes model improvement. Our empirical evaluations demonstrate that SRT significantly outperforms strong baselines across diverse tasks and model sizes. When applied to a 70B parameter model, SRT increases the win rate from 9.6{\%} to 25.8{\%} on the AlpacaEval 2.0 benchmark, surpassing well-established systems such as GPT-4-0314, Claude 2, and Gemini. Our analysis highlights the crucial role of language feedback in the success of SRT, suggesting potential for further exploration in this direction.",
}
| Aligning Large Language Models (LLMs) with human intentions and values is crucial yet challenging. Current methods primarily rely on human preferences, which are costly and insufficient in capturing nuanced feedback expressed in natural language. In this paper, we present Self-Refinement Tuning (SRT), a method that leverages model feedback for alignment, thereby reducing reliance on human annotations. SRT uses a base language model (e.g., Tulu2) to generate initial responses, which are critiqued and refined by a more advanced model (e.g., GPT-4-Turbo). This process enables the base model to self-evaluate and improve its outputs, facilitating continuous learning. SRT further optimizes the model by learning from its self-generated feedback and refinements, creating a feedback loop that promotes model improvement. Our empirical evaluations demonstrate that SRT significantly outperforms strong baselines across diverse tasks and model sizes. When applied to a 70B parameter model, SRT increases the win rate from 9.6{\%} to 25.8{\%} on the AlpacaEval 2.0 benchmark, surpassing well-established systems such as GPT-4-0314, Claude 2, and Gemini. Our analysis highlights the crucial role of language feedback in the success of SRT, suggesting potential for further exploration in this direction. | [
"Hu, Chi",
"Hu, Yimin",
"Cao, Hang",
"Xiao, Tong",
"Zhu, JingBo"
]
| Teaching Language Models to Self-Improve by Learning from Language Feedback | findings-acl.364 | Poster | 2404.10490v2 |
https://aclanthology.org/2024.findings-acl.365.bib | @inproceedings{wicke-wachowiak-2024-exploring,
title = "Exploring Spatial Schema Intuitions in Large Language and Vision Models",
author = "Wicke, Philipp and
Wachowiak, Lennart",
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.365",
pages = "6102--6117",
abstract = "Despite the ubiquity of large language models (LLMs) in AI research, the question of embodiment in LLMs remains underexplored, distinguishing them from embodied systems in robotics where sensory perception directly informs physical action.Our investigation navigates the intriguing terrain of whether LLMs, despite their non-embodied nature, effectively capture implicit human intuitions about fundamental, spatial building blocks of language. We employ insights from spatial cognitive foundations developed through early sensorimotor experiences, guiding our exploration through the reproduction of three psycholinguistic experiments. Surprisingly, correlations between model outputs and human responses emerge, revealing adaptability without a tangible connection to embodied experiences. Notable distinctions include polarized language model responses and reduced correlations in vision language models. This research contributes to a nuanced understanding of the interplay between language, spatial experiences, and the computations made by large language models.Project Website: https://cisnlp.github.io/Spatial{\_}Schemas/",
}
| Despite the ubiquity of large language models (LLMs) in AI research, the question of embodiment in LLMs remains underexplored, distinguishing them from embodied systems in robotics where sensory perception directly informs physical action.Our investigation navigates the intriguing terrain of whether LLMs, despite their non-embodied nature, effectively capture implicit human intuitions about fundamental, spatial building blocks of language. We employ insights from spatial cognitive foundations developed through early sensorimotor experiences, guiding our exploration through the reproduction of three psycholinguistic experiments. Surprisingly, correlations between model outputs and human responses emerge, revealing adaptability without a tangible connection to embodied experiences. Notable distinctions include polarized language model responses and reduced correlations in vision language models. This research contributes to a nuanced understanding of the interplay between language, spatial experiences, and the computations made by large language models.Project Website: https://cisnlp.github.io/Spatial{\_}Schemas/ | [
"Wicke, Philipp",
"Wachowiak, Lennart"
]
| Exploring Spatial Schema Intuitions in Large Language and Vision Models | findings-acl.365 | Poster | 2402.00956v2 |
https://aclanthology.org/2024.findings-acl.366.bib | @inproceedings{miao-etal-2024-efficient,
title = "Efficient Detection of {LLM}-generated Texts with a {B}ayesian Surrogate Model",
author = "Miao, Yibo and
Gao, Hongcheng and
Zhang, Hao and
Deng, Zhijie",
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.366",
pages = "6118--6130",
abstract = "The detection of machine-generated text, especially from large language models (LLMs), is crucial in preventing serious social problems resulting from their misuse. Some methods train dedicated detectors on specific datasets but fall short in generalizing to unseen test data, while other zero-shot ones often yield suboptimal performance. Although the recent DetectGPT has shown promising detection performance, it suffers from significant inefficiency issues, as detecting a single candidate requires querying the source LLM with hundreds of its perturbations. This paper aims to bridge this gap. Concretely, we propose to incorporate a Bayesian surrogate model, which allows us to select typical samples based on Bayesian uncertainty and interpolate scores from typical samples to other samples, to improve query efficiency. Empirical results demonstrate that our method significantly outperforms existing approaches under a low query budget. Notably, when detecting the text generated by LLaMA family models, our method with just 2 or 3 queries can outperform DetectGPT with 200 queries.",
}
| The detection of machine-generated text, especially from large language models (LLMs), is crucial in preventing serious social problems resulting from their misuse. Some methods train dedicated detectors on specific datasets but fall short in generalizing to unseen test data, while other zero-shot ones often yield suboptimal performance. Although the recent DetectGPT has shown promising detection performance, it suffers from significant inefficiency issues, as detecting a single candidate requires querying the source LLM with hundreds of its perturbations. This paper aims to bridge this gap. Concretely, we propose to incorporate a Bayesian surrogate model, which allows us to select typical samples based on Bayesian uncertainty and interpolate scores from typical samples to other samples, to improve query efficiency. Empirical results demonstrate that our method significantly outperforms existing approaches under a low query budget. Notably, when detecting the text generated by LLaMA family models, our method with just 2 or 3 queries can outperform DetectGPT with 200 queries. | [
"Miao, Yibo",
"Gao, Hongcheng",
"Zhang, Hao",
"Deng, Zhijie"
]
| Efficient Detection of {LLM}-generated Texts with a {B}ayesian Surrogate Model | findings-acl.366 | Poster | 2305.16617v3 |
https://aclanthology.org/2024.findings-acl.367.bib | @inproceedings{bouzoubaa-etal-2024-decoding,
title = "Decoding the Narratives: Analyzing Personal Drug Experiences Shared on {R}eddit",
author = "Bouzoubaa, Layla and
Aghakhani, Elham and
Song, Max and
Trinh, Quang and
Rezapour, Shadi",
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.367",
pages = "6131--6148",
abstract = "Online communities such as drug-related subreddits serve as safe spaces for people who use drugs (PWUD), fostering discussions on substance use experiences, harm reduction, and addiction recovery. Users{'} shared narratives on these forums provide insights into the likelihood of developing a substance use disorder (SUD) and recovery potential. Our study aims to develop a multi-level, multi-label classification model to analyze online user-generated texts about substance use experiences. For this purpose, we first introduce a novel taxonomy to assess the nature of posts, including their intended connections (Inquisition or Disclosure), subjects (e.g., Recovery, Dependency), and specific objectives (e.g., Relapse, Quality, Safety). Using various multi-label classification algorithms on a set of annotated data, we show that GPT-4, when prompted with instructions, definitions, and examples, outperformed all other models. We apply this model to label an additional 1,000 posts and analyze the categories of linguistic expression used within posts in each class. Our analysis shows that topics such as Safety, Combination of Substances, and Mental Health see more disclosure, while discussions about physiological Effects focus on harm reduction. Our work enriches the understanding of PWUD{'}s experiences and informs the broader knowledge base on SUD and drug use.",
}
| Online communities such as drug-related subreddits serve as safe spaces for people who use drugs (PWUD), fostering discussions on substance use experiences, harm reduction, and addiction recovery. Users{'} shared narratives on these forums provide insights into the likelihood of developing a substance use disorder (SUD) and recovery potential. Our study aims to develop a multi-level, multi-label classification model to analyze online user-generated texts about substance use experiences. For this purpose, we first introduce a novel taxonomy to assess the nature of posts, including their intended connections (Inquisition or Disclosure), subjects (e.g., Recovery, Dependency), and specific objectives (e.g., Relapse, Quality, Safety). Using various multi-label classification algorithms on a set of annotated data, we show that GPT-4, when prompted with instructions, definitions, and examples, outperformed all other models. We apply this model to label an additional 1,000 posts and analyze the categories of linguistic expression used within posts in each class. Our analysis shows that topics such as Safety, Combination of Substances, and Mental Health see more disclosure, while discussions about physiological Effects focus on harm reduction. Our work enriches the understanding of PWUD{'}s experiences and informs the broader knowledge base on SUD and drug use. | [
"Bouzoubaa, Layla",
"Aghakhani, Elham",
"Song, Max",
"Trinh, Quang",
"Rezapour, Shadi"
]
| Decoding the Narratives: Analyzing Personal Drug Experiences Shared on {R}eddit | findings-acl.367 | Poster | 2406.12117v1 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.