Authors: Xuefei Ning, Zifu Wang, Shiyao Li, Zinan Lin, Peiran Yao, Tianyu Fu, Matthew B. Blaschko, Guohao Dai, Huazhong Yang, Yu Wang
Abstract: Teaching to improve student models (e.g., knowledge distillation) is an extensively studied methodology in LLMs. However, for humans, teaching not only improves students but also improves teachers. We ask: Can LLMs also learn by teaching (LbT)? If yes, we can potentially unlock the possibility of continuously advancing the models without solely relying on human-produced data or stronger models. In this paper, we provide a preliminary exploration of this ambitious agenda. We show that LbT ideas can be incorporated into existing LLM training/prompting pipelines and provide noticeable improvements. Specifically, we design three methods, each mimicking one of the three levels of LbT in humans: observing students' feedback, learning from the feedback, and learning iteratively, with the goals of improving answer accuracy without training and improving models' inherent capability with fine-tuning. The findings are encouraging. For example, similar to LbT in human, we see that: (1) LbT can induce weak-to-strong generalization: strong models can improve themselves by teaching other weak models; (2) Diversity in students might help: teaching multiple students could be better than teaching one student or the teacher itself. We hope that this early promise can inspire future research on LbT and more broadly adopting the advanced techniques in education to improve LLMs. The code is available at https://github.com/imagination-research/lbt.
Authors: Chunyuan Deng, Yilun Zhao, Yuzhao Heng, Yitong Li, Jiannan Cao, Xiangru Tang, Arman Cohan
Abstract: Data contamination has garnered increased attention in the era of large language models (LLMs) due to the reliance on extensive internet-derived training corpora. The issue of training corpus overlap with evaluation benchmarks--referred to as contamination--has been the focus of significant recent research. This body of work aims to identify contamination, understand its impacts, and explore mitigation strategies from diverse perspectives. However, comprehensive studies that provide a clear pathway from foundational concepts to advanced insights are lacking in this nascent field. Therefore, we present a comprehensive survey in the field of data contamination, laying out the key issues, methodologies, and findings to date, and highlighting areas in need of further research and development. In particular, we begin by examining the effects of data contamination across various stages and forms. We then provide a detailed analysis of current contamination detection methods, categorizing them to highlight their focus, assumptions, strengths, and limitations. We also discuss mitigation strategies, offering a clear guide for future research. This survey serves as a succinct overview of the most recent advancements in data contamination research, providing a straightforward guide for the benefit of future research endeavors.
Authors: Kawshik Manikantan (CVIT, IIIT Hyderabad), Shubham Toshniwal (NVIDIA), Makarand Tapaswi (CVIT, IIIT Hyderabad), Vineet Gandhi (CVIT, IIIT Hyderabad)
Abstract: The limited generalization of coreference resolution (CR) models has been a major bottleneck in the task's broad application. Prior work has identified annotation differences, especially for mention detection, as one of the main reasons for the generalization gap and proposed using additional annotated target domain data. Rather than relying on this additional annotation, we propose an alternative formulation of the CR task, Major Entity Identification (MEI), where we: (a) assume the target entities to be specified in the input, and (b) limit the task to only the frequent entities. Through extensive experiments, we demonstrate that MEI models generalize well across domains on multiple datasets with supervised models and LLM-based few-shot prompting. Additionally, the MEI task fits the classification framework, which enables the use of classification-based metrics that are more robust than the current CR metrics. Finally, MEI is also of practical use as it allows a user to search for all mentions of a particular entity or a group of entities of interest.
Authors: Allen Roush, Yusuf Shabazz, Arvind Balaji, Peter Zhang, Stefano Mezza, Markus Zhang, Sanjay Basu, Sriram Vishwanath, Mehdi Fatemi, Ravid Schwartz-Ziv
Abstract: We introduce OpenDebateEvidence, a comprehensive dataset for argument mining and summarization sourced from the American Competitive Debate community. This dataset includes over 3.5 million documents with rich metadata, making it one of the most extensive collections of debate evidence. OpenDebateEvidence captures the complexity of arguments in high school and college debates, providing valuable resources for training and evaluation. Our extensive experiments demonstrate the efficacy of fine-tuning state-of-the-art large language models for argumentative abstractive summarization across various methods, models, and datasets. By providing this comprehensive resource, we aim to advance computational argumentation and support practical applications for debaters, educators, and researchers. OpenDebateEvidence is publicly available to support further research and innovation in computational argumentation. Access it here: https://huggingface.co/datasets/Yusuf5/OpenCaselist
Authors: Mobashir Sadat, Cornelia Caragea
Abstract: Scientific Natural Language Inference (NLI) is the task of predicting the semantic relation between a pair of sentences extracted from research articles. The automatic annotation method based on distant supervision for the training set of SciNLI (Sadat and Caragea, 2022b), the first and most popular dataset for this task, results in label noise which inevitably degenerates the performance of classifiers. In this paper, we propose a novel co-training method that assigns weights based on the training dynamics of the classifiers to the distantly supervised labels, reflective of the manner they are used in the subsequent training epochs. That is, unlike the existing semi-supervised learning (SSL) approaches, we consider the historical behavior of the classifiers to evaluate the quality of the automatically annotated labels. Furthermore, by assigning importance weights instead of filtering out examples based on an arbitrary threshold on the predicted confidence, we maximize the usage of automatically labeled data, while ensuring that the noisy labels have a minimal impact on model training. The proposed method obtains an improvement of 1.5% in Macro F1 over the distant supervision baseline, and substantial improvements over several other strong SSL baselines. We make our code and data available on Github.
Authors: Atula Tejaswi, Nilesh Gupta, Eunsol Choi
Abstract: Despite rapid progress in large language models (LLMs), their performance on a vast majority of languages remain unsatisfactory. In this paper, we study building language-specific LLMs by adapting monolingual and multilingual LLMs. We conduct systematic experiments on how design choices (base model selection, vocabulary extension, and continued fine-tuning) impact the adapted LLM, both in terms of efficiency (how many tokens are needed to encode the same amount of information) and end task performance. We find that (1) the initial performance before the adaptation is not always indicative of the final performance. (2) Efficiency can easily improved with simple vocabulary extension and continued fine-tuning in most LLMs we study, and (3) The optimal adaptation method is highly language-dependent, and the simplest approach works well across various experimental settings. Adapting English-centric models can yield better results than adapting multilingual models despite their worse initial performance on low-resource languages. Together, our work lays foundations on efficiently building language-specific LLMs by adapting existing LLMs.
Authors: Taiming Lu, Muhan Gao, Kuai Yu, Adam Byerly, Daniel Khashabi
Abstract: Large Language Models (LLMs) exhibit positional bias, struggling to utilize information from the middle or end of long contexts. Our study explores LLMs' long-context reasoning by probing their hidden representations. We find that while LLMs encode the position of target information, they often fail to leverage this in generating accurate responses. This reveals a disconnect between information retrieval and utilization, a "know but don't tell" phenomenon. We further analyze the relationship between extraction time and final accuracy, offering insights into the underlying mechanics of transformer models.
Authors: Pamela D. Rivi\`ere (Department of Cognitive Science UC San Diego), Anne L. Beatty-Mart\'inez (Department of Cognitive Science UC San Diego), Sean Trott (Department of Cognitive Science UC San Diego, Computational Social Science UC San Diego)
Abstract: Lexical ambiguity -- where a single wordform takes on distinct, context-dependent meanings -- serves as a useful tool to compare across different large language models' (LLMs') ability to form distinct, contextualized representations of the same stimulus. Few studies have systematically compared LLMs' contextualized word embeddings for languages beyond English. Here, we evaluate multiple bidirectional transformers' (BERTs') semantic representations of Spanish ambiguous nouns in context. We develop a novel dataset of minimal-pair sentences evoking the same or different sense for a target ambiguous noun. In a pre-registered study, we collect contextualized human relatedness judgments for each sentence pair. We find that various BERT-based LLMs' contextualized semantic representations capture some variance in human judgments but fall short of the human benchmark, and for Spanish -- unlike English -- model scale is uncorrelated with performance. We also identify stereotyped trajectories of target noun disambiguation as a proportion of traversal through a given LLM family's architecture, which we partially replicate in English. We contribute (1) a dataset of controlled, Spanish sentence stimuli with human relatedness norms, and (2) to our evolving understanding of the impact that LLM specification (architectures, training protocols) exerts on contextualized embeddings.
Authors: Taraka Rama, Sowmya Vajjala
Abstract: The Universal Dependencies (UD) project aims to create a cross-linguistically consistent dependency annotation for multiple languages, to facilitate multilingual NLP. It currently supports 114 languages. Dravidian languages are spoken by over 200 million people across the word, and yet there are only two languages from this family in UD. This paper examines some of the morphological and syntactic features of Dravidian languages and explores how they can be annotated in the UD framework.
Authors: Alkis Koudounas, Flavio Giobergia, Eliana Pastor, Elena Baralis
Abstract: Speech models may be affected by performance imbalance in different population subgroups, raising concerns about fair treatment across these groups. Prior attempts to mitigate unfairness either focus on user-defined subgroups, potentially overlooking other affected subgroups, or do not explicitly improve the internal representation at the subgroup level. This paper proposes the first adoption of contrastive learning to mitigate speech model bias in underperforming subgroups. We employ a three-level learning technique that guides the model in focusing on different scopes for the contrastive loss, i.e., task, subgroup, and the errors within subgroups. The experiments on two spoken language understanding datasets and two languages demonstrate that our approach improves internal subgroup representations, thus reducing model bias and enhancing performance.
Authors: Parker Seegmiller, Joseph Gatto, Sarah Masud Preum
Abstract: Recent evaluations of cross-domain text classification models aim to measure the ability of a model to obtain domain-invariant performance in a target domain given labeled samples in a source domain. The primary strategy for this evaluation relies on assumed differences between source domain samples and target domain samples in benchmark datasets. This evaluation strategy fails to account for the similarity between source and target domains, and may mask when models fail to transfer learning to specific target samples which are highly dissimilar from the source domain. We introduce Depth $F_1$, a novel cross-domain text classification performance metric. Designed to be complementary to existing classification metrics such as $F_1$, Depth $F_1$ measures how well a model performs on target samples which are dissimilar from the source domain. We motivate this metric using standard cross-domain text classification datasets and benchmark several recent cross-domain text classification models, with the goal of enabling in-depth evaluation of the semantic generalizability of cross-domain text classification models.
Authors: Seungbeen Lee, Seungwon Lim, Seungju Han, Giyeong Oh, Hyungjoo Chae, Jiwan Chung, Minju Kim, Beong-woo Kwak, Yeonsoo Lee, Dongha Lee, Jinyoung Yeo, Youngjae Yu
Abstract: The idea of personality in descriptive psychology, traditionally defined through observable behavior, has now been extended to Large Language Models (LLMs) to better understand their behavior. This raises a question: do LLMs exhibit distinct and consistent personality traits, similar to humans? Existing self-assessment personality tests, while applicable, lack the necessary validity and reliability for precise personality measurements. To address this, we introduce TRAIT, a new tool consisting of 8K multi-choice questions designed to assess the personality of LLMs with validity and reliability. TRAIT is built on the psychometrically validated human questionnaire, Big Five Inventory (BFI) and Short Dark Triad (SD-3), enhanced with the ATOMIC10X knowledge graph for testing personality in a variety of real scenarios. TRAIT overcomes the reliability and validity issues when measuring personality of LLM with self-assessment, showing the highest scores across three metrics: refusal rate, prompt sensitivity, and option order sensitivity. It reveals notable insights into personality of LLM: 1) LLMs exhibit distinct and consistent personality, which is highly influenced by their training data (i.e., data used for alignment tuning), and 2) current prompting techniques have limited effectiveness in eliciting certain traits, such as high psychopathy or low conscientiousness, suggesting the need for further research in this direction.
Authors: Rongxin Zhu, Jey Han Lau, Jianzhong Qi
Abstract: Factual consistency is an important quality in dialogue summarization. Large language model (LLM)-based automatic text summarization models generate more factually consistent summaries compared to those by smaller pretrained language models, but they face deployment challenges in real-world applications due to privacy or resource constraints. In this paper, we investigate the use of symbolic knowledge distillation to improve the factual consistency of smaller pretrained models for dialogue summarization. We employ zero-shot learning to extract symbolic knowledge from LLMs, generating both factually consistent (positive) and inconsistent (negative) summaries. We then apply two contrastive learning objectives on these summaries to enhance smaller summarization models. Experiments with BART, PEGASUS, and Flan-T5 indicate that our approach surpasses strong baselines that rely on complex data augmentation strategies. Our approach achieves better factual consistency while maintaining coherence, fluency, and relevance, as confirmed by various automatic evaluation metrics. We also provide access to the data and code to facilitate future research.
Authors: Alfonso Amayuelas, Xianjun Yang, Antonis Antoniades, Wenyue Hua, Liangming Pan, William Wang
Abstract: Large Language Models (LLMs) have shown exceptional results on current benchmarks when working individually. The advancement in their capabilities, along with a reduction in parameter size and inference times, has facilitated the use of these models as agents, enabling interactions among multiple models to execute complex tasks. Such collaborations offer several advantages, including the use of specialized models (e.g. coding), improved confidence through multiple computations, and enhanced divergent thinking, leading to more diverse outputs. Thus, the collaborative use of language models is expected to grow significantly in the coming years. In this work, we evaluate the behavior of a network of models collaborating through debate under the influence of an adversary. We introduce pertinent metrics to assess the adversary's effectiveness, focusing on system accuracy and model agreement. Our findings highlight the importance of a model's persuasive ability in influencing others. Additionally, we explore inference-time methods to generate more compelling arguments and evaluate the potential of prompt-based mitigation as a defensive strategy.
Authors: Yue Huang, Chenrui Fan, Yuan Li, Siyuan Wu, Tianyi Zhou, Xiangliang Zhang, Lichao Sun
Abstract: Large Language Models (LLMs) have garnered significant attention due to their remarkable ability to process information across various languages. Despite their capabilities, they exhibit inconsistencies in handling identical queries in different languages, presenting challenges for further advancement. This paper introduces a method to enhance the multilingual performance of LLMs by aggregating knowledge from diverse languages. This approach incorporates a low-resource knowledge detector specific to a language, a language selection process, and mechanisms for answer replacement and integration. Our experiments demonstrate notable performance improvements, particularly in reducing language performance disparity. An ablation study confirms that each component of our method significantly contributes to these enhancements. This research highlights the inherent potential of LLMs to harmonize multilingual capabilities and offers valuable insights for further exploration.
Authors: Jayetri Bardhan, Bushi Xiao, Daisy Zhe Wang
Abstract: Question answering (QA) over tables and text has gained much popularity over the years. Multi-hop table-text QA requires multiple hops between the table and text, making it a challenging QA task. Although several works have attempted to solve the table-text QA task, most involve training the models and requiring labeled data. In this paper, we have proposed a model - TTQA-RS: A break-down prompting approach for Multi-hop Table-Text Question Answering with Reasoning and Summarization. Our model uses augmented knowledge including table-text summary with decomposed sub-question with answer for a reasoning-based table-text QA. Using open-source language models our model outperformed all existing prompting methods for table-text QA tasks on existing table-text QA datasets like HybridQA and OTT-QA's development set. Our results are comparable with the training-based state-of-the-art models, demonstrating the potential of prompt-based approaches using open-source LLMs. Additionally, by using GPT-4 with LLaMA3-70B, our model achieved state-of-the-art performance for prompting-based methods on multi-hop table-text QA.
Authors: Zhiqiang Pi, Annapurna Vadaparty, Benjamin K. Bergen, Cameron R. Jones
Abstract: Recent empirical results have sparked a debate about whether or not Large Language Models (LLMs) are capable of Theory of Mind (ToM). While some have found LLMs to be successful on ToM evaluations such as the False Belief task (Kosinski, 2023), others have argued that LLMs solve these tasks by exploiting spurious correlations -- not representing beliefs -- since they fail on trivial alterations to these tasks (Ullman, 2023). In this paper, we introduce SCALPEL: a technique to generate targeted modifications for False Belief tasks to test different specific hypotheses about why LLMs fail. We find that modifications which make explicit common inferences -- such as that looking at a transparent object implies recognizing its contents -- preserve LLMs' performance. This suggests that LLMs' failures on modified ToM tasks could result from a lack of more general commonsense reasoning, rather than a failure to represent mental states. We argue that SCALPEL could be helpful for explaining LLM successes and failures in other cases.
Authors: Yunmo Chen, Tongfei Chen, Harsh Jhamtani, Patrick Xia, Richard Shin, Jason Eisner, Benjamin Van Durme
Abstract: We introduce iterative retrieval, a novel framework that empowers retrievers to make iterative decisions through policy optimization. Finding an optimal portfolio of retrieved items is a combinatorial optimization problem, generally considered NP-hard. This approach provides a learned approximation to such a solution, meeting specific task requirements under a given family of large language models (LLMs). We propose a training procedure based on reinforcement learning, incorporating feedback from LLMs. We instantiate an iterative retriever for composing in-context learning (ICL) exemplars and apply it to various semantic parsing tasks that demand synthesized programs as outputs. By adding only 4M additional parameters for state encoding, we convert an off-the-shelf dense retriever into a stateful iterative retriever, outperforming previous methods in selecting ICL exemplars on semantic parsing datasets such as CalFlow, TreeDST, and MTOP. Additionally, the trained iterative retriever generalizes across different inference LLMs beyond the one used during training.
Authors: Sefika Efeoglu, Adrian Paschke
Abstract: Information Extraction (IE) is crucial for converting unstructured data into structured formats like Knowledge Graphs (KGs). A key task within IE is Relation Extraction (RE), which identifies relationships between entities in text. Various RE methods exist, including supervised, unsupervised, weakly supervised, and rule-based approaches. Recent studies leveraging pre-trained language models (PLMs) have shown significant success in this area. In the current era dominated by Large Language Models (LLMs), fine-tuning these models can overcome limitations associated with zero-shot LLM prompting-based RE methods, especially regarding domain adaptation challenges and identifying implicit relations between entities in sentences. These implicit relations, which cannot be easily extracted from a sentence's dependency tree, require logical inference for accurate identification. This work explores the performance of fine-tuned LLMs and their integration into the Retrieval Augmented-based (RAG) RE approach to address the challenges of identifying implicit relations at the sentence level, particularly when LLMs act as generators within the RAG framework. Empirical evaluations on the TACRED, TACRED-Revisited (TACREV), Re-TACRED, and SemEVAL datasets show significant performance improvements with fine-tuned LLMs, including Llama2-7B, Mistral-7B, and T5 (Large). Notably, our approach achieves substantial gains on SemEVAL, where implicit relations are common, surpassing previous results on this dataset. Additionally, our method outperforms previous works on TACRED, TACREV, and Re-TACRED, demonstrating exceptional performance across diverse evaluation scenarios.
Authors: Varsha Suresh, Salah A\"it-Mokhtar, Caroline Brun, Ioan Calapodescu
Abstract: Self-supervised learning models have revolutionized the field of speech processing. However, the process of fine-tuning these models on downstream tasks requires substantial computational resources, particularly when dealing with multiple speech-processing tasks. In this paper, we explore the potential of adapter-based fine-tuning in developing a unified model capable of effectively handling multiple spoken language processing tasks. The tasks we investigate are Automatic Speech Recognition, Phoneme Recognition, Intent Classification, Slot Filling, and Spoken Emotion Recognition. We validate our approach through a series of experiments on the SUPERB benchmark, and our results indicate that adapter-based fine-tuning enables a single encoder-decoder model to perform multiple speech processing tasks with an average improvement of 18.4% across the five target tasks while staying efficient in terms of parameter updates.
Authors: Lexin Zhou, Youmna Farag, Andreas Vlachos
Abstract: Research on dialogue constructiveness assessment focuses on (i) analysing conversational factors that influence individuals to take specific actions, win debates, change their perspectives or broaden their open-mindedness and (ii) predicting constructive outcomes following dialogues for such use cases. These objectives can be achieved by training either interpretable feature-based models (which often involve costly human annotations) or neural models such as pre-trained language models (which have empirically shown higher task accuracy but lack interpretability). We propose a novel LLM feature-based framework that combines the strengths of feature-based and neural approaches while mitigating their downsides, in assessing dialogue constructiveness. The framework first defines a set of dataset-independent and interpretable linguistic features, which can be extracted by both prompting an LLM and simple heuristics. Such features are then used to train LLM feature-based models. We apply this framework to three datasets of dialogue constructiveness and find that our LLM feature-based models significantly outperform standard feature-based models and neural models, and tend to learn more robust prediction rules instead of relying on superficial shortcuts (as seen with neural models). Further, we demonstrate that interpreting these LLM feature-based models can yield valuable insights into what makes a dialogue constructive.
Authors: Lingxi Zhang, Jing Zhang, Yanling Wang, Cuiping Li, Hong Chen
Abstract: Large-scale knowledge bases (KBs) like Freebase and Wikidata house millions of structured knowledge. Knowledge Base Question Answering (KBQA) provides a user-friendly way to access these valuable KBs via asking natural language questions. In order to improve the generalization capabilities of KBQA models, extensive research has embraced a retrieve-then-reason framework to retrieve relevant evidence for logical expression generation. These multi-stage efforts prioritize acquiring external sources but overlook the incorporation of new knowledge into their model parameters. In effect, even advanced language models and retrievers have knowledge boundaries, thereby limiting the generalization capabilities of previous KBQA models. Therefore, this paper develops KBLLaMA, which follows a learn-then-reason framework to inject new KB knowledge into a large language model for flexible end-to-end KBQA. At the core of KBLLaMA, we study (1) how to organize new knowledge about KBQA and (2) how to facilitate the learning of the organized knowledge. Extensive experiments on various KBQA generalization tasks showcase the state-of-the-art performance of KBLLaMA. Especially on the general benchmark GrailQA and domain-specific benchmark Bio-chemical, KBLLaMA respectively derives a performance gain of up to 3.8% and 9.8% compared to the baselines.
Authors: Gaurav Ghosal, Tatsunori Hashimoto, Aditi Raghunathan
Abstract: In this work, we study the impact of QA fine-tuning data on downstream factuality. We show that fine-tuning on lesser-known facts that are poorly stored during pretraining yields significantly worse factuality than fine-tuning on well-known facts, even when all facts are seen during pretraining. We prove this phenomenon theoretically, showing that training on lesser-known facts can lead the model to ignore subject entity names and instead output a generic plausible response even when the relevant factual knowledge is encoded in the model. On three question answering benchmarks (PopQA, Entity Questions, and MMLU) and two language models (Llama-2-7B and Mistral-7B), we find that (i) finetuning on a completely factual but lesser-known subset of the data deteriorates downstream factuality (5-10%) and (ii) finetuning on a subset of better-known examples matches or outperforms finetuning on the entire dataset. Ultimately, our results shed light on the interaction between pretrained knowledge and finetuning data and demonstrate the importance of taking into account how facts are stored in the pretrained model when fine-tuning for knowledge-intensive tasks.
Authors: Julia Kharchenko, Tanya Roosta, Aman Chadha, Chirag Shah
Abstract: Large Language Models (LLMs) attempt to imitate human behavior by responding to humans in a way that pleases them, including by adhering to their values. However, humans come from diverse cultures with different values. It is critical to understand whether LLMs showcase different values to the user based on the stereotypical values of a user's known country. We prompt different LLMs with a series of advice requests based on 5 Hofstede Cultural Dimensions -- a quantifiable way of representing the values of a country. Throughout each prompt, we incorporate personas representing 36 different countries and, separately, languages predominantly tied to each country to analyze the consistency in the LLMs' cultural understanding. Through our analysis of the responses, we found that LLMs can differentiate between one side of a value and another, as well as understand that countries have differing values, but will not always uphold the values when giving advice, and fail to understand the need to answer differently based on different cultural values. Rooted in these findings, we present recommendations for training value-aligned and culturally sensitive LLMs. More importantly, the methodology and the framework developed here can help further understand and mitigate culture and language alignment issues with LLMs.
Authors: Jing Yang, Yu Zhao, Yang Linyao, Xiao Wang, Fei-Yue Wang
Abstract: Temporal relation extraction (TRE) aims to grasp the evolution of events or actions, and thus shape the workflow of associated tasks, so it holds promise in helping understand task requests initiated by requesters in crowdsourcing systems. However, existing methods still struggle with limited and unevenly distributed annotated data. Therefore, inspired by the abundant global knowledge stored within pre-trained language models (PLMs), we propose a multi-task prompt learning framework for TRE (TemPrompt), incorporating prompt tuning and contrastive learning to tackle these issues. To elicit more effective prompts for PLMs, we introduce a task-oriented prompt construction approach that thoroughly takes the myriad factors of TRE into consideration for automatic prompt generation. In addition, we present temporal event reasoning as a supplement to bolster the model's focus on events and temporal cues. The experimental results demonstrate that TemPrompt outperforms all compared baselines across the majority of metrics under both standard and few-shot settings. A case study is provided to validate its effectiveness in crowdsourcing scenarios.
Authors: Yinghao Li, Siyu Miao, Heyan Huang, Yang Gao
Abstract: Domain adaptation aims to enable Large Language Models (LLMs) to generalize domain datasets unseen effectively during the training phase. However, factors such as the size of the model parameters and the scale of training data are general influencers and do not reflect the nuances of domain adaptation performance. This paper investigates the fine-grained factors affecting domain adaptation performance, analyzing the specific impact of `words' in training data on summarization tasks. We propose quantifying dataset learning difficulty as the learning difficulty of generative summarization, which is determined by two indicators: word-based compression rate and abstraction level. Our experiments conclude that, when considering dataset learning difficulty, the cross-domain overlap and the performance gain in summarization tasks exhibit an approximate linear relationship, which is not directly related to the number of words. Based on this finding, predicting a model's performance on unknown domain datasets is possible without undergoing training.
Authors: Pritika Ramu, Aparna Garimella, Sambaran Bandyopadhyay
Abstract: Understanding whether a generated table is of good quality is important to be able to use it in creating or editing documents using automatic methods. In this work, we underline that existing measures for table quality evaluation fail to capture the overall semantics of the tables, and sometimes unfairly penalize good tables and reward bad ones. We propose TabEval, a novel table evaluation strategy that captures table semantics by first breaking down a table into a list of natural language atomic statements and then compares them with ground truth statements using entailment-based measures. To validate our approach, we curate a dataset comprising of text descriptions for 1,250 diverse Wikipedia tables, covering a range of topics and structures, in contrast to the limited scope of existing datasets. We compare TabEval with existing metrics using unsupervised and supervised text-to-table generation methods, demonstrating its stronger correlation with human judgments of table quality across four datasets.
Authors: Yiduo Guo, Jie Fu, Huishuai Zhang, Dongyan Zhao, Yikang Shen
Abstract: Continual pre-training has increasingly become the predominant approach for adapting Large Language Models (LLMs) to new domains. This process involves updating the pre-trained LLM with a corpus from a new domain, resulting in a shift in the training distribution. To study the behavior of LLMs during this shift, we measured the model's performance throughout the continual pre-training process. we observed a temporary performance drop at the beginning, followed by a recovery phase, a phenomenon known as the "stability gap," previously noted in vision models classifying new classes. To address this issue and enhance LLM performance within a fixed compute budget, we propose three effective strategies: (1) Continually pre-training the LLM on a subset with a proper size for multiple epochs, resulting in faster performance recovery than pre-training the LLM on a large corpus in a single epoch; (2) Pre-training the LLM only on high-quality sub-corpus, which rapidly boosts domain performance; and (3) Using a data mixture similar to the pre-training data to reduce distribution gap. We conduct various experiments on Llama-family models to validate the effectiveness of our strategies in both medical continual pre-training and instruction tuning. For example, our strategies improve the average medical task performance of the OpenLlama-3B model from 36.2% to 40.7% with only 40% of the original training budget and enhance the average general task performance without causing forgetting. Furthermore, we apply our strategies to the Llama-3-8B model. The resulting model, Llama-3-Physician, achieves the best medical performance among current open-source models, and performs comparably to or even better than GPT-4 on several medical benchmarks. We release our models at \url{https://huggingface.co/YiDuo1999/Llama-3-Physician-8B-Instruct}.
URLs: https://huggingface.co/YiDuo1999/Llama-3-Physician-8B-Instruct
Authors: Tinh Son Luong, Thanh-Thien Le, Thang Viet Doan, Linh Ngo Van, Thien Huu Nguyen, Diep Thi-Ngoc Nguyen
Abstract: Existing toxic detection models face significant limitations, such as lack of transparency, customization, and reproducibility. These challenges stem from the closed-source nature of their training data and the paucity of explanations for their evaluation mechanism. To address these issues, we propose a dataset creation mechanism that integrates voting and chain-of-thought processes, producing a high-quality open-source dataset for toxic content detection. Our methodology ensures diverse classification metrics for each sample and includes both classification scores and explanatory reasoning for the classifications. We utilize the dataset created through our proposed mechanism to train our model, which is then compared against existing widely-used detectors. Our approach not only enhances transparency and customizability but also facilitates better fine-tuning for specific use cases. This work contributes a robust framework for developing toxic content detection models, emphasizing openness and adaptability, thus paving the way for more effective and user-specific content moderation solutions.
Authors: Qi Liu, Bo Wang, Nan Wang, Jiaxin Mao
Abstract: Recent studies have demonstrated the effectiveness of using large language language models (LLMs) in passage ranking. The listwise approaches, such as RankGPT, have become new state-of-the-art in this task. However, the efficiency of RankGPT models is limited by the maximum context length and relatively high latency of LLM inference. To address these issues, in this paper, we propose PE-Rank, leveraging the single passage embedding as a good context compression for efficient listwise passage reranking. By treating each passage as a special token, we can directly input passage embeddings into LLMs, thereby reducing input length. Additionally, we introduce an inference method that dynamically constrains the decoding space to these special tokens, accelerating the decoding process. For adapting the model to reranking, we employ listwise learning to rank loss for training. Evaluation results on multiple benchmarks demonstrate that PE-Rank significantly improves efficiency in both prefilling and decoding, while maintaining competitive ranking effectiveness. {The Code is available at \url{https://github.com/liuqi6777/pe_rank}.}
Authors: Siyuan Wang, Zhuohan Long, Zhihao Fan, Zhongyu Wei
Abstract: The rapid development of Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs) has exposed vulnerabilities to various adversarial attacks. This paper provides a comprehensive overview of jailbreaking research targeting both LLMs and MLLMs, highlighting recent advancements in evaluation benchmarks, attack techniques and defense strategies. Compared to the more advanced state of unimodal jailbreaking, multimodal domain remains underexplored. We summarize the limitations and potential research directions of multimodal jailbreaking, aiming to inspire future research and further enhance the robustness and security of MLLMs.
Authors: Wentao Shi, Mengqi Yuan, Junkang Wu, Qifan Wang, Fuli Feng
Abstract: Adapting Large Language Models (LLMs) for agent tasks is critical in developing language agents. Direct Preference Optimization (DPO) is a promising technique for this adaptation with the alleviation of compounding errors, offering a means to directly optimize Reinforcement Learning (RL) objectives. However, applying DPO to multi-turn tasks presents challenges due to the inability to cancel the partition function. Overcoming this obstacle involves making the partition function independent of the current state and addressing length disparities between preferred and dis-preferred trajectories. In this light, we replace the policy constraint with the state-action occupancy measure constraint in the RL objective and add length normalization to the Bradley-Terry model, yielding a novel loss function named DMPO for multi-turn agent tasks with theoretical explanations. Extensive experiments on three multi-turn agent task datasets confirm the effectiveness and superiority of the DMPO loss.
Authors: Zhengbang Yang, Haotian Xia, Jingxi Li, Zezhi Chen, Zhuangdi Zhu, Weining Shen
Abstract: Understanding sports is crucial for the advancement of Natural Language Processing (NLP) due to its intricate and dynamic nature. Reasoning over complex sports scenarios has posed significant challenges to current NLP technologies which require advanced cognitive capabilities. Toward addressing the limitations of existing benchmarks on sports understanding in the NLP field, we extensively evaluated mainstream large language models for various sports tasks. Our evaluation spans from simple queries on basic rules and historical facts to complex, context-specific reasoning, leveraging strategies from zero-shot to few-shot learning, and chain-of-thought techniques. In addition to unimodal analysis, we further assessed the sports reasoning capabilities of mainstream video language models to bridge the gap in multimodal sports understanding benchmarking. Our findings highlighted the critical challenges of sports understanding for NLP. We proposed a new benchmark based on a comprehensive overview of existing sports datasets and provided extensive error analysis which we hope can help identify future research priorities in this field.
Authors: Issey Sukeda, Risa Kishikawa, Satoshi Kodera
Abstract: Since the rise of large language models (LLMs), the domain adaptation has been one of the hot topics in various domains. Many medical LLMs trained with English medical dataset have made public recently. However, Japanese LLMs in medical domain still lack its research. Here we utilize multiple 70B-parameter LLMs for the first time and show that instruction tuning using Japanese medical question-answering dataset significantly improves the ability of Japanese LLMs to solve Japanese medical license exams, surpassing 50\% in accuracy. In particular, the Japanese-centric models exhibit a more significant leap in improvement through instruction tuning compared to their English-centric counterparts. This underscores the importance of continual pretraining and the adjustment of the tokenizer in our local language. We also examine two slightly different prompt formats, resulting in non-negligible performance improvement.
Authors: Jaspreet Ranjit, Brihi Joshi, Rebecca Dorn, Laura Petry, Olga Koumoundouros, Jayne Bottarini, Peichen Liu, Eric Rice, Swabha Swayamdipta
Abstract: Warning: Contents of this paper may be upsetting. Public attitudes towards key societal issues, expressed on online media, are of immense value in policy and reform efforts, yet challenging to understand at scale. We study one such social issue: homelessness in the U.S., by leveraging the remarkable capabilities of large language models to assist social work experts in analyzing millions of posts from Twitter. We introduce a framing typology: Online Attitudes Towards Homelessness (OATH) Frames: nine hierarchical frames capturing critiques, responses and perceptions. We release annotations with varying degrees of assistance from language models, with immense benefits in scaling: 6.5x speedup in annotation time while only incurring a 3 point F1 reduction in performance with respect to the domain experts. Our experiments demonstrate the value of modeling OATH-Frames over existing sentiment and toxicity classifiers. Our large-scale analysis with predicted OATH-Frames on 2.4M posts on homelessness reveal key trends in attitudes across states, time periods and vulnerable populations, enabling new insights on the issue. Our work provides a general framework to understand nuanced public attitudes at scale, on issues beyond homelessness.
Authors: Ruixuan Xiao, Wentao Ma, Ke Wang, Yuchuan Wu, Junbo Zhao, Haobo Wang, Fei Huang, Yongbin Li
Abstract: LLM-based agents have emerged as promising tools, which are crafted to fulfill complex tasks by iterative planning and action. However, these agents are susceptible to undesired planning hallucinations when lacking specific knowledge for expertise-intensive tasks. To address this, preliminary attempts are made to enhance planning reliability by incorporating external workflow-related knowledge. Despite the promise, such infused knowledge is mostly disorganized and diverse in formats, lacking rigorous formalization and comprehensive comparisons. Motivated by this, we formalize different formats of workflow knowledge and present FlowBench, the first benchmark for workflow-guided planning. FlowBench covers 51 different scenarios from 6 domains, with knowledge presented in diverse formats. To assess different LLMs on FlowBench, we design a multi-tiered evaluation framework. We evaluate the efficacy of workflow knowledge across multiple formats, and the results indicate that current LLM agents need considerable improvements for satisfactory planning. We hope that our challenging benchmark can pave the way for future agent planning research.
Authors: Zhiwei Fei, Songyang Zhang, Xiaoyu Shen, Dawei Zhu, Xiao Wang, Maosong Cao, Fengzhe Zhou, Yining Li, Wenwei Zhang, Dahua Lin, Kai Chen, Jidong Ge
Abstract: While large language models (LLMs) have showcased impressive capabilities, they struggle with addressing legal queries due to the intricate complexities and specialized expertise required in the legal field. In this paper, we introduce InternLM-Law, a specialized LLM tailored for addressing diverse legal queries related to Chinese laws, spanning from responding to standard legal questions (e.g., legal exercises in textbooks) to analyzing complex real-world legal situations. We meticulously construct a dataset in the Chinese legal domain, encompassing over 1 million queries, and implement a data filtering and processing pipeline to ensure its diversity and quality. Our training approach involves a novel two-stage process: initially fine-tuning LLMs on both legal-specific and general-purpose content to equip the models with broad knowledge, followed by exclusive fine-tuning on high-quality legal data to enhance structured output generation. InternLM-Law achieves the highest average performance on LawBench, outperforming state-of-the-art models, including GPT-4, on 13 out of 20 subtasks. We make InternLM-Law and our dataset publicly available to facilitate future research in applying LLMs within the legal domain.
Authors: Yu Nakagome, Michael Hentschel
Abstract: Despite recent advances in end-to-end speech recognition methods, their output is biased to the training data's vocabulary, resulting in inaccurate recognition of unknown terms or proper nouns. To improve the recognition accuracy for a given set of such terms, we propose an adaptation parameter-free approach based on Self-conditioned CTC. Our method improves the recognition accuracy of misrecognized target keywords by substituting their intermediate CTC predictions with corrected labels, which are then passed on to the subsequent layers. First, we create pairs of correct labels and recognition error instances for a keyword list using Text-to-Speech and a recognition model. We use these pairs to replace intermediate prediction errors by the labels. Conditioning the subsequent layers of the encoder on the labels, it is possible to acoustically evaluate the target keywords. Experiments conducted in Japanese demonstrated that our method successfully improved the F1 score for unknown words.
Authors: Zhengliang Shi, Shuo Zhang, Weiwei Sun, Shen Gao, Pengjie Ren, Zhumin Chen, Zhaochun Ren
Abstract: Multi-Hop Question Answering (MHQA) tasks present a significant challenge for large language models (LLMs) due to the intensive knowledge required. Current solutions, like Retrieval-Augmented Generation, typically retrieve potential documents from an external corpus to read an answer. However, the performance of this retrieve-then-read paradigm is constrained by the retriever and the inevitable noise in the retrieved documents. To mitigate these challenges, we introduce a novel generate-then-ground (GenGround) framework, synergizing the parametric knowledge of LLMs and external documents to solve a multi-hop question. GenGround empowers LLMs to alternate two phases until the final answer is derived: (1) formulate a simpler, single-hop question and directly generate the answer; (2) ground the question-answer pair in retrieved documents, amending any wrong predictions in the answer. We also propose an instructional grounding distillation method to generalize our method into smaller models. Extensive experiments conducted on four datasets illustrate the superiority of our method.
Authors: Candida M. Greco, Lucio La Cava, Andrea Tagarelli
Abstract: Verbs form the backbone of language, providing the structure and meaning to sentences. Yet, their intricate semantic nuances pose a longstanding challenge. Understanding verb relations through the concept of lexical entailment is crucial for comprehending sentence meanings and grasping verb dynamics. This work investigates the capabilities of eight Large Language Models in recognizing lexical entailment relations among verbs through differently devised prompting strategies and zero-/few-shot settings over verb pairs from two lexical databases, namely WordNet and HyperLex. Our findings unveil that the models can tackle the lexical entailment recognition task with moderately good performance, although at varying degree of effectiveness and under different conditions. Also, utilizing few-shot prompting can enhance the models' performance. However, perfectly solving the task arises as an unmet challenge for all examined LLMs, which raises an emergence for further research developments on this topic.
Authors: Yannis Tevissen, Khalil Guetari, Fr\'ed\'eric Petitpont
Abstract: Video content creators need efficient tools to repurpose content, a task that often requires complex manual or automated searches. Crafting a new video from large video libraries remains a challenge. In this paper we introduce the task of Video Library Question Answering (VLQA) through an interoperable architecture that applies Retrieval Augmented Generation (RAG) to video libraries. We propose a system that uses large language models (LLMs) to generate search queries, retrieving relevant video moments indexed by speech and visual metadata. An answer generation module then integrates user queries with this metadata to produce responses with specific video timestamps. This approach shows promise in multimedia content retrieval, and AI-assisted video content creation.
Authors: Haiquan Zhao, Lingyu Li, Shisong Chen, Shuqi Kong, Jiaan Wang, Kexing Huang, Tianle Gu, Yixu Wang, Dandan Liang, Zhixu Li, Tan Teng, Yanghua Xiao, Yingchun Wang
Abstract: Emotion Support Conversation (ESC) is a crucial application, which aims to reduce human stress, offer emotional guidance, and ultimately enhance human mental and physical well-being. With the advancement of Large Language Models (LLMs), many researchers have employed LLMs as the ESC models. However, the evaluation of these LLM-based ESCs remains uncertain. Inspired by the awesome development of role-playing agents, we propose an ESC Evaluation framework (ESC-Eval), which uses a role-playing agent to interact with ESC models, followed by a manual evaluation of the interactive dialogues. In detail, we first re-organize 2,801 role-playing cards from seven existing datasets to define the roles of the role-playing agent. Second, we train a specific role-playing model called ESC-Role which behaves more like a confused person than GPT-4. Third, through ESC-Role and organized role cards, we systematically conduct experiments using 14 LLMs as the ESC models, including general AI-assistant LLMs (ChatGPT) and ESC-oriented LLMs (ExTES-Llama). We conduct comprehensive human annotations on interactive multi-turn dialogues of different ESC models. The results show that ESC-oriented LLMs exhibit superior ESC abilities compared to general AI-assistant LLMs, but there is still a gap behind human performance. Moreover, to automate the scoring process for future ESC models, we developed ESC-RANK, which trained on the annotated data, achieving a scoring performance surpassing 35 points of GPT-4. Our data and code are available at https://github.com/haidequanbu/ESC-Eval.
Authors: Wentong Chen, Yankai Lin, ZhenHao Zhou, HongYun Huang, Yantao Jia, Zhao Cao, Ji-Rong Wen
Abstract: In-Context Learning (ICL) is a critical capability of Large Language Models (LLMs) as it empowers them to comprehend and reason across interconnected inputs. Evaluating the ICL ability of LLMs can enhance their utilization and deepen our understanding of how this ability is acquired at the training stage. However, existing evaluation frameworks primarily focus on language abilities and knowledge, often overlooking the assessment of ICL ability. In this work, we introduce the ICLEval benchmark to evaluate the ICL abilities of LLMs, which encompasses two key sub-abilities: exact copying and rule learning. Through the ICLEval benchmark, we demonstrate that ICL ability is universally present in different LLMs, and model size is not the sole determinant of ICL efficacy. Surprisingly, we observe that ICL abilities, particularly copying, develop early in the pretraining process and stabilize afterward. Our source codes and benchmark are released at https://github.com/yiye3/ICLEval.
Authors: Shamane Siriwardhana, Mark McQuade, Thomas Gauthier, Lucas Atkins, Fernando Fernandes Neto, Luke Meyers, Anneketh Vij, Tyler Odenthal, Charles Goddard, Mary MacCarthy, Jacob Solawetz
Abstract: We conducted extensive experiments on domain adaptation of the Meta-Llama-3-70B-Instruct model on SEC data, exploring its performance on both general and domain-specific benchmarks. Our focus included continual pre-training (CPT) and model merging, aiming to enhance the model's domain-specific capabilities while mitigating catastrophic forgetting. Through this study, we evaluated the impact of integrating financial regulatory data into a robust language model and examined the effectiveness of our model merging techniques in preserving and improving the model's instructive abilities. The model is accessible at hugging face: https://huggingface.co/arcee-ai/Llama-3-SEC-Base, arcee-ai/Llama-3-SEC-Base. This is an intermediate checkpoint of our final model, which has seen 20B tokens so far. The full model is still in the process of training. This is a preprint technical report with thorough evaluations to understand the entire process.
Authors: Florin Cuconasu, Giovanni Trappolini, Nicola Tonellotto, Fabrizio Silvestri
Abstract: Retrieval Augmented Generation (RAG) represents a significant advancement in artificial intelligence combining a retrieval phase with a generative phase, with the latter typically being powered by large language models (LLMs). The current common practices in RAG involve using "instructed" LLMs, which are fine-tuned with supervised training to enhance their ability to follow instructions and are aligned with human preferences using state-of-the-art techniques. Contrary to popular belief, our study demonstrates that base models outperform their instructed counterparts in RAG tasks by 20% on average under our experimental settings. This finding challenges the prevailing assumptions about the superiority of instructed LLMs in RAG applications. Further investigations reveal a more nuanced situation, questioning fundamental aspects of RAG and suggesting the need for broader discussions on the topic; or, as Fromm would have it, "Seldom is a glance at the statistics enough to understand the meaning of the figures".
Authors: Yuanjie Lyu, Zihan Niu, Zheyong Xie, Chao Zhang, Tong Xu, Yang Wang, Enhong Chen
Abstract: Despite the significant progress of large language models (LLMs) in various tasks, they often produce factual errors due to their limited internal knowledge. Retrieval-Augmented Generation (RAG), which enhances LLMs with external knowledge sources, offers a promising solution. However, these methods can be misled by irrelevant paragraphs in retrieved documents. Due to the inherent uncertainty in LLM generation, inputting the entire document may introduce off-topic information, causing the model to deviate from the central topic and affecting the relevance of the generated content. To address these issues, we propose the Retrieve-Plan-Generation (RPG) framework. RPG generates plan tokens to guide subsequent generation in the plan stage. In the answer stage, the model selects relevant fine-grained paragraphs based on the plan and uses them for further answer generation. This plan-answer process is repeated iteratively until completion, enhancing generation relevance by focusing on specific topics. To implement this framework efficiently, we utilize a simple but effective multi-task prompt-tuning method, enabling the existing LLMs to handle both planning and answering. We comprehensively compare RPG with baselines across 5 knowledge-intensive generation tasks, demonstrating the effectiveness of our approach.
Authors: Zeyao Ma, Bohan Zhang, Jing Zhang, Jifan Yu, Xiaokang Zhang, Xiaohan Zhang, Sijia Luo, Xi Wang, Jie Tang
Abstract: We introduce SpreadsheetBench, a challenging spreadsheet manipulation benchmark exclusively derived from real-world scenarios, designed to immerse current large language models (LLMs) in the actual workflow of spreadsheet users. Unlike existing benchmarks that rely on synthesized queries and simplified spreadsheet files, SpreadsheetBench is built from 912 real questions gathered from online Excel forums, which reflect the intricate needs of users. The associated spreadsheets from the forums contain a variety of tabular data such as multiple tables, non-standard relational tables, and abundant non-textual elements. Furthermore, we propose a more reliable evaluation metric akin to online judge platforms, where multiple spreadsheet files are created as test cases for each instruction, ensuring the evaluation of robust solutions capable of handling spreadsheets with varying values. Our comprehensive evaluation of various LLMs under both single-round and multi-round inference settings reveals a substantial gap between the state-of-the-art (SOTA) models and human performance, highlighting the benchmark's difficulty.
Authors: Lichao Zhang, Jia Yu, Shuai Zhang, Long Li, Yangyang Zhong, Guanbao Liang, Yuming Yan, Qing Ma, Fangsheng Weng, Fayu Pan, Jing Li, Renjun Xu, Zhenzhong Lan
Abstract: Large Language Models (LLMs) have significantly advanced user-bot interactions, enabling more complex and coherent dialogues. However, the prevalent text-only modality might not fully exploit the potential for effective user engagement. This paper explores the impact of multi-modal interactions, which incorporate images and audio alongside text, on user engagement in chatbot conversations. We conduct a comprehensive analysis using a diverse set of chatbots and real-user interaction data, employing metrics such as retention rate and conversation length to evaluate user engagement. Our findings reveal a significant enhancement in user engagement with multi-modal interactions compared to text-only dialogues. Notably, the incorporation of a third modality significantly amplifies engagement beyond the benefits observed with just two modalities. These results suggest that multi-modal interactions optimize cognitive processing and facilitate richer information comprehension. This study underscores the importance of multi-modality in chatbot design, offering valuable insights for creating more engaging and immersive AI communication experiences and informing the broader AI community about the benefits of multi-modal interactions in enhancing user engagement.
Authors: Yongqi Fan, Hongli Sun, Kui Xue, Xiaofan Zhang, Shaoting Zhang, Tong Ruan
Abstract: Numerous advanced Large Language Models (LLMs) now support context lengths up to 128K, and some extend to 200K. Some benchmarks in the generic domain have also followed up on evaluating long-context capabilities. In the medical domain, tasks are distinctive due to the unique contexts and need for domain expertise, necessitating further evaluation. However, despite the frequent presence of long texts in medical scenarios, evaluation benchmarks of long-context capabilities for LLMs in this field are still rare. In this paper, we propose MedOdyssey, the first medical long-context benchmark with seven length levels ranging from 4K to 200K tokens. MedOdyssey consists of two primary components: the medical-context "needles in a haystack" task and a series of tasks specific to medical applications, together comprising 10 datasets. The first component includes challenges such as counter-intuitive reasoning and novel (unknown) facts injection to mitigate knowledge leakage and data contamination of LLMs. The second component confronts the challenge of requiring professional medical expertise. Especially, we design the ``Maximum Identical Context'' principle to improve fairness by guaranteeing that different LLMs observe as many identical contexts as possible. Our experiment evaluates advanced proprietary and open-source LLMs tailored for processing long contexts and presents detailed performance analyses. This highlights that LLMs still face challenges and need for further research in this area. Our code and data are released in the repository: \url{https://github.com/JOHNNY-fans/MedOdyssey.}
Authors: Giulio Salierno, Rosamaria Bert\`e, Luca Attias, Carla Morrone, Dario Pettazzoni, Daniela Battisti
Abstract: Recent advances in Natural Language Processing have demonstrated the effectiveness of pretrained language models like BERT for a variety of downstream tasks. We present GiusBERTo, the first BERT-based model specialized for anonymizing personal data in Italian legal documents. GiusBERTo is trained on a large dataset of Court of Auditors decisions to recognize entities to anonymize, including names, dates, locations, while retaining contextual relevance. We evaluate GiusBERTo on a held-out test set and achieve 97% token-level accuracy. GiusBERTo provides the Italian legal community with an accurate and tailored BERT model for de-identification, balancing privacy and data protection.
Authors: Jinge Wu, Zhaolong Wu, Abul Hasan, Yunsoo Kim, Jason P. Y. Cheung, Teng Zhang, Honghan Wu
Abstract: This study proposes an approach for error correction in clinical radiology reports, leveraging large language models (LLMs) and retrieval-augmented generation (RAG) techniques. The proposed framework employs internal and external retrieval mechanisms to extract relevant medical entities and relations from the report and external knowledge sources. A three-stage inference process is introduced, decomposing the task into error detection, localization, and correction subtasks, which enhances the explainability and performance of the system. The effectiveness of the approach is evaluated using a benchmark dataset created by corrupting real-world radiology reports with realistic errors, guided by domain experts. Experimental results demonstrate the benefits of the proposed methods, with the combination of internal and external retrieval significantly improving the accuracy of error detection, localization, and correction across various state-of-the-art LLMs. The findings contribute to the development of more robust and reliable error correction systems for clinical documentation.
Authors: Ishaan Watts, Varun Gumma, Aditya Yadavalli, Vivek Seshadri, Manohar Swaminathan, Sunayana Sitaram
Abstract: Evaluation of multilingual Large Language Models (LLMs) is challenging due to a variety of factors -- the lack of benchmarks with sufficient linguistic diversity, contamination of popular benchmarks into LLM pre-training data and the lack of local, cultural nuances in translated benchmarks. In this work, we study human and LLM-based evaluation in a multilingual, multi-cultural setting. We evaluate 30 models across 10 Indic languages by conducting 90K human evaluations and 30K LLM-based evaluations and find that models such as GPT-4o and Llama-3 70B consistently perform best for most Indic languages. We build leaderboards for two evaluation settings - pairwise comparison and direct assessment and analyse the agreement between humans and LLMs. We find that humans and LLMs agree fairly well in the pairwise setting but the agreement drops for direct assessment evaluation especially for languages such as Bengali and Odia. We also check for various biases in human and LLM-based evaluation and find evidence of self-bias in the GPT-based evaluator. Our work presents a significant step towards scaling up multilingual evaluation of LLMs.
Authors: Inessa Fedorova, Aleksei Musatow
Abstract: The paraphrase identification task involves measuring semantic similarity between two short sentences. It is a tricky task, and multilingual paraphrase identification is even more challenging. In this work, we train a bi-encoder model in a contrastive manner to detect hard paraphrases across multiple languages. This approach allows us to use model-produced embeddings for various tasks, such as semantic search. We evaluate our model on downstream tasks and also assess embedding space quality. Our performance is comparable to state-of-the-art cross-encoders, with only a minimal relative drop of 7-10% on the chosen dataset, while keeping decent quality of embeddings.
Authors: Jingyi Sun, Pepa Atanasova, Isabelle Augenstein
Abstract: Explaining the decision-making process of machine learning models is crucial for ensuring their reliability and fairness. One popular explanation form highlights key input features, such as i) tokens (e.g., Shapley Values and Integrated Gradients), ii) interactions between tokens (e.g., Bivariate Shapley and Attention-based methods), or iii) interactions between spans of the input (e.g., Louvain Span Interactions). However, these explanation types have only been studied in isolation, making it difficult to judge their respective applicability. To bridge this gap, we propose a unified framework that facilitates a direct comparison between highlight and interactive explanations comprised of four diagnostic properties. Through extensive analysis across these three types of input feature explanations--each utilizing three different explanation techniques--across two datasets and two models, we reveal that each explanation type excels in terms of different diagnostic properties. In our experiments, highlight explanations are the most faithful to a model's prediction, and interactive explanations provide better utility for learning to simulate a model's predictions. These insights further highlight the need for future research to develop combined methods that enhance all diagnostic properties.
Authors: Badr AlKhamissi, Greta Tuckute, Antoine Bosselut, Martin Schrimpf
Abstract: Large Language Models (LLMs) have been shown to be effective models of the human language system, with some models predicting most explainable variance of brain activity in current datasets. Even in untrained models, the representations induced by architectural priors can exhibit reasonable alignment to brain data. In this work, we investigate the key architectural components driving the surprising alignment of untrained models. To estimate LLM-to-brain similarity, we first select language-selective units within an LLM, similar to how neuroscientists identify the language network in the human brain. We then benchmark the brain alignment of these LLM units across five different brain recording datasets. By isolating critical components of the Transformer architecture, we identify tokenization strategy and multihead attention as the two major components driving brain alignment. A simple form of recurrence further improves alignment. We further demonstrate this quantitative brain alignment of our model by reproducing landmark studies in the language neuroscience field, showing that localized model units -- just like language voxels measured empirically in the human brain -- discriminate more reliably between lexical than syntactic differences, and exhibit similar response profiles under the same experimental conditions. Finally, we demonstrate the utility of our model's representations for language modeling, achieving improved sample and parameter efficiency over comparable architectures. Our model's estimates of surprisal sets a new state-of-the-art in the behavioral alignment to human reading times. Taken together, we propose a highly brain- and behaviorally-aligned model that conceptualizes the human language system as an untrained shallow feature encoder, with structural priors, combined with a trained decoder to achieve efficient and performant language processing.
Authors: Lin Long, Rui Wang, Ruixuan Xiao, Junbo Zhao, Xiao Ding, Gang Chen, Haobo Wang
Abstract: Within the evolving landscape of deep learning, the dilemma of data quantity and quality has been a long-standing problem. The recent advent of Large Language Models (LLMs) offers a data-centric solution to alleviate the limitations of real-world data with synthetic data generation. However, current investigations into this field lack a unified framework and mostly stay on the surface. Therefore, this paper provides an organization of relevant studies based on a generic workflow of synthetic data generation. By doing so, we highlight the gaps within existing research and outline prospective avenues for future study. This work aims to shepherd the academic and industrial communities towards deeper, more methodical inquiries into the capabilities and applications of LLMs-driven synthetic data generation.
Authors: Victor Hugo Nascimento Rocha, Igor Cataneo Silveira, Paulo Pirozelli, Denis Deratani Mau\'a, Fabio Gagliardi Cozman
Abstract: The recent success of Large Language Models (LLMs) has sparked concerns about their potential to spread misinformation. As a result, there is a pressing need for tools to identify ``fake arguments'' generated by such models. To create these tools, examples of texts generated by LLMs are needed. This paper introduces a methodology to obtain good, bad and ugly arguments from argumentative essays produced by ChatGPT, OpenAI's LLM. We then describe a novel dataset containing a set of diverse arguments, ArGPT. We assess the effectiveness of our dataset and establish baselines for several argumentation-related tasks. Finally, we show that the artificially generated data relates well to human argumentation and thus is useful as a tool to train and test systems for the defined tasks.
Authors: Muhammad Imran, Olga Kellert, Carlos G\'omez-Rodr\'iguez
Abstract: Sentiment Analysis (SA) is a crucial aspect of Natural Language Processing (NLP), addressing subjective assessments in textual content. Syntactic parsing is useful in SA because explicit syntactic information can improve accuracy while providing explainability, but it tends to be a computational bottleneck in practice due to the slowness of parsing algorithms. This paper addresses said bottleneck by using a SEquence Labeling Syntactic Parser (SELSP) to inject syntax into SA. By treating dependency parsing as a sequence labeling problem, we greatly enhance the speed of syntax-based SA. SELSP is trained and evaluated on a ternary polarity classification task, demonstrating its faster performance and better accuracy in polarity prediction tasks compared to conventional parsers like Stanza and to heuristic approaches that use shallow syntactic rules for SA like VADER. This increased speed and improved accuracy make SELSP particularly appealing to SA practitioners in both research and industry. In addition, we test several sentiment dictionaries on our SELSP to see which one improves the performance in polarity prediction tasks. Moreover, we compare the SELSP with Transformer-based models trained on a 5-label classification task. The results show that dictionaries that capture polarity judgment variation provide better results than dictionaries that ignore polarity judgment variation. Moreover, we show that SELSP is considerably faster than Transformer-based models in polarity prediction tasks.
Authors: Wei He, Marco Idiart, Carolina Scarton, Aline Villavicencio
Abstract: Accurately modeling idiomatic or non-compositional language has been a longstanding challenge in Natural Language Processing (NLP). This is partly because these expressions do not derive their meanings solely from their constituent words, but also due to the scarcity of relevant data resources, and their impact on the performance of downstream tasks such as machine translation and simplification. In this paper we propose an approach to model idiomaticity effectively using a triplet loss that incorporates the asymmetric contribution of components words to an idiomatic meaning for training language models by using adaptive contrastive learning and resampling miners to build an idiomatic-aware learning objective. Our proposed method is evaluated on a SemEval challenge and outperforms previous alternatives significantly in many metrics.
Authors: Chenglong Wang, Hang Zhou, Kaiyan Chang, Bei Li, Yongyu Mu, Tong Xiao, Tongran Liu, Jingbo Zhu
Abstract: Alignment training is crucial for enabling large language models (LLMs) to cater to human intentions and preferences. It is typically performed based on two stages with different objectives: instruction-following alignment and human-preference alignment. However, aligning LLMs with these objectives in sequence suffers from an inherent problem: the objectives may conflict, and the LLMs cannot guarantee to simultaneously align with the instructions and human preferences well. To response to these, in this work, we propose a Hybrid Alignment Training (Hbat) approach, based on alternating alignment and modified elastic weight consolidation methods. The basic idea is to alternate between different objectives during alignment training, so that better collaboration can be achieved between the two alignment tasks.We experiment with Hbat on summarization and dialogue tasks. Experimental results show that the proposed \textsc{Hbat} can significantly outperform all baselines. Notably, Hbat yields consistent performance gains over the traditional two-stage alignment training when using both proximal policy optimization and direct preference optimization.
Authors: Chia-Yu Hung, Navonil Majumder, Ambuj Mehrish, Soujanya Poria
Abstract: The widespread applicability and increasing omnipresence of LLMs have instigated a need to align LLM responses to user and stakeholder preferences. Many preference optimization approaches have been proposed that fine-tune LLM parameters to achieve good alignment. However, such parameter tuning is known to interfere with model performance on many tasks. Moreover, keeping up with shifting user preferences is tricky in such a situation. Decoding-time alignment with reward model guidance solves these issues at the cost of increased inference time. However, most of such methods fail to strike the right balance between exploration and exploitation of reward -- often due to the conflated formulation of these two aspects - to give well-aligned responses. To remedy this we decouple these two aspects and implement them in an evolutionary fashion: exploration is enforced by decoding from mutated instructions and exploitation is represented as the periodic replacement of poorly-rewarded generations with well-rewarded ones. Empirical evidences indicate that this strategy outperforms many preference optimization and decode-time alignment approaches on two widely accepted alignment benchmarks AlpacaEval 2 and MT-Bench. Our implementation will be available at: https://darwin-alignment.github.io.
Authors: Subhankar Maity, Aniket Deroy, Sudeshna Sarkar
Abstract: We evaluate the effectiveness of GPT-4 Turbo in generating educational questions from NCERT textbooks in zero-shot mode. Our study highlights GPT-4 Turbo's ability to generate questions that require higher-order thinking skills, especially at the "understanding" level according to Bloom's Revised Taxonomy. While we find a notable consistency between questions generated by GPT-4 Turbo and those assessed by humans in terms of complexity, there are occasional differences. Our evaluation also uncovers variations in how humans and machines evaluate question quality, with a trend inversely related to Bloom's Revised Taxonomy levels. These findings suggest that while GPT-4 Turbo is a promising tool for educational question generation, its efficacy varies across different cognitive levels, indicating a need for further refinement to fully meet educational standards.
Authors: Makesh Narsimhan Sreedhar, Traian Rebedea, Christopher Parisien
Abstract: Dialogue policies play a crucial role in developing task-oriented dialogue systems, yet their development and maintenance are challenging and typically require substantial effort from experts in dialogue modeling. While in many situations, large amounts of conversational data are available for the task at hand, people lack an effective solution able to extract dialogue policies from this data. In this paper, we address this gap by first illustrating how Large Language Models (LLMs) can be instrumental in extracting dialogue policies from datasets, through the conversion of conversations into a unified intermediate representation consisting of canonical forms. We then propose a novel method for generating dialogue policies utilizing a controllable and interpretable graph-based methodology. By combining canonical forms across conversations into a flow network, we find that running graph traversal algorithms helps in extracting dialogue flows. These flows are a better representation of the underlying interactions than flows extracted by prompting LLMs. Our technique focuses on giving conversation designers greater control, offering a productivity tool to improve the process of developing dialogue policies.
Authors: Irune Zubiaga, Aitor Soroa, Rodrigo Agerri
Abstract: The proliferation of misinformation and harmful narratives in online discourse has underscored the critical need for effective Counter Narrative (CN) generation techniques. However, existing automatic evaluation methods often lack interpretability and fail to capture the nuanced relationship between generated CNs and human perception. Aiming to achieve a higher correlation with human judgments, this paper proposes a novel approach to asses generated CNs that consists on the use of a Large Language Model (LLM) as a evaluator. By comparing generated CNs pairwise in a tournament-style format, we establish a model ranking pipeline that achieves a correlation of $0.88$ with human preference. As an additional contribution, we leverage LLMs as zero-shot (ZS) CN generators and conduct a comparative analysis of chat, instruct, and base models, exploring their respective strengths and limitations. Through meticulous evaluation, including fine-tuning experiments, we elucidate the differences in performance and responsiveness to domain-specific data. We conclude that chat-aligned models in ZS are the best option for carrying out the task, provided they do not refuse to generate an answer due to security concerns.
Authors: Yanis Labrak, Gabriel Meseguer-Brocal, Elena V. Epure
Abstract: In recent years, generated content in music has gained significant popularity, with large language models being effectively utilized to produce human-like lyrics in various styles, themes, and linguistic structures. This technological advancement supports artists in their creative processes but also raises issues of authorship infringement, consumer satisfaction and content spamming. To address these challenges, methods for detecting generated lyrics are necessary. However, existing works have not yet focused on this specific modality or on creative text in general regarding machine-generated content detection methods and datasets. In response, we have curated the first dataset of high-quality synthetic lyrics and conducted a comprehensive quantitative evaluation of various few-shot content detection approaches, testing their generalization capabilities and complementing this with a human evaluation. Our best few-shot detector, based on LLM2Vec, surpasses stylistic and statistical methods, which are shown competitive in other domains at distinguishing human-written from machine-generated content. It also shows good generalization capabilities to new artists and models, and effectively detects post-generation paraphrasing. This study emphasizes the need for further research on creative content detection, particularly in terms of generalization and scalability with larger song catalogs. All datasets, pre-processing scripts, and code are available publicly on GitHub and Hugging Face under the Apache 2.0 license.
Authors: Qingyang Zhu, Xiang Hu, Pengyu Ji, Wei Wu, Kewei Tu
Abstract: As a cornerstone in language modeling, tokenization involves segmenting text inputs into pre-defined atomic units. Conventional statistical tokenizers often disrupt constituent boundaries within words, thereby corrupting semantic information. To address this drawback, we introduce morphological structure guidance to tokenization and propose a deep model to induce character-level structures of words. Specifically, the deep model jointly encodes internal structures and representations of words with a mechanism named $\textit{MorphOverriding}$ to ensure the indecomposability of morphemes. By training the model with self-supervised objectives, our method is capable of inducing character-level structures that align with morphological rules without annotated training data. Based on the induced structures, our algorithm tokenizes words through vocabulary matching in a top-down manner. Empirical results indicate that the proposed method effectively retains complete morphemes and outperforms widely adopted methods such as BPE and WordPiece on both morphological segmentation tasks and language modeling tasks. The code will be released later.
Authors: Charlotte Pouw, Marianne de Heer Kloots, Afra Alishahi, Willem Zuidema
Abstract: Human listeners effortlessly compensate for phonological changes during speech perception, often unconsciously inferring the intended sounds. For example, listeners infer the underlying /n/ when hearing an utterance such as "clea[m] pan", where [m] arises from place assimilation to the following labial [p]. This article explores how the neural speech recognition model Wav2Vec2 perceives assimilated sounds, and identifies the linguistic knowledge that is implemented by the model to compensate for assimilation during Automatic Speech Recognition (ASR). Using psycholinguistic stimuli, we systematically analyze how various linguistic context cues influence compensation patterns in the model's output. Complementing these behavioral experiments, our probing experiments indicate that the model shifts its interpretation of assimilated sounds from their acoustic form to their underlying form in its final layers. Finally, our causal intervention experiments suggest that the model relies on minimal phonological context cues to accomplish this shift. These findings represent a step towards better understanding the similarities and differences in phonological processing between neural ASR models and humans.
Authors: Yanran Chen, Hannes Gr\"oner, Sina Zarrie{\ss}, Steffen Eger
Abstract: Natural Language Generation (NLG), and more generally generative AI, are among the currently most impactful research fields. Creative NLG, such as automatic poetry generation, is a fascinating niche in this area. While most previous research has focused on forms of the Turing test when evaluating automatic poetry generation - can humans distinguish between automatic and human generated poetry - we evaluate the diversity of automatically generated poetry, by comparing distributions of generated poetry to distributions of human poetry along structural, lexical, semantic and stylistic dimensions, assessing different model types (word vs. character-level, general purpose LLMs vs. poetry-specific models), including the very recent LLaMA3, and types of fine-tuning (conditioned vs. unconditioned). We find that current automatic poetry systems are considerably underdiverse along multiple dimensions - they often do not rhyme sufficiently, are semantically too uniform and even do not match the length distribution of human poetry. Our experiments reveal, however, that style-conditioning and character-level modeling clearly increases diversity across virtually all dimensions we explore. Our identified limitations may serve as the basis for more genuinely diverse future poetry generation models.
Authors: Doyoung Kim, Jongwon Lee, Jinho Park, Minjoon Seo
Abstract: Language models have demonstrated impressive capabilities across various natural language processing tasks, yet they struggle with planning tasks requiring multi-step simulations. Inspired by human cognitive processes, this paper investigates the optimal planning power of language models that can construct a cognitive map of a given environment. Our experiments demonstrate that cognitive map significantly enhances the performance of both optimal and reachable planning generation ability in the Gridworld path planning task. We observe that our method showcases two key characteristics similar to human cognition: \textbf{generalization of its planning ability to extrapolated environments and rapid adaptation with limited training data.} We hope our findings in the Gridworld task provide insights into modeling human cognitive processes in language models, potentially leading to the development of more advanced and robust systems that better resemble human cognition.
Authors: Georgios Paraskevopoulos, Chara Tsoukala, Athanasios Katsamanis, Vassilis Katsouros
Abstract: The development of speech technologies for languages with limited digital representation poses significant challenges, primarily due to the scarcity of available data. This issue is exacerbated in the era of large, data-intensive models. Recent research has underscored the potential of leveraging weak supervision to augment the pool of available data. In this study, we compile an 800-hour corpus of Modern Greek from podcasts and employ Whisper large-v3 to generate silver transcriptions. This corpus is utilized to fine-tune our models, aiming to assess the efficacy of this approach in enhancing ASR performance. Our analysis spans 16 distinct podcast domains, alongside evaluations on established datasets for Modern Greek. The findings indicate consistent WER improvements, correlating with increases in both data volume and model size. Our study confirms that assembling large, weakly supervised corpora serves as a cost-effective strategy for advancing speech technologies in under-resourced languages.
Authors: Tim Schopf, Florian Matthes
Abstract: Scientific literature searches are often exploratory, whereby users are not yet familiar with a particular field or concept but are interested in learning more about it. However, existing systems for scientific literature search are typically tailored to keyword-based lookup searches, limiting the possibilities for exploration. We propose NLP-KG, a feature-rich system designed to support the exploration of research literature in unfamiliar natural language processing (NLP) fields. In addition to a semantic search, NLP-KG allows users to easily find survey papers that provide a quick introduction to a field of interest. Further, a Fields of Study hierarchy graph enables users to familiarize themselves with a field and its related areas. Finally, a chat interface allows users to ask questions about unfamiliar concepts or specific articles in NLP and obtain answers grounded in knowledge retrieved from scientific publications. Our system provides users with comprehensive exploration possibilities, supporting them in investigating the relationships between different fields, understanding unfamiliar concepts in NLP, and finding relevant research literature. Demo, video, and code are available at: https://github.com/NLP-Knowledge-Graph/NLP-KG-WebApp.
Authors: Ziyan Jiang, Xueguang Ma, Wenhu Chen
Abstract: In traditional RAG framework, the basic retrieval units are normally short. The common retrievers like DPR normally work with 100-word Wikipedia paragraphs. Such a design forces the retriever to search over a large corpus to find the `needle' unit. In contrast, the readers only need to extract answers from the short retrieved units. Such an imbalanced `heavy' retriever and `light' reader design can lead to sub-optimal performance. In order to alleviate the imbalance, we propose a new framework LongRAG, consisting of a `long retriever' and a `long reader'. LongRAG processes the entire Wikipedia into 4K-token units, which is 30x longer than before. By increasing the unit size, we significantly reduce the total units from 22M to 700K. This significantly lowers the burden of retriever, which leads to a remarkable retrieval score: answer recall@1=71% on NQ (previously 52%) and answer recall@2=72% (previously 47%) on HotpotQA (full-wiki). Then we feed the top-k retrieved units ($\approx$ 30K tokens) to an existing long-context LLM to perform zero-shot answer extraction. Without requiring any training, LongRAG achieves an EM of 62.7% on NQ, which is the best known result. LongRAG also achieves 64.3% on HotpotQA (full-wiki), which is on par of the SoTA model. Our study offers insights into the future roadmap for combining RAG with long-context LLMs.
Authors: Nishant Balepur, Matthew Shu, Alexander Hoyle, Alison Robey, Shi Feng, Seraphina Goldfarb-Tarrant, Jordan Boyd-Graber
Abstract: Keyword mnemonics are memorable explanations that link new terms to simpler keywords. Prior works generate mnemonics for students, but they do not guide models toward mnemonics students prefer and aid learning. We build SMART, a mnemonic generator trained on feedback from real students learning new terms. To train SMART, we first fine-tune LLaMA-2 on a curated set of user-written mnemonics. We then use LLM alignment to enhance SMART: we deploy mnemonics generated by SMART in a flashcard app to find preferences on mnemonics students favor. We gather 2684 preferences from 45 students across two types: expressed (inferred from ratings) and observed (inferred from student learning), yielding three key findings. First, expressed and observed preferences disagree; what students think is helpful does not fully capture what is truly helpful. Second, Bayesian models can synthesize complementary data from multiple preference types into a single effectiveness signal. SMART is tuned via Direct Preference Optimization on this signal, which we show resolves ties and missing labels in the typical method of pairwise comparisons, augmenting data for LLM output quality gains. Third, mnemonic experts assess SMART as matching GPT-4, at much lower deployment costs, showing the utility of capturing diverse student feedback to align LLMs in education.
Authors: Lucie Guillou, Arnaud Sangnier, Nathalie Sznajder
Abstract: We study networks of processes that all execute the same finite protocol and communicate synchronously in two different ways: a process can broadcast one message to all other processes or send it to at most one other process. In both cases, if no process can receive the message, it will still be sent. We establish a precise complexity class for two coverability problems with a parameterised number of processes: the state coverability problem and the configuration coverability problem. It is already known that these problems are Ackermann-hard (but decidable) in the general case. We show that when the protocol is Wait-Only, i.e., it has no state from which a process can send and receive messages, the complexity drops to P and PSPACE, respectively.
Authors: Minqian Liu, Zhiyang Xu, Zihao Lin, Trevor Ashby, Joy Rimchala, Jiaxin Zhang, Lifu Huang
Abstract: Interleaved text-and-image generation has been an intriguing research direction, where the models are required to generate both images and text pieces in an arbitrary order. Despite the emerging advancements in interleaved generation, the progress in its evaluation still significantly lags behind. Existing evaluation benchmarks do not support arbitrarily interleaved images and text for both inputs and outputs, and they only cover a limited number of domains and use cases. Also, current works predominantly use similarity-based metrics which fall short in assessing the quality in open-ended scenarios. To this end, we introduce InterleavedBench, the first benchmark carefully curated for the evaluation of interleaved text-and-image generation. InterleavedBench features a rich array of tasks to cover diverse real-world use cases. In addition, we present InterleavedEval, a strong reference-free metric powered by GPT-4o to deliver accurate and explainable evaluation. We carefully define five essential evaluation aspects for InterleavedEval, including text quality, perceptual quality, image coherence, text-image coherence, and helpfulness, to ensure a comprehensive and fine-grained assessment. Through extensive experiments and rigorous human evaluation, we show that our benchmark and metric can effectively evaluate the existing models with a strong correlation with human judgments surpassing previous reference-based metrics. We also provide substantial findings and insights to foster future research in interleaved generation and its evaluation.
Authors: Jiarui Feng, Hao Liu, Lecheng Kong, Yixin Chen, Muhan Zhang
Abstract: In this report, we present TAGLAS, an atlas of text-attributed graph (TAG) datasets and benchmarks. TAGs are graphs with node and edge features represented in text, which have recently gained wide applicability in training graph-language or graph foundation models. In TAGLAS, we collect and integrate more than 23 TAG datasets with domains ranging from citation graphs to molecule graphs and tasks from node classification to graph question-answering. Unlike previous graph datasets and benchmarks, all datasets in TAGLAS have a unified node and edge text feature format, which allows a graph model to be simultaneously trained and evaluated on multiple datasets from various domains. Further, we provide a standardized, efficient, and simplified way to load all datasets and tasks. We also provide useful utils like text-to-embedding conversion, and graph-to-text conversion, which can facilitate different evaluation scenarios. Finally, we also provide standard and easy-to-use evaluation utils. The project is open-sourced at https://github.com/JiaruiFeng/TAGLAS and is still under construction. Please expect more datasets/features in the future.
Authors: Murali Karthick Baskar, Andrew Rosenberg, Bhuvana Ramabhadran, Neeraj Gaur, Zhong Meng
Abstract: In this paper, we focus on addressing the constraints faced when applying LLMs to ASR. Recent works utilize prefixLM-type models, which directly apply speech as a prefix to LLMs for ASR. We have found that optimizing speech prefixes leads to better ASR performance and propose applying RNNT loss to perform speech prefix-tuning. This is a simple approach and does not increase the model complexity or alter the inference pipeline. We also propose language-based soft prompting to further improve with frozen LLMs. Empirical analysis on realtime testset from 10 Indic languages demonstrate that our proposed speech prefix-tuning yields improvements with both frozen and fine-tuned LLMs. Our recognition results on an average of 10 Indics show that the proposed prefix-tuning with RNNT loss results in a 12\% relative improvement in WER over the baseline with a fine-tuned LLM. Our proposed approches with the frozen LLM leads to a 31\% relative improvement over basic soft-prompting prefixLM.
Authors: William Fleshman, Benjamin Van Durme
Abstract: Large language models (LLMs) fine-tuned for text-retrieval have demonstrated state-of-the-art results across several information retrieval (IR) benchmarks. However, supervised training for improving these models requires numerous labeled examples, which are generally unavailable or expensive to acquire. In this work, we explore the effectiveness of extending reverse engineered adaptation to the context of information retrieval (RE-AdaptIR). We use RE-AdaptIR to improve LLM-based IR models using only unlabeled data. We demonstrate improved performance both in training domains as well as zero-shot in domains where the models have seen no queries. We analyze performance changes in various fine-tuning scenarios and offer findings of immediate use to practitioners.
Authors: Steven A. Lehr, Aylin Caliskan, Suneragiri Liyanage, Mahzarin R. Banaji
Abstract: How good a research scientist is ChatGPT? We systematically probed the capabilities of GPT-3.5 and GPT-4 across four central components of the scientific process: as a Research Librarian, Research Ethicist, Data Generator, and Novel Data Predictor, using psychological science as a testing field. In Study 1 (Research Librarian), unlike human researchers, GPT-3.5 and GPT-4 hallucinated, authoritatively generating fictional references 36.0% and 5.4% of the time, respectively, although GPT-4 exhibited an evolving capacity to acknowledge its fictions. In Study 2 (Research Ethicist), GPT-4 (though not GPT-3.5) proved capable of detecting violations like p-hacking in fictional research protocols, correcting 88.6% of blatantly presented issues, and 72.6% of subtly presented issues. In Study 3 (Data Generator), both models consistently replicated patterns of cultural bias previously discovered in large language corpora, indicating that ChatGPT can simulate known results, an antecedent to usefulness for both data generation and skills like hypothesis generation. Contrastingly, in Study 4 (Novel Data Predictor), neither model was successful at predicting new results absent in their training data, and neither appeared to leverage substantially new information when predicting more versus less novel outcomes. Together, these results suggest that GPT is a flawed but rapidly improving librarian, a decent research ethicist already, capable of data generation in simple domains with known characteristics but poor at predicting novel patterns of empirical data to aid future experimentation.
Authors: Ivana Kaji\'c, Olivia Wiles, Isabela Albuquerque, Matthias Bauer, Su Wang, Jordi Pont-Tuset, Aida Nematzadeh
Abstract: Text-to-image generative models are capable of producing high-quality images that often faithfully depict concepts described using natural language. In this work, we comprehensively evaluate a range of text-to-image models on numerical reasoning tasks of varying difficulty, and show that even the most advanced models have only rudimentary numerical skills. Specifically, their ability to correctly generate an exact number of objects in an image is limited to small numbers, it is highly dependent on the context the number term appears in, and it deteriorates quickly with each successive number. We also demonstrate that models have poor understanding of linguistic quantifiers (such as "a few" or "as many as"), the concept of zero, and struggle with more advanced concepts such as partial quantities and fractional representations. We bundle prompts, generated images and human annotations into GeckoNum, a novel benchmark for evaluation of numerical reasoning.
Authors: Zackary Rackauckas, Arthur C\^amara, Jakub Zavrel
Abstract: Challenges in the automated evaluation of Retrieval-Augmented Generation (RAG) Question-Answering (QA) systems include hallucination problems in domain-specific knowledge and the lack of gold standard benchmarks for company internal tasks. This results in difficulties in evaluating RAG variations, like RAG-Fusion (RAGF), in the context of a product QA task at Infineon Technologies. To solve these problems, we propose a comprehensive evaluation framework, which leverages Large Language Models (LLMs) to generate large datasets of synthetic queries based on real user queries and in-domain documents, uses LLM-as-a-judge to rate retrieved documents and answers, evaluates the quality of answers, and ranks different variants of Retrieval-Augmented Generation (RAG) agents with RAGElo's automated Elo-based competition. LLM-as-a-judge rating of a random sample of synthetic queries shows a moderate, positive correlation with domain expert scoring in relevance, accuracy, completeness, and precision. While RAGF outperformed RAG in Elo score, a significance analysis against expert annotations also shows that RAGF significantly outperforms RAG in completeness, but underperforms in precision. In addition, Infineon's RAGF assistant demonstrated slightly higher performance in document relevance based on MRR@5 scores. We find that RAGElo positively aligns with the preferences of human annotators, though due caution is still required. Finally, RAGF's approach leads to more complete answers based on expert annotations and better answers overall based on RAGElo's evaluation criteria.
Authors: Mengdan Zhu, Raasikh Kanjiani, Jiahui Lu, Andrew Choi, Qirui Ye, Liang Zhao
Abstract: Deep generative models like VAEs and diffusion models have advanced various generation tasks by leveraging latent variables to learn data distributions and generate high-quality samples. Despite the field of explainable AI making strides in interpreting machine learning models, understanding latent variables in generative models remains challenging. This paper introduces LatentExplainer, a framework for automatically generating semantically meaningful explanations of latent variables in deep generative models. LatentExplainer tackles three main challenges: inferring the meaning of latent variables, aligning explanations with inductive biases, and handling varying degrees of explainability. By perturbing latent variables and interpreting changes in generated data, the framework provides a systematic approach to understanding and controlling the data generation process, enhancing the transparency and interpretability of deep generative models. We evaluate our proposed method on several real-world and synthetic datasets, and the results demonstrate superior performance in generating high-quality explanations of latent variables.
Authors: Kyle Wong, Alfonso Amayuelas, Liangming Pan, William Yang Wang
Abstract: Large language models (LLMs) have shown remarkable performance on code generation tasks. A recent application of LLMs for code generation is iterative code repair, where a model fixes an incorrect program by rationalizing about errors and generating a new program. However, code repair is primarily studied on high-resource languages like Python, and the framework's efficacy is under-explored on low-resource languages. To apply code repair for low-resource languages, we propose Distilling Low-Resource Repairs (DistiLRR), an approach that transfers the reasoning and code generation ability from a teacher model to a student model. Our results show that DistiLRR consistently outperforms baselines on low-resource languages, but has similar performance on high-resource languages. To investigate this behavior, we perform a further analysis and find that the correlation between rationale quality and code correctness is weaker than previously perceived. We hypothesize this weakness is magnified in low-resource settings where base models lack deep knowledge of a programming language, leading to wavering benefits of code repair between high-resource and low-resource languages.
Authors: JiaYing Zheng, HaiNan Zhang, LingXiang Wang, WangJie Qiu, HongWei Zheng, ZhiMing Zheng
Abstract: Private data, being larger and quality-higher than public data, can greatly improve large language models (LLM). However, due to privacy concerns, this data is often dispersed in multiple silos, making its secure utilization for LLM training a challenge. Federated learning (FL) is an ideal solution for training models with distributed private data, but traditional frameworks like FedAvg are unsuitable for LLM due to their high computational demands on clients. An alternative, split learning, offloads most training parameters to the server while training embedding and output layers locally, making it more suitable for LLM. Nonetheless, it faces significant challenges in security and efficiency. Firstly, the gradients of embeddings are prone to attacks, leading to potential reverse engineering of private data. Furthermore, the server's limitation of handle only one client's training request at a time hinders parallel training, severely impacting training efficiency. In this paper, we propose a Federated Learning framework for LLM, named FL-GLM, which prevents data leakage caused by both server-side and peer-client attacks while improving training efficiency. Specifically, we first place the input block and output block on local client to prevent embedding gradient attacks from server. Secondly, we employ key-encryption during client-server communication to prevent reverse engineering attacks from peer-clients. Lastly, we employ optimization methods like client-batching or server-hierarchical, adopting different acceleration methods based on the actual computational capabilities of the server. Experimental results on NLU and generation tasks demonstrate that FL-GLM achieves comparable metrics to centralized chatGLM model, validating the effectiveness of our federated learning framework.
Authors: Tianyu Fu, Haofeng Huang, Xuefei Ning, Genghan Zhang, Boju Chen, Tianqi Wu, Hongyi Wang, Zixiao Huang, Shiyao Li, Shengen Yan, Guohao Dai, Huazhong Yang, Yu Wang
Abstract: Sparse attention can effectively mitigate the significant memory and throughput demands of Large Language Models (LLMs) in long contexts. Existing methods typically employ a uniform sparse attention mask, applying the same sparse pattern across different attention heads and input lengths. However, this uniform approach fails to capture the diverse attention patterns inherent in LLMs, ignoring their distinct accuracy-latency trade-offs. To address this challenge, we propose the Mixture of Attention (MoA), which automatically tailors distinct sparse attention configurations to different heads and layers. MoA constructs and navigates a search space of various attention patterns and their scaling rules relative to input sequence lengths. It profiles the model, evaluates potential configurations, and pinpoints the optimal sparse attention compression plan. MoA adapts to varying input sizes, revealing that some attention heads expand their focus to accommodate longer sequences, while other heads consistently concentrate on fixed-length local contexts. Experiments show that MoA increases the effective context length by $3.9\times$ with the same average attention span, boosting retrieval accuracy by $1.5-7.1\times$ over the uniform-attention baseline across Vicuna-7B, Vicuna-13B, and Llama3-8B models. Moreover, MoA narrows the capability gaps between sparse and dense models, reducing the maximum relative performance drop from $9\%-36\%$ to within $5\%$ across two long-context understanding benchmarks. MoA achieves a $1.2-1.4\times$ GPU memory reduction and boosts decode throughput by $5.5-6.7 \times$ for 7B and 13B dense models on a single GPU, with minimal impact on performance.
Authors: Melvin Wong, Jiao Liu, Thiago Rios, Stefan Menzel, Yew Soon Ong
Abstract: The rapid research and development of generative artificial intelligence has enabled the generation of high-quality images, text, and 3D models from text prompts. This advancement impels an inquiry into whether these models can be leveraged to create digital artifacts for both creative and engineering applications. Drawing on innovative designs from other domains may be one answer to this question, much like the historical practice of ``bionics", where humans have sought inspiration from nature's exemplary designs. This raises the intriguing possibility of using generative models to simultaneously tackle design tasks across multiple domains, facilitating cross-domain learning and resulting in a series of innovative design solutions. In this paper, we propose LLM2FEA as the first attempt to discover novel designs in generative models by transferring knowledge across multiple domains. By utilizing a multi-factorial evolutionary algorithm (MFEA) to drive a large language model, LLM2FEA integrates knowledge from various fields to generate prompts that guide the generative model in discovering novel and practical objects. Experimental results in the context of 3D aerodynamic design verify the discovery capabilities of the proposed LLM2FEA. The designs generated by LLM2FEA not only satisfy practicality requirements to a certain degree but also feature novel and aesthetically pleasing shapes, demonstrating the potential applications of LLM2FEA in discovery tasks.
Authors: Wei Liu, Chenxi Wang, Yifei Wang, Zihao Xie, Rennai Qiu, Yufan Dang, Zhuoyun Du, Weize Chen, Cheng Yang, Chen Qian
Abstract: Large Language Model Multi-Agent Systems (LLM-MAS) have achieved great progress in solving complex tasks. It performs communication among agents within the system to collaboratively solve tasks, under the premise of shared information. However, when agents' communication is leveraged to enhance human cooperation, a new challenge arises due to information asymmetry, since each agent can only access the information of its human user. Previous MAS struggle to complete tasks under this condition. To address this, we propose a new MAS paradigm termed iAgents, which denotes Informative Multi-Agent Systems. In iAgents, the human social network is mirrored in the agent network, where agents proactively exchange human information necessary for task resolution, thereby overcoming information asymmetry. iAgents employs a novel agent reasoning mechanism, InfoNav, to navigate agents' communication towards effective information exchange. Together with InfoNav, iAgents organizes human information in a mixed memory to provide agents with accurate and comprehensive information for exchange. Additionally, we introduce InformativeBench, the first benchmark tailored for evaluating LLM agents' task-solving ability under information asymmetry. Experimental results show that iAgents can collaborate within a social network of 140 individuals and 588 relationships, autonomously communicate over 30 turns, and retrieve information from nearly 70,000 messages to complete tasks within 3 minutes.
Authors: Zixi Zhang, Cheng Zhang, Xitong Gao, Robert D. Mullins, George A. Constantinides, Yiren Zhao
Abstract: Low-rank Adaption (LoRA) has been the de-facto parameter-efficient fine-tuning technique for large language models. We present HeteroLoRA, a light-weight search algorithm that leverages zero-cost proxies to allocate the limited LoRA trainable parameters across the model for better fine-tuned performance. In addition to the allocation for the standard LoRA-adapted models, we also demonstrate the efficacy of HeteroLoRA by performing the allocation in a more challenging search space that includes LoRA modules and LoRA-adapted shortcut connections. Experiments show that HeteroLoRA enables improvements in model performance given the same parameter budge. For example, on MRPC, we see an improvement of 1.6% in accuracy with similar training parameter budget. We will open-source our algorithm once the paper is accepted.
Authors: Manuel Mondal, Ljiljana Dolamic, G\'er\^ome Bovet, Philippe Cudr\'e-Mauroux
Abstract: Prompting and Multiple Choices Questions (MCQ) have become the preferred approach to assess the capabilities of Large Language Models (LLMs), due to their ease of manipulation and evaluation. Such experimental appraisals have pointed toward the LLMs' apparent ability to perform causal reasoning or to grasp uncertainty. In this paper, we investigate whether these abilities are measurable outside of tailored prompting and MCQ by reformulating these issues as direct text completion - the foundation of LLMs. To achieve this goal, we define scenarios with multiple possible outcomes and we compare the prediction made by the LLM through prompting (their Stated Answer) to the probability distributions they compute over these outcomes during next token prediction (their Revealed Belief). Our findings suggest that the Revealed Belief of LLMs significantly differs from their Stated Answer and hint at multiple biases and misrepresentations that their beliefs may yield in many scenarios and outcomes. As text completion is at the core of LLMs, these results suggest that common evaluation methods may only provide a partial picture and that more research is needed to assess the extent and nature of their capabilities.
Authors: Yannis Tevissen
Abstract: Recent advancements in image generation technology have enabled widespread access to AI-generated imagery, prominently used in advertising, entertainment, and progressively in every form of visual content. However, these technologies often perpetuate societal biases. This study investigates the representation biases in popular image generation models towards people with disabilities (PWD). Through a comprehensive experiment involving several popular text-to-image models, we analyzed the depiction of disability. The results indicate a significant bias, with most generated images portraying disabled individuals as old, sad, and predominantly using manual wheelchairs. These findings highlight the urgent need for more inclusive AI development, ensuring diverse and accurate representation of PWD in generated images. This research underscores the importance of addressing and mitigating biases in AI models to foster equitable and realistic representations.
Authors: Fernando De Meer Pardo, Claude Lehmann, Dennis Gehrig, Andrea Nagy, Stefano Nicoli, Branka Hadji Misheva, Martin Braschler, Kurt Stockinger
Abstract: In this paper, we present an end-to-end multi-source Entity Matching problem, which we call entity group matching, where the goal is to assign to the same group, records originating from multiple data sources but representing the same real-world entity. We focus on the effects of transitively matched records, i.e. the records connected by paths in the graph G = (V,E) whose nodes and edges represent the records and whether they are a match or not. We present a real-world instance of this problem, where the challenge is to match records of companies and financial securities originating from different data providers. We also introduce two new multi-source benchmark datasets that present similar matching challenges as real-world records. A distinctive characteristic of these records is that they are regularly updated following real-world events, but updates are not applied uniformly across data sources. This phenomenon makes the matching of certain groups of records only possible through the use of transitive information. In our experiments, we illustrate how considering transitively matched records is challenging since a limited amount of false positive pairwise match predictions can throw off the group assignment of large quantities of records. Thus, we propose GraLMatch, a method that can partially detect and remove false positive pairwise predictions through graph-based properties. Finally, we showcase how fine-tuning a Transformer-based model (DistilBERT) on a reduced number of labeled samples yields a better final entity group matching than training on more samples and/or incorporating fine-tuning optimizations, illustrating how precision becomes the deciding factor in the entity group matching of large volumes of records.
Authors: Francisco de Arriba-P\'erez, Silvia Garc\'ia-M\'endez, F\'atima Leal, Benedita Malheiro, J. C. Burguillo
Abstract: Spam reviews are a pervasive problem on online platforms due to its significant impact on reputation. However, research into spam detection in data streams is scarce. Another concern lies in their need for transparency. Consequently, this paper addresses those problems by proposing an online solution for identifying and explaining spam reviews, incorporating data drift adaptation. It integrates (i) incremental profiling, (ii) data drift detection & adaptation, and (iii) identification of spam reviews employing Machine Learning. The explainable mechanism displays a visual and textual prediction explanation in a dashboard. The best results obtained reached up to 87 % spam F-measure.
Authors: Lin Fan, Xun Gong, Cenyang Zheng, Yafei Ou
Abstract: The intersection of medical Visual Question Answering (Med-VQA) is a challenging research topic with advantages including patient engagement and clinical expert involvement for second opinions. However, existing Med-VQA methods based on joint embedding fail to explain whether their provided results are based on correct reasoning or coincidental answers, which undermines the credibility of VQA answers. In this paper, we investigate the construction of a more cohesive and stable Med-VQA structure. Motivated by causal effect, we propose a novel Triangular Reasoning VQA (Tri-VQA) framework, which constructs reverse causal questions from the perspective of "Why this answer?" to elucidate the source of the answer and stimulate more reasonable forward reasoning processes. We evaluate our method on the Endoscopic Ultrasound (EUS) multi-attribute annotated dataset from five centers, and test it on medical VQA datasets. Experimental results demonstrate the superiority of our approach over existing methods. Our codes and pre-trained models are available at https://anonymous.4open.science/r/Tri_VQA.
Authors: Teo Guichoux, Laure Soulier, Nicolas Obin, Catherine Pelachaud
Abstract: Co-speech gestures play a crucial role in the interactions between humans and embodied conversational agents (ECA). Recent deep learning methods enable the generation of realistic, natural co-speech gestures synchronized with speech, but such approaches require large amounts of training data. "In-the-wild" datasets, which compile videos from sources such as YouTube through human pose detection models, offer a solution by providing 2D skeleton sequences that are paired with speech. Concurrently, innovative lifting models have emerged, capable of transforming these 2D pose sequences into their 3D counterparts, leading to large and diverse datasets of 3D gestures. However, the derived 3D pose estimation is essentially a pseudo-ground truth, with the actual ground truth being the 2D motion data. This distinction raises questions about the impact of gesture representation dimensionality on the quality of generated motions, a topic that, to our knowledge, remains largely unexplored. In this work, we evaluate the impact of the dimensionality of the training data, 2D or 3D joint coordinates, on the performance of a multimodal speech-to-gesture deep generative model. We use a lifting model to convert 2D-generated sequences of body pose to 3D. Then, we compare the sequence of gestures generated directly in 3D to the gestures generated in 2D and lifted to 3D as post-processing.
Authors: Weijia Zhang, Mohammad Aliannejadi, Yifei Yuan, Jiahuan Pei, Jia-Hong Huang, Evangelos Kanoulas
Abstract: Large language models (LLMs) often produce unsupported or unverifiable information, known as "hallucinations." To mitigate this, retrieval-augmented LLMs incorporate citations, grounding the content in verifiable sources. Despite such developments, manually assessing how well a citation supports the associated statement remains a major challenge. Previous studies use faithfulness metrics to estimate citation support automatically but are limited to binary classification, overlooking fine-grained citation support in practical scenarios. To investigate the effectiveness of faithfulness metrics in fine-grained scenarios, we propose a comparative evaluation framework that assesses the metric effectiveness in distinguishinging citations between three-category support levels: full, partial, and no support. Our framework employs correlation analysis, classification evaluation, and retrieval evaluation to measure the alignment between metric scores and human judgments comprehensively. Our results show no single metric consistently excels across all evaluations, revealing the complexity of assessing fine-grained support. Based on the findings, we provide practical recommendations for developing more effective metrics.
Authors: Siyin Wang, Xingsong Ye, Qinyuan Cheng, Junwen Duan, Shimin Li, Jinlan Fu, Xipeng Qiu, Xuanjing Huang
Abstract: As Artificial General Intelligence (AGI) becomes increasingly integrated into various facets of human life, ensuring the safety and ethical alignment of such systems is paramount. Previous studies primarily focus on single-modality threats, which may not suffice given the integrated and complex nature of cross-modality interactions. We introduce a novel safety alignment challenge called Safe Inputs but Unsafe Output (SIUO) to evaluate cross-modality safety alignment. Specifically, it considers cases where single modalities are safe independently but could potentially lead to unsafe or unethical outputs when combined. To empirically investigate this problem, we developed the SIUO, a cross-modality benchmark encompassing 9 critical safety domains, such as self-harm, illegal activities, and privacy violations. Our findings reveal substantial safety vulnerabilities in both closed- and open-source LVLMs, such as GPT-4V and LLaVA, underscoring the inadequacy of current models to reliably interpret and respond to complex, real-world scenarios.
Authors: Jinyin Wang, Haijing Zhang, Yihao Zhong, Yingbin Liang, Rongwei Ji, Yiru Cang
Abstract: Image-text matching is a key multimodal task that aims to model the semantic association between images and text as a matching relationship. With the advent of the multimedia information age, image, and text data show explosive growth, and how to accurately realize the efficient and accurate semantic correspondence between them has become the core issue of common concern in academia and industry. In this study, we delve into the limitations of current multimodal deep learning models in processing image-text pairing tasks. Therefore, we innovatively design an advanced multimodal deep learning architecture, which combines the high-level abstract representation ability of deep neural networks for visual information with the advantages of natural language processing models for text semantic understanding. By introducing a novel cross-modal attention mechanism and hierarchical feature fusion strategy, the model achieves deep fusion and two-way interaction between image and text feature space. In addition, we also optimize the training objectives and loss functions to ensure that the model can better map the potential association structure between images and text during the learning process. Experiments show that compared with existing image-text matching models, the optimized new model has significantly improved performance on a series of benchmark data sets. In addition, the new model also shows excellent generalization and robustness on large and diverse open scenario datasets and can maintain high matching performance even in the face of previously unseen complex situations.
Authors: Weihang Su, Yiran Hu, Anzhe Xie, Qingyao Ai, Zibing Que, Ning Zheng, Yun Liu, Weixing Shen, Yiqun Liu
Abstract: Statute retrieval aims to find relevant statutory articles for specific queries. This process is the basis of a wide range of legal applications such as legal advice, automated judicial decisions, legal document drafting, etc. Existing statute retrieval benchmarks focus on formal and professional queries from sources like bar exams and legal case documents, thereby neglecting non-professional queries from the general public, which often lack precise legal terminology and references. To address this gap, we introduce the STAtute Retrieval Dataset (STARD), a Chinese dataset comprising 1,543 query cases collected from real-world legal consultations and 55,348 candidate statutory articles. Unlike existing statute retrieval datasets, which primarily focus on professional legal queries, STARD captures the complexity and diversity of real queries from the general public. Through a comprehensive evaluation of various retrieval baselines, we reveal that existing retrieval approaches all fall short of these real queries issued by non-professional users. The best method only achieves a Recall@100 of 0.907, suggesting the necessity for further exploration and additional research in this area. All the codes and datasets are available at: https://github.com/oneal2000/STARD/tree/main
Authors: Haoling Li, Xin Zhang, Xiao Liu, Yeyun Gong, Yifan Wang, Yujiu Yang, Qi Chen, Peng Cheng
Abstract: Large language models (LLMs) have revolutionized lots of fields of research. Although it is well-known that fine-tuning is essential for enhancing the capabilities of LLMs, existing research suggests that there is potential redundancy in the fine-tuning process and therefore proposes to update only a subset of parameters. However, these methods fail to leverage the task-specific information to identify important parameters during training. Based on the insight that gradients inherently contain information on task-specific data, we propose Gradient-Mask Tuning (GMT), a method that selectively updates parameters during training based on their gradient information. Specifically, we compute the absolute values of the gradients and apply masking to those with relatively smaller magnitudes. Our empirical results across various tasks demonstrate that GMT not only outperforms traditional fine-tuning methods but also elevates the upper limits of LLM performance. Further analysis indicates that GMT exhibits insensitivity to mask ratio and possesses computational efficiency comparable to vanilla SFT.
Authors: Brandon Huang, Chancharik Mitra, Assaf Arbelle, Leonid Karlinsky, Trevor Darrell, Roei Herzig
Abstract: The recent success of interleaved Large Multimodal Models (LMMs) in few-shot learning suggests that in-context learning (ICL) with many examples can be promising for learning new tasks. However, this many-shot multimodal ICL setting has one crucial problem: it is fundamentally limited by the model's context length set at pretraining. The problem is especially prominent in the multimodal domain, which processes both text and images, requiring additional tokens. This motivates the need for a multimodal method to compress many shots into fewer tokens without finetuning. In this work, we enable LMMs to perform multimodal, many-shot in-context learning by leveraging Multimodal Task Vectors (MTV)--compact implicit representations of in-context examples compressed in the model's attention heads. Specifically, we first demonstrate the existence of such MTV in LMMs and then leverage these extracted MTV to enable many-shot in-context learning for various vision-and-language tasks. Our experiments suggest that MTV can scale in performance with the number of compressed shots and generalize to similar out-of-domain tasks without additional context length for inference.
Authors: Zhenyu Li, Xiuxing Li, Sunqi Fan, Jianyong Wang
Abstract: Structured tabular data is a fundamental data type in numerous fields, and the capacity to reason over tables is crucial for answering questions and validating hypotheses. However, constructing labeled data for complex reasoning tasks is labor intensive, and the quantity of annotated data remains insufficient to support the intricate demands of real-world applications. To address the insufficient annotation challenge, we present a self-training framework for unsupervised complex tabular reasoning (UCTR-ST) by generating diverse synthetic data with complex logic. Specifically, UCTR-ST incorporates several essential techniques: we aggregate diverse programs and execute them on tables based on a "Program-Management" component, and we bridge the gap between programs and text with a powerful "Program-Transformation" module that generates natural language sentences with complex logic. Furthermore, we optimize the procedure using a "Table-Text Manipulator" to handle joint table-text reasoning scenarios. The entire framework utilizes self-training techniques to leverage the unlabeled training data, which results in significant performance improvements when tested on real-world data. Experimental results demonstrate that UCTRST achieves above 90% of the supervised model performance on different tasks and domains, reducing the dependence on manual annotation. Additionally, our approach can serve as a data augmentation technique, significantly boosting the performance of supervised models in low-resourced domains.
Authors: Alfonso Amayuelas, Liangming Pan, Wenhu Chen, William Wang
Abstract: This paper investigates the capabilities of Large Language Models (LLMs) in the context of understanding their knowledge and uncertainty over questions. Specifically, we focus on addressing known-unknown questions, characterized by high uncertainty due to the absence of definitive answers. To facilitate our study, we collect a new dataset with Known-Unknown Questions (KUQ) and establish a categorization framework to clarify the origins of uncertainty in such queries. Subsequently, we examine the performance of open-source LLMs, fine-tuned using this dataset, in distinguishing between known and unknown queries within open-ended question-answering scenarios. The fine-tuned models demonstrated a significant improvement, achieving a considerable increase in F1-score relative to their pre-fine-tuning state. Through a comprehensive analysis, we reveal insights into the models' improved uncertainty articulation and their consequent efficacy in multi-agent debates. These findings help us understand how LLMs can be trained to identify and express uncertainty, improving our knowledge of how they understand and express complex or unclear information.
Authors: Cheng Qian, Chi Han, Yi R. Fung, Yujia Qin, Zhiyuan Liu, Heng Ji
Abstract: Large Language Models (LLMs) have made significant progress in utilizing tools, but their ability is limited by API availability and the instability of implicit reasoning, particularly when both planning and execution are involved. To overcome these limitations, we propose CREATOR, a novel framework that enables LLMs to create their own tools using documentation and code realization. CREATOR disentangles abstract tool creation and concrete decision execution, resulting in improved performance. We evaluate CREATOR on MATH and TabMWP benchmarks, respectively consisting of challenging math competition problems and diverse tabular contents. Remarkably, CREATOR outperforms existing chain-of-thought, program-of-thought, and tool-using baselines. Additionally, we introduce the Creation Challenge dataset, featuring 2K diverse questions, to emphasize the necessity and benefits of LLMs' tool creation ability. Further research demonstrates that leveraging LLMs as tool creators facilitates knowledge transfer, and LLMs exhibit varying levels of tool creation abilities, enabling them to adapt to diverse situations. The tool creation ability revolutionizes the LLM's problem-solving paradigm, driving us closer to the next frontier of artificial intelligence. All the codes and data are released.
Authors: Hao Fei, Meishan Zhang, Min Zhang, Tat-Seng Chua
Abstract: Structured Natural Language Processing (XNLP) is an important subset of NLP that entails understanding the underlying semantic or syntactic structure of texts, which serves as a foundational component for many downstream applications. Despite certain recent efforts to explore universal solutions for specific categories of XNLP tasks, a comprehensive and effective approach for unifying all XNLP tasks long remains underdeveloped. In the meanwhile, while XNLP demonstration systems are vital for researchers exploring various XNLP tasks, existing platforms can be limited to, e.g., supporting few XNLP tasks, lacking interactivity and universalness. To this end, we propose an advanced XNLP demonstration platform, where we propose leveraging LLM to achieve universal XNLP, with one model for all with high generalizability. Overall, our system advances in multiple aspects, including universal XNLP modeling, high performance, interpretability, scalability, and interactivity, providing a unified platform for exploring diverse XNLP tasks in the community. XNLP is online: https://xnlp.haofei.vip
URLs: https://xnlp.haofei.vip
Authors: Xin Zeng, Xiaoyu Wang, Tengxiang Zhang, Chun Yu, Shengdong Zhao, Yiqiang Chen
Abstract: Current gesture interfaces typically demand users to learn and perform gestures from a predefined set, which leads to a less natural experience. Interfaces supporting user-defined gestures eliminate the learning process, but users still need to demonstrate and associate the gesture to a specific system function themselves. We introduce GestureGPT, a free-form hand gesture understanding framework that does not require users to learn, demonstrate, or associate gestures. Our framework leverages the large language model's (LLM) astute common sense and strong inference ability to understand a spontaneously performed gesture from its natural language descriptions, and automatically maps it to a function provided by the interface. More specifically, our triple-agent framework involves a Gesture Description Agent that automatically segments and formulates natural language descriptions of hand poses and movements based on hand landmark coordinates. The description is deciphered by a Gesture Inference Agent through self-reasoning and querying about the interaction context (e.g., interaction history, gaze data), which a Context Management Agent organizes and provides. Following iterative exchanges, the Gesture Inference Agent discerns user intent, grounding it to an interactive function. We validated our conceptual framework under two real-world scenarios: smart home controlling and online video streaming. The average zero-shot Top-5 grounding accuracies are 83.59% for smart home tasks and 73.44% for video streaming. We also provided an extensive discussion of our framework including model selection rationale, generated description quality, generalizability etc.
Authors: Jiale Cheng, Xiao Liu, Kehan Zheng, Pei Ke, Hongning Wang, Yuxiao Dong, Jie Tang, Minlie Huang
Abstract: Large language models (LLMs) have shown impressive success in various applications. However, these models are often not well aligned with human intents, which calls for additional treatments on them; that is, the alignment problem. To make LLMs better follow user instructions, existing alignment methods primarily focus on further training them. However, the extra training of LLMs is usually expensive in terms of GPU computing; even worse, some LLMs are not accessible for user-demanded training, such as GPTs. In this work, we take a different perspective -- Black-Box Prompt Optimization (BPO) -- to perform alignments. The idea is to optimize user prompts to suit LLMs' input understanding, so as to best realize users' intents without updating LLMs' parameters. BPO leverages human preferences to optimize prompts, thus making it superior to LLM (e.g., ChatGPT) as a prompt engineer. Moreover, BPO is model-agnostic, and the empirical results demonstrate that the BPO-aligned ChatGPT yields a 22% increase in the win rate against its original version and 10% for GPT-4. Notably, the BPO-aligned LLMs can outperform the same models aligned by PPO and DPO, and it also brings additional performance gains when combining BPO with PPO or DPO. Code and datasets are released at https://github.com/thu-coai/BPO.
Authors: Sheng Lu, Hendrik Schuff, Iryna Gurevych
Abstract: In-context learning (ICL) has become one of the most popular learning paradigms. While there is a growing body of literature focusing on prompt engineering, there is a lack of systematic analysis comparing the effects of prompts across different models and tasks. To address this gap, we present a comprehensive prompt analysis based on the sensitivity of a function. Our analysis reveals that sensitivity is an unsupervised proxy for model performance, as it exhibits a strong negative correlation with accuracy. We use gradient-based saliency scores to empirically demonstrate how different prompts affect the relevance of input tokens to the output, resulting in different levels of sensitivity. Furthermore, we introduce sensitivity-aware decoding which incorporates sensitivity estimation as a penalty term in the standard greedy decoding. We show that this approach is particularly helpful when information in the input is scarce. Our work provides a fresh perspective on the analysis of prompts, and contributes to a better understanding of the mechanism of ICL.
Authors: Leonardo Ranaldi, Giulia Pucci, Federico Ranaldi, Elena Sofia Ruzzetti, Fabio Massimo Zanzotto
Abstract: Reasoning methods, best exemplified by the well-known Chain-of-Thought (CoT), empower the reasoning abilities of Large Language Models (LLMs) by eliciting them to solve complex tasks in a step-by-step manner. Although they are achieving significant success, the ability to deliver multi-step reasoning remains limited to English because of the imbalance in the distribution of pre-training data, which makes other languages a barrier. In this paper, we propose Cross-lingual Tree-of-Thoughts (Cross-ToT), a method for aligning Cross-lingual CoT reasoning across languages. The proposed method, through a self-consistent cross-lingual prompting mechanism inspired by the Tree-of-Thoughts approach, provides multi-step reasoning paths in different languages that, during the steps, lead to the final solution. Experimental evaluations show that our method significantly outperforms existing prompting methods by reducing the number of interactions and achieving state-of-the-art performance.
Authors: Chenxi Dong
Abstract: This paper proposes a low-code solution to build an AI tutor that leverages advanced AI techniques to provide accurate and contextually relevant responses in a personalized learning environment. The OpenAI Assistants API allows AI Tutor to easily embed, store, retrieve, and manage files and chat history, enabling a low-code solution. Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) technology generate sophisticated answers based on course-specific materials. The application efficiently organizes and retrieves relevant information through vector embedding and similarity-based retrieval algorithms. The AI Tutor prototype demonstrates its ability to generate relevant, accurate answers with source citations. It represents a significant advancement in technology-enhanced tutoring systems, democratizing access to high-quality, customized educational support in higher education.
Authors: Matthew Dahl, Varun Magesh, Mirac Suzgun, Daniel E. Ho
Abstract: Do large language models (LLMs) know the law? These models are increasingly being used to augment legal practice, education, and research, yet their revolutionary potential is threatened by the presence of hallucinations -- textual output that is not consistent with legal facts. We present the first systematic evidence of these hallucinations, documenting LLMs' varying performance across jurisdictions, courts, time periods, and cases. Our work makes four key contributions. First, we develop a typology of legal hallucinations, providing a conceptual framework for future research in this area. Second, we find that legal hallucinations are alarmingly prevalent, occurring between 58% of the time with ChatGPT 4 and 88% with Llama 2, when these models are asked specific, verifiable questions about random federal court cases. Third, we illustrate that LLMs often fail to correct a user's incorrect legal assumptions in a contra-factual question setup. Fourth, we provide evidence that LLMs cannot always predict, or do not always know, when they are producing legal hallucinations. Taken together, our findings caution against the rapid and unsupervised integration of popular LLMs into legal tasks. Even experienced lawyers must remain wary of legal hallucinations, and the risks are highest for those who stand to benefit from LLMs the most -- pro se litigants or those without access to traditional legal resources.
Authors: Li Lucy, Suchin Gururangan, Luca Soldaini, Emma Strubell, David Bamman, Lauren F. Klein, Jesse Dodge
Abstract: Large language models' (LLMs) abilities are drawn from their pretraining data, and model development begins with data curation. However, decisions around what data is retained or removed during this initial stage are under-scrutinized. In our work, we ground web text, which is a popular pretraining data source, to its social and geographic contexts. We create a new dataset of 10.3 million self-descriptions of website creators, and extract information about who they are and where they are from: their topical interests, social roles, and geographic affiliations. Then, we conduct the first study investigating how ten "quality" and English language identification (langID) filters affect webpages that vary along these social dimensions. Our experiments illuminate a range of implicit preferences in data curation: we show that some quality classifiers act like topical domain filters, and langID can overlook English content from some regions of the world. Overall, we hope that our work will encourage a new line of research on pretraining data curation practices and its social implications.
Authors: Minghan Wang, Thuy-Trang Vu, Yuxia Wang, Ehsan Shareghi, Gholamreza Haffari
Abstract: Simultaneous machine translation (SimulMT) presents a challenging trade-off between translation quality and latency. Recent studies have shown that LLMs can achieve good performance in SimulMT tasks. However, this often comes at the expense of high inference cost and latency. In this paper, we propose a conversational SimulMT framework to enhance the inference efficiency of LLM-based SimulMT through multi-turn-dialogue-based decoding. Our experiments with Llama2-7b-chat on two SimulMT benchmarks demonstrate the superiority of LLM in translation quality while achieving comparable computational latency to specialized SimulMT models.
Authors: Siyuan Wang, Zhongyu Wei, Yejin Choi, Xiang Ren
Abstract: Large language models (LLMs) have achieved impressive human-like performance across various reasoning tasks. However, their mastery of underlying inferential rules still falls short of human capabilities. To investigate this, we propose a logic scaffolding inferential rule generation framework, to construct an inferential rule base, ULogic, comprising both primitive and compositional rules across five domains. Our analysis of GPT-series models over a rule subset reveals significant gaps in LLMs' logic understanding compared to human performance, especially in compositional and structural complex rules with certain bias patterns. We further distill these rules into a smaller-scale inference engine for flexible rule generation and enhancing downstream reasoning. Through a multi-judger evaluation, our inference engine proves effective in generating accurate, complex and abstract conclusions and premises, and improve various commonsense reasoning tasks. Overall, our work sheds light on LLMs' limitations in grasping inferential rule and suggests ways to enhance their logical reasoning abilities~\footnote{Code and data are available at \url{https://github.com/SiyuanWangw/ULogic}.}.
Authors: Keyang Xuan, Li Yi, Fan Yang, Ruochen Wu, Yi R. Fung, Heng Ji
Abstract: The rise of multimodal misinformation on social platforms poses significant challenges for individuals and societies. Its increased credibility and broader impact compared to textual misinformation make detection complex, requiring robust reasoning across diverse media types and profound knowledge for accurate verification. The emergence of Large Vision Language Model (LVLM) offers a potential solution to this problem. Leveraging their proficiency in processing visual and textual information, LVLM demonstrates promising capabilities in recognizing complex information and exhibiting strong reasoning skills. In this paper, we first investigate the potential of LVLM on multimodal misinformation detection. We find that even though LVLM has a superior performance compared to LLMs, its profound reasoning may present limited power with a lack of evidence. Based on these observations, we propose LEMMA: LVLM-Enhanced Multimodal Misinformation Detection with External Knowledge Augmentation. LEMMA leverages LVLM intuition and reasoning capabilities while augmenting them with external knowledge to enhance the accuracy of misinformation detection. Our method improves the accuracy over the top baseline LVLM by 7% and 13% on Twitter and Fakeddit datasets respectively.
Authors: Oliver Bentham, Nathan Stringham, Ana Marasovi\'c
Abstract: Understanding the extent to which Chain-of-Thought (CoT) generations align with a large language model's (LLM) internal computations is critical for deciding whether to trust an LLM's output. As a proxy for CoT faithfulness, Lanham et al. (2023) propose a metric that measures a model's dependence on its CoT for producing an answer. Within a single family of proprietary models, they find that LLMs exhibit a scaling-then-inverse-scaling relationship between model size and their measure of faithfulness, and that a 13 billion parameter model exhibits increased faithfulness compared to models ranging from 810 million to 175 billion parameters in size. We evaluate whether these results generalize as a property of all LLMs. We replicate the experimental setup in their section focused on scaling experiments with three different families of models and, under specific conditions, successfully reproduce the scaling trends for CoT faithfulness they report. However, after normalizing the metric to account for a model's bias toward certain answer choices, unfaithfulness drops significantly for smaller less-capable models. This normalized faithfulness metric is also strongly correlated ($R^2$=0.74) with accuracy, raising doubts about its validity for evaluating faithfulness.
Authors: Zhaopeng Feng, Yan Zhang, Hao Li, Bei Wu, Jiayu Liao, Wenqiang Liu, Jun Lang, Yang Feng, Jian Wu, Zuozhu Liu
Abstract: Large Language Models (LLMs) have achieved impressive results in Machine Translation (MT). However, careful evaluations by human reveal that the translations produced by LLMs still contain multiple errors. Importantly, feeding back such error information into the LLMs can lead to self-refinement and result in improved translation performance. Motivated by these insights, we introduce a systematic LLM-based self-refinement translation framework, named \textbf{TEaR}, which stands for \textbf{T}ranslate, \textbf{E}stimate, \textbf{a}nd \textbf{R}efine, marking a significant step forward in this direction. Our findings demonstrate that 1) our self-refinement framework successfully assists LLMs in improving their translation quality across a wide range of languages, whether it's from high-resource languages to low-resource ones or whether it's English-centric or centered around other languages; 2) TEaR exhibits superior systematicity and interpretability; 3) different estimation strategies yield varied impacts, directly affecting the effectiveness of the final corrections. Additionally, traditional neural translation models and evaluation models operate separately, often focusing on singular tasks due to their limited capabilities, while general-purpose LLMs possess the capability to undertake both tasks simultaneously. We further conduct cross-model correction experiments to investigate the potential relationship between the translation and evaluation capabilities of general-purpose LLMs. Our code and data are available at https://github.com/fzp0424/self_correct_mt
Authors: Zhenting Qi, Hanlin Zhang, Eric Xing, Sham Kakade, Himabindu Lakkaraju
Abstract: Retrieval-Augmented Generation (RAG) improves pre-trained models by incorporating external knowledge at test time to enable customized adaptation. We study the risk of datastore leakage in Retrieval-In-Context RAG Language Models (LMs). We show that an adversary can exploit LMs' instruction-following capabilities to easily extract text data verbatim from the datastore of RAG systems built with instruction-tuned LMs via prompt injection. The vulnerability exists for a wide range of modern LMs that span Llama2, Mistral/Mixtral, Vicuna, SOLAR, WizardLM, Qwen1.5, and Platypus2, and the exploitability exacerbates as the model size scales up. Extending our study to production RAG models GPTs, we design an attack that can cause datastore leakage with a 100% success rate on 25 randomly selected customized GPTs with at most 2 queries, and we extract text data verbatim at a rate of 41% from a book of 77,000 words and 3% from a corpus of 1,569,000 words by prompting the GPTs with only 100 queries generated by themselves.
Authors: Zachary Horvitz, Jingru Chen, Rahul Aditya, Harshvardhan Srivastava, Robert West, Zhou Yu, Kathleen McKeown
Abstract: Humor is a fundamental facet of human cognition and interaction. Yet, despite recent advances in natural language processing, humor detection remains a challenging task that is complicated by the scarcity of datasets that pair humorous texts with similar non-humorous counterparts. In our work, we investigate whether large language models (LLMs), can generate synthetic data for humor detection via editing texts. We benchmark LLMs on an existing human dataset and show that current LLMs display an impressive ability to 'unfun' jokes, as judged by humans and as measured on the downstream task of humor detection. We extend our approach to a code-mixed English-Hindi humor dataset, where we find that GPT-4's synthetic data is highly rated by bilingual annotators and provides challenging adversarial examples for humor classifiers.
Authors: Makesh Narsimhan Sreedhar, Traian Rebedea, Shaona Ghosh, Jiaqi Zeng, Christopher Parisien
Abstract: Recent advancements in instruction-tuning datasets have predominantly focused on specific tasks like mathematical or logical reasoning. There has been a notable gap in data designed for aligning language models to maintain topic relevance in conversations - a critical aspect for deploying chatbots to production. We introduce the CantTalkAboutThis dataset to help language models remain focused on the subject at hand during task-oriented interactions. It consists of synthetic dialogues on a wide range of conversation topics from different domains. These dialogues are interspersed with distractor turns that intentionally divert the chatbot from the predefined topic. Fine-tuning language models on this dataset helps make them resilient to deviating from the role assigned and improves their ability to maintain topical coherence compared to general-purpose instruction-tuned LLMs like GPT-4-turbo and Mixtral-Instruct. Additionally, preliminary observations suggest that training models on this dataset also enhance their performance on fine-grained instruction following tasks, including safety alignment.
Authors: Zhihao Xu, Ruixuan Huang, Shuai Wang, Xiting Wang
Abstract: Despite careful safety alignment, current large language models (LLMs) remain vulnerable to various attacks. To further unveil the safety risks of LLMs, we introduce a Safety Concept Activation Vector (SCAV) framework, which effectively guides the attacks by accurately interpreting LLMs' safety mechanisms. We then develop an SCAV-guided attack method that can generate both attack prompts and embedding-level attacks with automatically selected perturbation hyperparameters. Both automatic and human evaluations demonstrate that our attack method significantly improves the attack success rate and response quality while requiring less training data. Additionally, we find that our generated attack prompts may be transferable to GPT-4, and the embedding-level attacks may also be transferred to other white-box LLMs whose parameters are known. Our experiments further uncover the safety risks present in current LLMs. For example, we find that six out of seven open-source LLMs that we attack consistently provide relevant answers to more than 85\% malicious instructions. Finally, we provide insights into the safety mechanism of LLMs.
Authors: Jincheng Dai, Zhuowei Huang, Haiyun Jiang, Chen Chen, Deng Cai, Wei Bi, Shuming Shi
Abstract: Large Language Models (LLMs), despite their remarkable performance across a wide range of tasks, necessitate substantial GPU memory and consume significant computational resources. Beyond the memory taken up by model weights, the memory used by the KV cache rises linearly with sequence length, becoming a primary bottleneck for inference. In this paper, we introduce an innovative method for optimizing the KV cache, which considerably minimizes its memory footprint. Upon thorough investigation, we discover that in most Transformer models, (i) there is a striking similarity between adjacent tokens' query vectors, and (ii) the attention calculation of the current query can rely exclusively on the attention information of a small fraction of preceding queries. Based on these observations, we present CORM, a KV cache eviction policy that dynamically retains essential key-value pairs for inference without the need for model fine-tuning. Our validation shows that CORM reduces the inference memory usage of KV cache by up to 70\% with negligible performance degradation across six tasks in LongBench. Furthermore, we demonstrate that CORM is compatible with GQA for further compression rate.
Authors: Lotem Golany, Filippo Galgani, Maya Mamo, Nimrod Parasol, Omer Vandsburger, Nadav Bar, Ido Dagan
Abstract: Automating data generation with Large Language Models (LLMs) has become increasingly popular. In this work, we investigate the feasibility and effectiveness of LLM-based data generation in the challenging setting of source-grounded information-seeking dialogs, with response attribution, over long documents. Our source texts consist of long and noisy meeting transcripts, adding to the task complexity. Since automating attribution remains difficult, we propose a semi-automatic approach: dialog queries and responses are generated with LLMs, followed by human verification and identification of attribution spans. Using this approach, we created MISeD -- Meeting Information Seeking Dialogs dataset -- a dataset of information-seeking dialogs focused on meeting transcripts. Models finetuned with MISeD demonstrate superior performance compared to off-the-shelf models, even those of larger size. Finetuning on MISeD gives comparable response generation quality to finetuning on fully manual data, while improving attribution quality and reducing time and effort.
Authors: Everlyn Asiko Chimoto, Jay Gala, Orevaoghene Ahia, Julia Kreutzer, Bruce A. Bassett, Sara Hooker
Abstract: Neural Machine Translation models are extremely data and compute-hungry. However, not all data points contribute equally to model training and generalization. Data pruning to remove the low-value data points has the benefit of drastically reducing the compute budget without significant drop in model performance. In this paper, we propose a new data pruning technique: Checkpoints Across Time (CAT), that leverages early model training dynamics to identify the most relevant data points for model performance. We benchmark CAT against several data pruning techniques including COMET-QE, LASER and LaBSE. We find that CAT outperforms the benchmarks on Indo-European languages on multiple test sets. When applied to English-German, English-French and English-Swahili translation tasks, CAT achieves comparable performance to using the full dataset, while pruning up to 50% of training data. We inspect the data points that CAT selects and find that it tends to favour longer sentences and sentences with unique or rare words.
Authors: Kaixuan Huang, Xudong Guo, Mengdi Wang
Abstract: Speculative decoding reduces the inference latency of a target large language model via utilizing a smaller and faster draft model. Its performance depends on a hyperparameter K -- the candidate length, i.e., the number of candidate tokens for the target model to verify in each round. However, previous methods often use simple heuristics to choose K, which may result in sub-optimal performance. We study the choice of the candidate length K and formulate it as a Markov Decision Process. We theoretically show that the optimal policy of this Markov decision process takes the form of a threshold policy, i.e., the current speculation should stop and be verified when the probability of getting a rejection exceeds a threshold value. Motivated by this theory, we propose SpecDec++, an enhanced version of speculative decoding that adaptively determines the candidate length on the fly. We augment the draft model with a trained acceptance prediction head to predict the conditional acceptance probability of the candidate tokens. SpecDec++ will stop the current speculation when the predicted probability that at least one token gets rejected exceeds a threshold. We implement SpecDec++ and apply it to the llama-2-chat 7B & 70B model pair. Our adaptive method achieves a 2.04x speedup on the Alpaca dataset (an additional 7.2% improvement over the baseline speculative decoding). On the GSM8K and HumanEval datasets, our method achieves a 2.26x speedup (9.4% improvement) and 2.23x speedup (11.1% improvement), respectively.
Authors: Xuyang Wu, Zhiyuan Peng, Krishna Sravanthi Rajanala Sai, Hsin-Tai Wu, Yi Fang
Abstract: Effective passage retrieval and reranking methods have been widely utilized to identify suitable candidates in open-domain question answering tasks, recent studies have resorted to LLMs for reranking the retrieved passages by the log-likelihood of the question conditioned on each passage. Although these methods have demonstrated promising results, the performance is notably sensitive to the human-written prompt (or hard prompt), and fine-tuning LLMs can be computationally intensive and time-consuming. Furthermore, this approach limits the leverage of question-passage relevance pairs and passage-specific knowledge to enhance the ranking capabilities of LLMs. In this paper, we propose passage-specific prompt tuning for reranking in open-domain question answering (PSPT): a parameter-efficient method that fine-tunes learnable passage-specific soft prompts, incorporating passage-specific knowledge from a limited set of question-passage relevance pairs. The method involves ranking retrieved passages based on the log-likelihood of the model generating the question conditioned on each passage and the learned soft prompt. We conducted extensive experiments utilizing the Llama-2-chat-7B model across three publicly available open-domain question answering datasets and the results demonstrate the effectiveness of the proposed approach.
Authors: Ebrahim Feghhi, Nima Hadidi, Bryan Song, Idan A. Blank, Jonathan C. Kao
Abstract: Given the remarkable capabilities of large language models (LLMs), there has been a growing interest in evaluating their similarity to the human brain. One approach towards quantifying this similarity is by measuring how well a model predicts neural signals, also called "brain score". Internal representations from LLMs achieve state-of-the-art brain scores, leading to speculation that they share computational principles with human language processing. This inference is only valid if the subset of neural activity predicted by LLMs reflects core elements of language processing. Here, we question this assumption by analyzing three neural datasets used in an impactful study on LLM-to-brain mappings, with a particular focus on an fMRI dataset where participants read short passages. We first find that when using shuffled train-test splits, as done in previous studies with these datasets, a trivial feature that encodes temporal autocorrelation not only outperforms LLMs but also accounts for the majority of neural variance that LLMs explain. We therefore use contiguous splits moving forward. Second, we explain the surprisingly high brain scores of untrained LLMs by showing they do not account for additional neural variance beyond two simple features: sentence length and sentence position. This undermines evidence used to claim that the transformer architecture biases computations to be more brain-like. Third, we find that brain scores of trained LLMs on this dataset can largely be explained by sentence length, position, and pronoun-dereferenced static word embeddings; a small, additional amount is explained by sense-specific embeddings and contextual representations of sentence structure. We conclude that over-reliance on brain scores can lead to over-interpretations of similarity between LLMs and brains, and emphasize the importance of deconstructing what LLMs are mapping to in neural signals.
Authors: Yifan Tang, Yihao Wang, Ru Zhang, Jianyi Liu
Abstract: To detect stego (steganographic text) in complex scenarios, linguistic steganalysis (LS) with various motivations has been proposed and achieved excellent performance. However, with the development of generative steganography, some stegos have strong concealment, especially after the emergence of LLMs-based steganography, the existing LS has low detection or cannot detect them. We designed a novel LS with two modes called LSGC. In the generation mode, we created an LS-task "description" and used the generation ability of LLM to explain whether texts to be detected are stegos. On this basis, we rethought the principle of LS and LLMs, and proposed the classification mode. In this mode, LSGC deleted the LS-task "description" and used the "causalLM" LLMs to extract steganographic features. The LS features can be extracted by only one pass of the model, and a linear layer with initialization weights is added to obtain the classification probability. Experiments on strongly concealed stegos show that LSGC significantly improves detection and reaches SOTA performance. Additionally, LSGC in classification mode greatly reduces training time while maintaining high performance.
Authors: Rajiv Movva, Pang Wei Koh, Emma Pierson
Abstract: To what extent to do LLMs align with human perceptions of safety? We study this question via *annotation alignment*, the extent to which LLMs and humans agree when annotating the safety of user-chatbot conversations. We leverage the recent DICES dataset (Aroyo et al., 2023), in which 350 conversations are each rated for safety by 112 annotators spanning 10 race-gender groups. GPT-4 achieves a Pearson correlation of $r = 0.59$ with the average annotator rating, higher than the median annotator's correlation with the average ($r=0.51$). We show that larger datasets are needed to resolve whether GPT-4 exhibits disparities in how well it correlates with demographic groups. Also, there is substantial idiosyncratic variation in correlation *within* groups, suggesting that race & gender do not fully capture differences in alignment. Finally, we find that GPT-4 cannot predict when one demographic group finds a conversation more unsafe than another.
Authors: Andong Chen, Lianzhang Lou, Kehai Chen, Xuefeng Bai, Yang Xiang, Muyun Yang, Tiejun Zhao, Min Zhang
Abstract: Recently, large language models (LLMs) enhanced by self-reflection have achieved promising performance on machine translation. The key idea is guiding LLMs to generate translation with human-like feedback. However, existing self-reflection methods lack effective feedback information, limiting the translation performance. To address this, we introduce a DUAL-REFLECT framework, leveraging the dual learning of translation tasks to provide effective feedback, thereby enhancing the models' self-reflective abilities and improving translation performance. The application of this method across various translation tasks has proven its effectiveness in improving translation accuracy and eliminating ambiguities, especially in translation tasks with low-resource language pairs.
Authors: David Wadden, Kejian Shi, Jacob Morrison, Aakanksha Naik, Shruti Singh, Nitzan Barzilay, Kyle Lo, Tom Hope, Luca Soldaini, Shannon Zejiang Shen, Doug Downey, Hannaneh Hajishirzi, Arman Cohan
Abstract: We present SciRIFF (Scientific Resource for Instruction-Following and Finetuning), a dataset of 137K instruction-following demonstrations for 54 tasks covering five essential scientific literature understanding capabilities: information extraction, summarization, question answering, claim verification, and classification. SciRIFF demonstrations are notable for their long input contexts, detailed task specifications, and complex structured outputs. While instruction-following resources are available in specific domains such as clinical medicine and chemistry, SciRIFF is the first dataset focused on extracting and synthesizing information from research literature across a wide range of scientific fields. To demonstrate the utility of SciRIFF, we develop a sample-efficient strategy to adapt a general instruction-following model for science by performing additional finetuning on a mix of general-domain and SciRIFF demonstrations. In evaluations on nine held-out scientific tasks, our model -- called SciTulu -- improves over a strong LLM baseline by 28.1% and 6.5% at the 7B and 70B scales respectively, while maintaining general instruction-following performance within 2% of the baseline. We are optimistic that SciRIFF will facilitate the development and evaluation of LLMs to help researchers navigate the ever-growing body of scientific literature. We release our dataset, model checkpoints, and data processing and evaluation code to enable further research.
Authors: Bowen Cao, Deng Cai, Zhisong Zhang, Yuexian Zou, Wai Lam
Abstract: The performance of large language models (LLMs) is acutely sensitive to the phrasing of prompts, which raises significant concerns about their reliability in real-world scenarios. Existing studies often divide prompts into task-level instructions and case-level inputs and primarily focus on evaluating and improving robustness against variations in tasks-level instructions. However, this setup fails to fully address the diversity of real-world user queries and assumes the existence of task-specific datasets. To address these limitations, we introduce RobustAlpacaEval, a new benchmark that consists of semantically equivalent case-level queries and emphasizes the importance of using the worst prompt performance to gauge the lower bound of model performance. Extensive experiments on RobustAlpacaEval with ChatGPT and six open-source LLMs from the Llama, Mistral, and Gemma families uncover substantial variability in model performance; for instance, a difference of 45.48% between the worst and best performance for the Llama-2-70B-chat model, with its worst performance dipping as low as 9.38%. We further illustrate the difficulty in identifying the worst prompt from both model-agnostic and model-dependent perspectives, emphasizing the absence of a shortcut to characterize the worst prompt. We also attempt to enhance the worst prompt performance using existing prompt engineering and prompt consistency methods, but find that their impact is limited. These findings underscore the need to create more resilient LLMs that can maintain high performance across diverse prompts. Data and code are available at https://github.com/cbwbuaa/On-the-Worst-Prompt- Performance-of-LLMs.
Authors: Samuel S. Sohn, Danrui Li, Sen Zhang, Che-Jui Chang, Mubbasir Kapadia
Abstract: Digital storytelling, essential in entertainment, education, and marketing, faces challenges in production scalability and flexibility. The StoryAgent framework, introduced in this paper, utilizes Large Language Models and generative tools to automate and refine digital storytelling. Employing a top-down story drafting and bottom-up asset generation approach, StoryAgent tackles key issues such as manual intervention, interactive scene orchestration, and narrative consistency. This framework enables efficient production of interactive and consistent narratives across multiple modalities, democratizing content creation and enhancing engagement. Our results demonstrate the framework's capability to produce coherent digital stories without reference videos, marking a significant advancement in automated digital storytelling.
Authors: Jiashuo Wang, Yang Xiao, Yanran Li, Changhe Song, Chunpu Xu, Chenhao Tan, Wenjie Li
Abstract: Although there is a growing belief that LLMs can be used as therapists, exploring LLMs' capabilities and inefficacy, particularly from the client's perspective, is limited. This work focuses on a client-centered assessment of LLM therapists with the involvement of simulated clients, a standard approach in clinical medical education. However, there are two challenges when applying the approach to assess LLM therapists at scale. Ethically, asking humans to frequently mimic clients and exposing them to potentially harmful LLM outputs can be risky and unsafe. Technically, it can be difficult to consistently compare the performances of different LLM therapists interacting with the same client. To this end, we adopt LLMs to simulate clients and propose ClientCAST, a client-centered approach to assessing LLM therapists by client simulation. Specifically, the simulated client is utilized to interact with LLM therapists and complete questionnaires related to the interaction. Based on the questionnaire results, we assess LLM therapists from three client-centered aspects: session outcome, therapeutic alliance, and self-reported feelings. We conduct experiments to examine the reliability of ClientCAST and use it to evaluate LLMs therapists implemented by Claude-3, GPT-3.5, LLaMA3-70B, and Mixtral 8*7B. Codes are released at https://github.com/wangjs9/ClientCAST.
Authors: Abhinav Rao, Monojit Choudhury, Somak Aditya
Abstract: We introduce two paradoxes concerning jailbreak of foundation models: First, it is impossible to construct a perfect jailbreak classifier, and second, a weaker model cannot consistently detect whether a stronger (in a pareto-dominant sense) model is jailbroken or not. We provide formal proofs for these paradoxes and a short case study on Llama and GPT4-o to demonstrate this. We discuss broader theoretical and practical repercussions of these results.
Authors: Roman Koshkin, Katsuhito Sudoh, Satoshi Nakamura
Abstract: The advent of transformers has fueled progress in machine translation. More recently large language models (LLMs) have come to the spotlight thanks to their generality and strong performance in a wide range of language tasks, including translation. Here we show that open-source LLMs perform on par with or better than some state-of-the-art baselines in simultaneous machine translation (SiMT) tasks, zero-shot. We also demonstrate that injection of minimal background information, which is easy with an LLM, brings further performance gains, especially on challenging technical subject-matter. This highlights LLMs' potential for building next generation of massively multilingual, context-aware and terminologically accurate SiMT systems that require no resource-intensive training or fine-tuning.
Authors: Mandar Sharma, Nikhil Muralidhar, Shengzhe Xu, Raquib Bin Yousuf, Naren Ramakrishnan
Abstract: The pretraining-fine-tuning paradigm has been the de facto strategy for transfer learning in modern language modeling. With the understanding that task adaptation in LMs is often a function of parameters shared across tasks, we argue that a more surgical approach to regularization needs to exist for smoother transfer learning. Towards this end, we investigate how the pretraining loss landscape is affected by these task-sensitive parameters through an information-theoretic lens. We then leverage the findings from our investigations to devise a novel approach to dropout for improved model regularization and better downstream generalization. This approach, named guided dropout, is both task & architecture agnostic and adds no computational overhead to the fine-tuning process. Through empirical evaluations, we showcase that our approach to regularization yields consistently better performance, even in scenarios of data paucity, compared to standardized baselines.
Authors: Haiping Huang
Abstract: A good theory of mathematical beauty is more practical than any current observation, as new predictions of physical reality can be verified self-consistently. This belief applies to the current status of understanding deep neural networks including large language models and even the biological intelligence. Toy models provide a metaphor of physical reality, allowing mathematically formulating that reality (i.e., the so-called theory), which can be updated as more conjectures are justified or refuted. One does not need to pack all details into a model, but rather, more abstract models are constructed, as complex systems like brains or deep networks have many sloppy dimensions but much less stiff dimensions that strongly impact macroscopic observables. This kind of bottom-up mechanistic modeling is still promising in the modern era of understanding the natural or artificial intelligence. Here, we shed light on eight challenges in developing theory of intelligence following this theoretical paradigm. Theses challenges are representation learning, generalization, adversarial robustness, continual learning, causal learning, internal model of the brain, next-token prediction, and finally the mechanics of subjective experience.
Authors: Haochen Zhang, Yuyang Dong, Chuan Xiao, Masafumi Oyamada
Abstract: This paper explores the utilization of LLMs for data preprocessing (DP), a crucial step in the data mining pipeline that transforms raw data into a clean format conducive to easy processing. Whereas the use of LLMs has sparked interest in devising universal solutions to DP, recent initiatives in this domain typically rely on GPT APIs, raising inevitable data breach concerns. Unlike these approaches, we consider instruction-tuning local LLMs (7 -- 13B models) as universal DP task solvers that operate on a local, single, and low-priced GPU, ensuring data security and enabling further customization. We select a collection of datasets across four representative DP tasks and construct instruction tuning data using data configuration, knowledge injection, and reasoning data distillation techniques tailored to DP. By tuning Mistral-7B, Llama 3-8B, and OpenOrca-Platypus2-13B, our models, namely, Jellyfish-7B/8B/13B, deliver competitiveness compared to GPT-3.5/4 models and strong generalizability to unseen tasks while barely compromising the base models' abilities in NLP tasks. Meanwhile, Jellyfish offers enhanced reasoning capabilities compared to GPT-3.5. Our models are available at: https://huggingface.co/NECOUDBFM/Jellyfish . Our instruction dataset is available at: https://huggingface.co/datasets/NECOUDBFM/Jellyfish-Instruct .
URLs: https://huggingface.co/NECOUDBFM/Jellyfish, https://huggingface.co/datasets/NECOUDBFM/Jellyfish-Instruct
Authors: Zhengqing Yuan, Zhaoxu Li, Weiran Huang, Yanfang Ye, Lichao Sun
Abstract: In recent years, multimodal large language models (MLLMs) such as GPT-4V have demonstrated remarkable advancements, excelling in a variety of vision-language tasks. Despite their prowess, the closed-source nature and computational demands of such models limit their accessibility and applicability. This study introduces TinyGPT-V, a novel open-source MLLM, designed for efficient training and inference across various vision-language tasks, including image captioning (IC) and visual question answering (VQA). Leveraging a compact yet powerful architecture, TinyGPT-V integrates the Phi-2 language model with pre-trained vision encoders, utilizing a unique mapping module for visual and linguistic information fusion. With a training regimen optimized for small backbones and employing a diverse dataset amalgam, TinyGPT-V requires significantly lower computational resources 24GB for training and as little as 8GB for inference without compromising on performance. Our experiments demonstrate that TinyGPT-V, with its language model 2.8 billion parameters, achieves comparable results in VQA and image inference tasks to its larger counterparts while being uniquely suited for deployment on resource-constrained devices through innovative quantization techniques. This work not only paves the way for more accessible and efficient MLLMs but also underscores the potential of smaller, optimized models in bridging the gap between high performance and computational efficiency in real-world applications. Additionally, this paper introduces a new approach to multimodal large language models using smaller backbones. Our code and training weights are available in the supplementary material.
Authors: Daniela A. Wiepert, Rene L. Utianski, Joseph R. Duffy, John L. Stricker, Leland R. Barnard, David T. Jones, Hugo Botha
Abstract: Accurately extracting clinical information from speech is critical to the diagnosis and treatment of many neurological conditions. As such, there is interest in leveraging AI for automatic, objective assessments of clinical speech to facilitate diagnosis and treatment of speech disorders. We explore transfer learning using foundation models, focusing on the impact of layer selection for the downstream task of predicting pathological speech features. We find that selecting an optimal layer can greatly improve performance (~15.8% increase in balanced accuracy per feature as compared to worst layer, ~13.6% increase as compared to final layer), though the best layer varies by predicted feature and does not always generalize well to unseen data. A learned weighted sum offers comparable performance to the average best layer in-distribution (only ~1.2% lower) and had strong generalization for out-of-distribution data (only 1.5% lower than the average best layer).
Authors: Xinyi Wang, Alfonso Amayuelas, Kexun Zhang, Liangming Pan, Wenhu Chen, William Yang Wang
Abstract: Pre-trained language models (LMs) are able to perform complex reasoning without explicit fine-tuning. To understand how pre-training with a next-token prediction objective contributes to the emergence of such reasoning capability, we propose that we can view an LM as deriving new conclusions by aggregating indirect reasoning paths seen at pre-training time. We found this perspective effective in two important cases of reasoning: logic reasoning with knowledge graphs (KGs) and chain-of-thought (CoT) reasoning. More specifically, we formalize the reasoning paths as random walk paths on the knowledge/reasoning graphs. Analyses of learned LM distributions suggest that a weighted sum of relevant random walk path probabilities is a reasonable way to explain how LMs reason. Experiments and analysis on multiple KG and CoT datasets reveal the effect of training on random walk paths and suggest that augmenting unlabeled random walk reasoning paths can improve real-world multi-step reasoning performance. code: https://github.com/WANGXinyiLinda/LM_random_walk
Authors: Baohao Liao, Christian Herold, Shahram Khadivi, Christof Monz
Abstract: Memory-efficient finetuning of large language models (LLMs) has recently attracted huge attention with the increasing size of LLMs, primarily due to the constraints posed by GPU memory limitations and the effectiveness of these methods compared to full finetuning. Despite the advancements, current strategies for memory-efficient finetuning, such as QLoRA, exhibit inconsistent performance across diverse bit-width quantizations and multifaceted tasks. This inconsistency largely stems from the detrimental impact of the quantization process on preserved knowledge, leading to catastrophic forgetting and undermining the utilization of pretrained models for finetuning purposes. In this work, we introduce a novel quantization framework, ApiQ, designed to restore the lost information from quantization by concurrently initializing the LoRA components and quantizing the weights of LLMs. This approach ensures the maintenance of the original LLM's activation precision while mitigating the error propagation from shallower into deeper layers. Through comprehensive evaluations conducted on a spectrum of language tasks with various LLMs, ApiQ demonstrably minimizes activation error during quantization. Consequently, it consistently achieves superior finetuning results across various bit-widths.
Authors: Junho Kim, Yeon Ju Kim, Yong Man Ro
Abstract: This paper presents a way of enhancing the reliability of Large Multi-modal Models (LMMs) in addressing hallucination, where the models generate cross-modal inconsistent responses. Without additional training, we propose Counterfactual Inception, a novel method that implants counterfactual thinking into LMMs using self-generated counterfactual keywords. Our method is grounded in the concept of counterfactual thinking, a cognitive process where human considers alternative realities, enabling more extensive context exploration. Bridging the human cognition mechanism into LMMs, we aim for the models to engage with and generate responses that span a wider contextual scene understanding, mitigating hallucinatory outputs. We further introduce Plausibility Verification Process (PVP), a simple yet robust keyword constraint that effectively filters out sub-optimal keywords to enable the consistent triggering of counterfactual thinking in the model responses. Comprehensive analyses across various LMMs, including both open-source and proprietary models, corroborate that counterfactual thinking significantly reduces hallucination and helps to broaden contextual understanding based on true visual clues.
Authors: Yingying Gao, Shilei Zhang, Chao Deng, Junlan Feng
Abstract: Pre-trained speech language models such as HuBERT and WavLM leverage unlabeled speech data for self-supervised learning and offer powerful representations for numerous downstream tasks. Despite the success of these models, their high requirements for memory and computing resource hinder their application on resource restricted devices. Therefore, this paper introduces GenDistiller, a novel knowledge distillation framework which generates the hidden representations of the pre-trained teacher model directly by a much smaller student network. The proposed method takes the previous hidden layer as history and implements a layer-by-layer prediction of the teacher model autoregressively. Experiments on SUPERB reveal the advantage of GenDistiller over the baseline distilling method without an autoregressive framework, with 33% fewer parameters, similar time consumption and better performance on most of the SUPERB tasks. Ultimately, the proposed GenDistiller reduces the size of WavLM by 82%.
Authors: Jeffrey Li, Alex Fang, Georgios Smyrnis, Maor Ivgi, Matt Jordan, Samir Gadre, Hritik Bansal, Etash Guha, Sedrick Keh, Kushal Arora, Saurabh Garg, Rui Xin, Niklas Muennighoff, Reinhard Heckel, Jean Mercat, Mayee Chen, Suchin Gururangan, Mitchell Wortsman, Alon Albalak, Yonatan Bitton, Marianna Nezhurina, Amro Abbas, Cheng-Yu Hsieh, Dhruba Ghosh, Josh Gardner, Maciej Kilian, Hanlin Zhang, Rulin Shao, Sarah Pratt, Sunny Sanyal, Gabriel Ilharco, Giannis Daras, Kalyani Marathe, Aaron Gokaslan, Jieyu Zhang, Khyathi Chandu, Thao Nguyen, Igor Vasiljevic, Sham Kakade, Shuran Song, Sujay Sanghavi, Fartash Faghri, Sewoong Oh, Luke Zettlemoyer, Kyle Lo, Alaaeldin El-Nouby, Hadi Pouransari, Alexander Toshev, Stephanie Wang, Dirk Groeneveld, Luca Soldaini, Pang Wei Koh, Jenia Jitsev, Thomas Kollar, Alexandros G. Dimakis, Yair Carmon, Achal Dave, Ludwig Schmidt, Vaishaal Shankar
Abstract: We introduce DataComp for Language Models (DCLM), a testbed for controlled dataset experiments with the goal of improving language models. As part of DCLM, we provide a standardized corpus of 240T tokens extracted from Common Crawl, effective pretraining recipes based on the OpenLM framework, and a broad suite of 53 downstream evaluations. Participants in the DCLM benchmark can experiment with data curation strategies such as deduplication, filtering, and data mixing at model scales ranging from 412M to 7B parameters. As a baseline for DCLM, we conduct extensive experiments and find that model-based filtering is key to assembling a high-quality training set. The resulting dataset, DCLM-Baseline enables training a 7B parameter language model from scratch to 64% 5-shot accuracy on MMLU with 2.6T training tokens. Compared to MAP-Neo, the previous state-of-the-art in open-data language models, DCLM-Baseline represents a 6.6 percentage point improvement on MMLU while being trained with 40% less compute. Our baseline model is also comparable to Mistral-7B-v0.3 and Llama 3 8B on MMLU (63% & 66%), and performs similarly on an average of 53 natural language understanding tasks while being trained with 6.6x less compute than Llama 3 8B. Our results highlight the importance of dataset design for training language models and offer a starting point for further research on data curation.
Authors: Wenbin An, Feng Tian, Sicong Leng, Jiahao Nie, Haonan Lin, QianYing Wang, Guang Dai, Ping Chen, Shijian Lu
Abstract: Despite their great success across various multimodal tasks, Large Vision-Language Models (LVLMs) are facing a prevalent problem with object hallucinations, where the generated textual responses are inconsistent with ground-truth objects in the given image. This paper investigates various LVLMs and pinpoints attention deficiency toward discriminative local image features as one root cause of object hallucinations. Specifically, LVLMs predominantly attend to prompt-independent global image features, while failing to capture prompt-relevant local features, consequently undermining the visual grounding capacity of LVLMs and leading to hallucinations. To this end, we propose Assembly of Global and Local Attention (AGLA), a training-free and plug-and-play approach that mitigates object hallucinations by exploring an ensemble of global features for response generation and local features for visual discrimination simultaneously. Our approach exhibits an image-prompt matching scheme that captures prompt-relevant local features from images, leading to an augmented view of the input image where prompt-relevant content is reserved while irrelevant distractions are masked. With the augmented view, a calibrated decoding distribution can be derived by integrating generative global features from the original image and discriminative local features from the augmented image. Extensive experiments show that AGLA consistently mitigates object hallucinations and enhances general perception capability for LVLMs across various discriminative and generative benchmarks. Our code will be released at https://github.com/Lackel/AGLA.
Authors: Ahmed Abdeen Hamed, Tamer E. Fandy
Abstract: The objective of this research is to introduce a network specialized in predicting drugs that can be repurposed by investigating real-world evidence sources, such as clinical trials and biomedical literature. Specifically, it aims to generate drug combination therapies for complex diseases (e.g., cancer, Alzheimer's). We present a multilayered network medicine approach, empowered by a highly configured ChatGPT prompt engineering system, which is constructed on the fly to extract drug mentions in clinical trials. Additionally, we introduce a novel algorithm that connects real-world evidence with disease-specific signaling pathways (e.g., KEGG database). This sheds light on the repurposability of drugs if they are found to bind with one or more protein constituents of a signaling pathway. To demonstrate, we instantiated the framework for breast cancer and found that, out of 46 breast cancer signaling pathways, the framework identified 38 pathways that were covered by at least two drugs. This evidence signals the potential for combining those drugs. Specifically, the most covered signaling pathway, ID hsa:2064, was covered by 108 drugs, some of which can be combined. Conversely, the signaling pathway ID hsa:1499 was covered by only two drugs, indicating a significant gap for further research. Our network medicine framework, empowered by GenAI, shows promise in identifying drug combinations with a high degree of specificity, knowing the exact signaling pathways and proteins that serve as targets. It is noteworthy that ChatGPT successfully accelerated the process of identifying drug mentions in clinical trials, though further investigations are required to determine the relationships among the drug mentions.
Authors: Alessandro Suglia, Claudio Greco, Katie Baker, Jose L. Part, Ioannis Papaioannou, Arash Eshghi, Ioannis Konstas, Oliver Lemon
Abstract: AI personal assistants deployed via robots or wearables require embodied understanding to collaborate with humans effectively. However, current Vision-Language Models (VLMs) primarily focus on third-person view videos, neglecting the richness of egocentric perceptual experience. To address this gap, we propose three key contributions. First, we introduce the Egocentric Video Understanding Dataset (EVUD) for training VLMs on video captioning and question answering tasks specific to egocentric videos. Second, we present AlanaVLM, a 7B parameter VLM trained using parameter-efficient methods on EVUD. Finally, we evaluate AlanaVLM's capabilities on OpenEQA, a challenging benchmark for embodied video question answering. Our model achieves state-of-the-art performance, outperforming open-source models including strong Socratic models using GPT-4 as a planner by 3.6%. Additionally, we outperform Claude 3 and Gemini Pro Vision 1.0 and showcase competitive results compared to Gemini Pro 1.5 and GPT-4V, even surpassing the latter in spatial reasoning. This research paves the way for building efficient VLMs that can be deployed in robots or wearables, leveraging embodied video understanding to collaborate seamlessly with humans in everyday tasks, contributing to the next generation of Embodied AI.
Authors: Xiaohan Lin, Qingxing Cao, Yinya Huang, Haiming Wang, Jianqiao Lu, Zhengying Liu, Linqi Song, Xiaodan Liang
Abstract: Formal verification (FV) has witnessed growing significance with current emerging program synthesis by the evolving large language models (LLMs). However, current formal verification mainly resorts to symbolic verifiers or hand-craft rules, resulting in limitations for extensive and flexible verification. On the other hand, formal languages for automated theorem proving, such as Isabelle, as another line of rigorous verification, are maintained with comprehensive rules and theorems. In this paper, we propose FVEL, an interactive Formal Verification Environment with LLMs. Specifically, FVEL transforms a given code to be verified into Isabelle, and then conducts verification via neural automated theorem proving with an LLM. The joined paradigm leverages the rigorous yet abundant formulated and organized rules in Isabelle and is also convenient for introducing and adjusting cutting-edge LLMs. To achieve this goal, we extract a large-scale FVELER3. The FVELER dataset includes code dependencies and verification processes that are formulated in Isabelle, containing 758 theories, 29,125 lemmas, and 200,646 proof steps in total with in-depth dependencies. We benchmark FVELER in the FVEL environment by first fine-tuning LLMs with FVELER and then evaluating them on Code2Inv and SV-COMP. The results show that FVEL with FVELER fine-tuned Llama3- 8B solves 17.39% (69 -> 81) more problems, and Mistral-7B 12% (75 -> 84) more problems in SV-COMP. And the proportion of proof errors is reduced. Project page: https://fveler.github.io/.