Authors: Gracjan G\'oral, Alicja Ziarko, Michal Nauman, Maciej Wo{\l}czyk
Abstract: Visual perspective-taking (VPT), the ability to understand the viewpoint of another person, enables individuals to anticipate the actions of other people. For instance, a driver can avoid accidents by assessing what pedestrians see. Humans typically develop this skill in early childhood, but it remains unclear whether the recently emerging Vision Language Models (VLMs) possess such capability. Furthermore, as these models are increasingly deployed in the real world, understanding how they perform nuanced tasks like VPT becomes essential. In this paper, we introduce two manually curated datasets, Isle-Bricks and Isle-Dots for testing VPT skills, and we use it to evaluate 12 commonly used VLMs. Across all models, we observe a significant performance drop when perspective-taking is required. Additionally, we find performance in object detection tasks is poorly correlated with performance on VPT tasks, suggesting that the existing benchmarks might not be sufficient to understand this problem. The code and the dataset will be available at https://sites.google.com/view/perspective-taking
Authors: Shivam Shandilya, Menglin Xia, Supriyo Ghosh, Huiqiang Jiang, Jue Zhang, Qianhui Wu, Victor R\"uhle
Abstract: The increasing prevalence of large language models (LLMs) such as GPT-4 in various applications has led to a surge in the size of prompts required for optimal performance, leading to challenges in computational efficiency. Prompt compression aims to reduce the inference cost by minimizing input tokens without compromising on the task performance. However, existing prompt compression techniques either rely on sub-optimal metrics such as information entropy or model it as a task-agnostic token classification problem that fails to capture task-specific information. To address these issues, we propose a novel and efficient reinforcement learning (RL) based task-aware prompt compression method. To ensure low latency requirements, we leverage existing Transformer encoder-based token classification model while guiding the learning process with task-specific reward signals using lightweight REINFORCE algorithm. We evaluate the performance of our method on three diverse and challenging tasks including text summarization, question answering and code summarization. We demonstrate that our RL-guided compression method improves the task performance by 8% - 260% across these three scenarios over state-of-the-art compression techniques while satisfying the same compression rate and latency requirements.
Authors: Akshaj Kumar Veldanda, Shi-Xiong Zhang, Anirban Das, Supriyo Chakraborty, Stephen Rawls, Sambit Sahu, Milind Naphade
Abstract: Large language models (LLMs) have revolutionized various domains, yet their utility comes with significant challenges related to outdated or problematic knowledge embedded during pretraining. This paper addresses the challenge of modifying LLMs to unlearn problematic and outdated information while efficiently integrating new knowledge without retraining from scratch. Here, we propose LLM Surgery, a framework to efficiently modify LLM behaviour by optimizing a three component objective function that: (1) Performs reverse gradient on unlearning dataset (problematic and outdated information), (2) Performs gradient descent on the update dataset (new and updated information), and (3) Minimizes the KL divergence on the retain dataset (small subset of unchanged text), ensuring alignment between pretrained and modified model outputs. Due to the lack of publicly available datasets specifically tailored for our novel task, we compiled a new dataset and an evaluation benchmark. Using Llama2-7B, we demonstrate that LLM Surgery can achieve significant forgetting on the unlearn set, a 20\% increase in accuracy on the update set, and maintain performance on the retain set.
Authors: Jiarui Zhang
Abstract: In modern commercial systems, including Recommendation, Ranking, and E-Commerce platforms, there is a trend towards improving customer experiences by incorporating Personalization context as input into Large Language Models (LLMs). However, LLMs often struggle to effectively parse and utilize sparse and complex personal context without additional processing or contextual enrichment, underscoring the need for more sophisticated context understanding mechanisms. In this work, we propose Guided Profile Generation (GPG), a general method designed to generate personal profiles in natural language. As is observed, intermediate guided profile generation enables LLMs to summarize, and extract the important, distinctive features from the personal context into concise, descriptive sentences, precisely tailoring their generation more closely to an individual's unique habits and preferences. Our experimental results show that GPG improves LLM's personalization ability across different tasks, for example, it increases 37% accuracy in predicting personal preference compared to directly feeding the LLMs with raw personal context.
Authors: Anindita Kundu, Denilson Barbosa
Abstract: We evaluate the effectiveness of Large Language Models (LLMs) in assessing essay quality, focusing on their alignment with human grading. More precisely, we evaluate ChatGPT and Llama in the Automated Essay Scoring (AES) task, a crucial natural language processing (NLP) application in Education. We consider both zero-shot and few-shot learning and different prompting approaches. We compare the numeric grade provided by the LLMs to human rater-provided scores utilizing the ASAP dataset, a well-known benchmark for the AES task. Our research reveals that both LLMs generally assign lower scores compared to those provided by the human raters; moreover, those scores do not correlate well with those provided by the humans. In particular, ChatGPT tends to be harsher and further misaligned with human evaluations than Llama. We also experiment with a number of essay features commonly used by previous AES methods, related to length, usage of connectives and transition words, and readability metrics, including the number of spelling and grammar mistakes. We find that, generally, none of these features correlates strongly with human or LLM scores. Finally, we report results on Llama 3, which are generally better across the board, as expected. Overall, while LLMs do not seem an adequate replacement for human grading, our results are somewhat encouraging for their use as a tool to assist humans in the grading of written essays in the future.
Authors: Tianqi Liu, Wei Xiong, Jie Ren, Lichang Chen, Junru Wu, Rishabh Joshi, Yang Gao, Jiaming Shen, Zhen Qin, Tianhe Yu, Daniel Sohn, Anastasiia Makarova, Jeremiah Liu, Yuan Liu, Bilal Piot, Abe Ittycheriah, Aviral Kumar, Mohammad Saleh
Abstract: Reward models (RMs) play a pivotal role in aligning large language models (LLMs) with human preferences. However, traditional RM training, which relies on response pairs tied to specific prompts, struggles to disentangle prompt-driven preferences from prompt-independent artifacts, such as response length and format. In this work, we expose a fundamental limitation of current RM training methods, where RMs fail to effectively distinguish between contextual signals and irrelevant artifacts when determining preferences. To address this, we introduce a causal framework that learns preferences independent of these artifacts and propose a novel data augmentation technique designed to eliminate them. Extensive experiments show that our approach successfully filters out undesirable artifacts, yielding a more robust reward model (RRM). Our RRM improves the performance of a pairwise reward model trained on Gemma-2-9b-it, on RewardBench, increasing accuracy from 80.61% to 84.15%. Additionally, we train two DPO policies using both the RM and RRM, demonstrating that the RRM significantly enhances DPO-aligned policies, improving MT-Bench scores from 7.27 to 8.31 and length-controlled win-rates in AlpacaEval-2 from 33.46% to 52.49%.
Authors: Huanxuan Liao, Shizhu He, Yupu Hao, Xiang Li, Yuanzhe Zhang, Kang Liu, Jun Zhao
Abstract: Small Language Models (SLMs) are attracting attention due to the high computational demands and privacy concerns of Large Language Models (LLMs). Some studies fine-tune SLMs using Chains of Thought (CoT) data distilled from LLMs, aiming to enhance their reasoning ability. Furthermore, Some CoT distillation methods introduce external symbolic knowledge into the generation process to improve the limited knowledge memory, reasoning ability and out-of-domain (OOD) generalization of SLMs. However, the introduction of symbolic knowledge increases computational overhead and introduces potential noise. In this paper, we introduce $\textit{SKIntern}$, an innovative approach that empowers SLMs to internalize symbolic knowledge and few-shot examples gradually through a progressive fine-tuning process, guided by a predefined linear decay schedule under curriculum learning. By efficiently internalizing knowledge, $\textit{SKIntern}$ reduces computational overhead and speeds up the reasoning process by focusing solely on the question during inference. It outperforms state-of-the-art baselines by over 5\%, while reducing inference costs (measured in FLOPs) by up to $4\times$ across a wide range of SLMs in both in-domain (ID) and out-of-domain (OOD) tasks. Our code will be available at \url{https://github.com/Xnhyacinth/SKIntern}.
Authors: Lai Wei, Zhen Ying, Muyang He, Yutong Chen, Qian Yang, Yanzhe Hong, Jiaping Lu, Xiaoying Li, Weiran Huang, Ying Chen
Abstract: Diabetes is a chronic disease that poses a significant global health burden, and optimizing diabetes management requires multi-stakeholder collaboration. Large language models (LLMs) have shown promise in various healthcare scenarios, but their effectiveness across a diverse range of diabetes tasks remains unproven. In this study, we introduced a framework to train and validate diabetes-specific LLMs. We first developed a comprehensive data processing pipeline that includes data collection, filtering, augmentation and refinement. This approach contributes to creating a high-quality, diabetes-specific dataset, and several evaluation benchmarks entirely from scratch. Utilizing the collected training dataset, we fine-tuned a diabetes-specific LLM family that demonstrated state-of-the-art proficiency in understanding and processing various diabetes tasks compared to other LLMs. Furthermore, clinical studies showed the potential applications of our models in diabetes care, including providing personalized healthcare, assisting medical education, and streamlining clinical tasks. In conclusion, our study introduced a framework to develop and evaluate a diabetes-specific LLM family, and highlighted its potential to enhance clinical practice and provide personalized, data-driven support for diabetes support when facing different end users. The code is provided via GitHub at https://github.com/waltonfuture/Diabetica.
Authors: Qiaozhi He, Xiaomin Zhuang, Zhihua Wu
Abstract: This paper investigates scaling laws for local SGD in LLM training, a distributed optimization algorithm that facilitates training on loosely connected devices. Through extensive experiments, we show that local SGD achieves competitive results compared to conventional methods, given equivalent model parameters, datasets, and computational resources. Furthermore, we explore the application of local SGD in various practical scenarios, including multi-cluster setups and edge computing environments. Our findings elucidate the necessary conditions for effective multi-cluster LLM training and examine the potential and limitations of leveraging edge computing resources in the LLM training process. This demonstrates its viability as an alternative to single large-cluster training.
Authors: Yuxin Wang, Minghua Ma, Zekun Wang, Jingchang Chen, Huiming Fan, Liping Shan, Qing Yang, Dongliang Xu, Ming Liu, Bing Qin
Abstract: The colossal parameters and computational overhead of Large Language Models (LLMs) challenge their real-world applications. Network pruning, which targets unstructured or structured sparsity by removing redundant parameters, has recently been explored for LLM acceleration. Existing LLM pruning works focus on unstructured pruning, which typically requires special hardware support for a practical speed-up. In contrast, structured pruning can reduce latency on general devices. However, it remains a challenge to perform structured pruning efficiently and maintain performance, especially at high sparsity ratios. To this end, we introduce an efficient structured pruning framework named CFSP, which leverages both Coarse (interblock) and Fine-grained (intrablock) activation information as an importance criterion to guide pruning. The pruning is highly efficient, as it only requires one forward pass to compute feature activations. Specifically, we first allocate the sparsity budget across blocks based on their importance and then retain important weights within each block. In addition, we introduce a recovery fine-tuning strategy that adaptively allocates training overhead based on coarse-grained importance to further improve performance. Experimental results demonstrate that CFSP outperforms existing methods on diverse models across various sparsity budgets. Our code will be available at https://github.com/wyxscir/CFSP.
Authors: Yupu Hao, Pengfei Cao, Zhuoran Jin, Huanxuan Liao, ubo Chen, Kang Liu, Jun Zhao
Abstract: Tool learning enables the Large Language Models (LLMs) to interact with the external environment by invoking tools, enriching the accuracy and capability scope of LLMs. However, previous works predominantly focus on improving model's tool-utilizing accuracy and the ability to generalize to new, unseen tools, excessively forcing LLMs to adjust specific tool-invoking pattern without considering the harm to model's general performance. This deviates from the actual applications and original intention of integrating tools to enhance model. To tackle this problem, we dissect the capability trade-offs by examining the hidden representation changes and the gradient-based importance score of model's components. Based on the analysis result, we propose a Component Importance-based Tool-utilizing ability Injection method (CITI). According to the gradient-based importance score of different components, it alleviates the capability conflicts caused by fine-tuning process by applying distinct training strategies to different components. CITI applies Mixture-Of-LoRA (MOLoRA) for important components. Meanwhile, it fine-tunes the parameters of few components deemed less important in the backbone of the LLM, while keeping other parameters frozen. CITI can effectively enhance the model's tool-utilizing capability without excessively compromising its general performance. Experimental results demonstrate that our approach achieves outstanding performance across a range of evaluation metrics.
Authors: Huanxuan Liao, Shizhu He, Yao Xu, Yuanzhe Zhang, Kang Liu, Jun Zhao
Abstract: In this paper, we propose $\textbf{Ne}$ural-$\textbf{Sy}$mbolic $\textbf{C}$ollaborative $\textbf{D}$istillation ($\textbf{NesyCD}$), a novel knowledge distillation method for learning the complex reasoning abilities of Large Language Models (LLMs, e.g., \textgreater 13B). We argue that complex reasoning tasks are difficult for Small Language Models (SLMs, e.g., $\leq$ 7B), as these tasks demand not only general cognitive abilities but also specialized knowledge, which is often sparse and difficult for these neural-based SLMs to effectively capture. Therefore, NesyCD distills the general capabilities and specialized knowledge in LLMs using different manners. On the one hand, we distill only general abilities from teacher LLMs into the student SLMs of parameterized neural networks. On the other hand, for the specialized abilities and uncommon knowledge of a complex reasoning task, we employ a symbolic knowledge distillation approach to obtain and store the specialized knowledge within a symbolic knowledge base (KB). By decoupling general and specialized capabilities, the proposed NesyCD can achieve superior performance cost-effectively, utilizing smaller models and blending parameterized neural networks with symbolic KB. Moreover, the specialized KB generalizes well and is comprehended and manipulated by humans. Our experiments show that NesyCD significantly boosts SLMs' complex reasoning performance on in-domain (BBH, GSM8K) and out-of-domain (AGIEval, ARC) datasets. Notably, our approach enabled the LLaMA3-8B and Qwen2-7B to surpass GPT-3.5-turbo in performance and come close to matching LLaMA3-70B, despite the latter having nine times more parameters. Our code will be available at https://github.com/Xnhyacinth/NesyCD.
Authors: Yuang Li, Xiaosong Qiao, Xiaofeng Zhao, Huan Zhao, Wei Tang, Min Zhang, Hao Yang
Abstract: Large language models can enhance automatic speech recognition systems through generative error correction. In this paper, we propose Pinyin-enhanced GEC, which leverages Pinyi, the phonetic representation of Mandarin Chinese, as supplementary information to improve Chinese ASR error correction. Our approach only utilizes synthetic errors for training and employs the one-best hypothesis during inference. Additionally, we introduce a multitask training approach involving conversion tasks between Pinyin and text to align their feature spaces. Experiments on the Aishell-1 and the Common Voice datasets demonstrate that our approach consistently outperforms GEC with text-only input. More importantly, we provide intuitive explanations for the effectiveness of PY-GEC and multitask training from two aspects: 1) increased attention weight on Pinyin features; and 2) aligned feature space between Pinyin and text hidden states.
Authors: Yu Wang, Chi Han, Tongtong Wu, Xiaoxin He, Wangchunshu Zhou, Nafis Sadeq, Xiusi Chen, Zexue He, Wei Wang, Gholamreza Haffari, Heng Ji, Julian McAuley
Abstract: Building a human-like system that continuously interacts with complex environments -- whether simulated digital worlds or human society -- presents several key challenges. Central to this is enabling continuous, high-frequency interactions, where the interactions are termed experiences. We refer to this envisioned system as the LifeSpan Cognitive System (LSCS). A critical feature of LSCS is its ability to engage in incremental and rapid updates while retaining and accurately recalling past experiences. We identify two major challenges in achieving this: (1) Abstraction and Experience Merging, and (2) Long-term Retention with Accurate Recall. These properties are essential for storing new experiences, organizing past experiences, and responding to the environment in ways that leverage relevant historical data. Unlike language models with continual learning, which typically rely on large corpora for fine-tuning and focus on improving performance within specific domains or tasks, LSCS must rapidly and incrementally update with new information from its environment at a high frequency. Existing technologies with the potential of solving the above two major challenges can be classified into four classes based on a conceptual metric called Storage Complexity, which measures the relative space required to store past experiences. Each of these four classes of technologies has its own strengths and limitations. Given that none of the existing technologies can achieve LSCS alone, we propose a novel paradigm for LSCS that integrates all four classes of technologies. The new paradigm operates through two core processes: Absorbing Experiences and Generating Responses.
Authors: Florian Boudin, Akiko Aizawa
Abstract: Adapting keyphrase generation models to new domains typically involves few-shot fine-tuning with in-domain labeled data. However, annotating documents with keyphrases is often prohibitively expensive and impractical, requiring expert annotators. This paper presents silk, an unsupervised method designed to address this issue by extracting silver-standard keyphrases from citation contexts to create synthetic labeled data for domain adaptation. Extensive experiments across three distinct domains demonstrate that our method yields high-quality synthetic samples, resulting in significant and consistent improvements in in-domain performance over strong baselines.
Authors: Ximing Wen, Wenjuan Tan, Rosina O. Weber
Abstract: Pretrained transformer-based Language Models (LMs) are well-known for their ability to achieve significant improvement on text classification tasks with their powerful word embeddings, but their black-box nature, which leads to a lack of interpretability, has been a major concern. In this work, we introduce GAProtoNet, a novel white-box Multi-head Graph Attention-based Prototypical Network designed to explain the decisions of text classification models built with LM encoders. In our approach, the input vector and prototypes are regarded as nodes within a graph, and we utilize multi-head graph attention to selectively construct edges between the input node and prototype nodes to learn an interpretable prototypical representation. During inference, the model makes decisions based on a linear combination of activated prototypes weighted by the attention score assigned for each prototype, allowing its choices to be transparently explained by the attention weights and the prototypes projected into the closest matching training examples. Experiments on multiple public datasets show our approach achieves superior results without sacrificing the accuracy of the original black-box LMs. We also compare with four alternative prototypical network variations and our approach achieves the best accuracy and F1 among all. Our case study and visualization of prototype clusters also demonstrate the efficiency in explaining the decisions of black-box models built with LMs.
Authors: Junfeng Jiang, Jiahao Huang, Akiko Aizawa
Abstract: Recent developments in Japanese large language models (LLMs) primarily focus on general domains, with fewer advancements in Japanese biomedical LLMs. One obstacle is the absence of a comprehensive, large-scale benchmark for comparison. Furthermore, the resources for evaluating Japanese biomedical LLMs are insufficient. To advance this field, we propose a new benchmark including eight LLMs across four categories and 20 Japanese biomedical datasets across five tasks. Experimental results indicate that: (1) LLMs with a better understanding of Japanese and richer biomedical knowledge achieve better performance in Japanese biomedical tasks, (2) LLMs that are not mainly designed for Japanese biomedical domains can still perform unexpectedly well, and (3) there is still much room for improving the existing LLMs in certain Japanese biomedical tasks. Moreover, we offer insights that could further enhance development in this field. Our evaluation tools tailored to our benchmark as well as the datasets are publicly available in https://huggingface.co/datasets/Coldog2333/JMedBench to facilitate future research.
Authors: Md Abdur Rahman, Hossain Shahriar, Fan Wu, Alfredo Cuzzocrea
Abstract: Large language models (LLMs) are renowned for their exceptional capabilities, and applying to a wide range of applications. However, this widespread use brings significant vulnerabilities. Also, it is well observed that there are huge gap which lies in the need for effective detection and mitigation strategies against malicious prompt injection attacks in large language models, as current approaches may not adequately address the complexity and evolving nature of these vulnerabilities in real-world applications. Therefore, this work focuses the impact of malicious prompt injection attacks which is one of most dangerous vulnerability on real LLMs applications. It examines to apply various BERT (Bidirectional Encoder Representations from Transformers) like multilingual BERT, DistilBert for classifying malicious prompts from legitimate prompts. Also, we observed how tokenizing the prompt texts and generating embeddings using multilingual BERT contributes to improve the performance of various machine learning methods: Gaussian Naive Bayes, Random Forest, Support Vector Machine, and Logistic Regression. The performance of each model is rigorously analyzed with various parameters to improve the binary classification to discover malicious prompts. Multilingual BERT approach to embed the prompts significantly improved and outperformed the existing works and achieves an outstanding accuracy of 96.55% by Logistic regression. Additionally, we investigated the incorrect predictions of the model to gain insights into its limitations. The findings can guide researchers in tuning various BERT for finding the most suitable model for diverse LLMs vulnerabilities.
Authors: Ariadna Sanchez, Alice Ross, Nina Markl
Abstract: This paper presents a review of 107 research papers relating to speech and sex or gender in ISCA Interspeech publications between 2013 and 2023. We note the scarcity of work on this topic and find that terminology, particularly the word \textit{gender}, is used in ways that are underspecified and often out of step with the prevailing view in social sciences that gender is socially constructed and is a spectrum as opposed to a binary category. We draw attention to the potential problems that this can cause for already marginalised groups, and suggest some questions for researchers to ask themselves when undertaking work on speech and gender.
Authors: David Herel, Vojtech Bartek, Tomas Mikolov
Abstract: Who is the US President? The answer changes depending on when the question is asked. While large language models (LLMs) are evaluated on various reasoning tasks, they often miss a crucial dimension: time. In real-world scenarios, the correctness of answers is frequently tied to temporal context. In this paper, we introduce a novel dataset designed to rigorously test LLMs' ability to handle time-sensitive facts. Our benchmark offers a systematic way to measure how well LLMs align their knowledge with the correct time context, filling a key gap in current evaluation methods and offering a valuable tool for improving real-world applicability in future models.
Authors: Yuyan Chen, Yanghua Xiao
Abstract: Emotion cognition in large language models (LLMs) is crucial for enhancing performance across various applications, such as social media, human-computer interaction, and mental health assessment. We explore the current landscape of research, which primarily revolves around emotion classification, emotionally rich response generation, and Theory of Mind assessments, while acknowledge the challenges like dependency on annotated data and complexity in emotion processing. In this paper, we present a detailed survey of recent progress in LLMs for emotion cognition. We explore key research studies, methodologies, outcomes, and resources, aligning them with Ulric Neisser's cognitive stages. Additionally, we outline potential future directions for research in this evolving field, including unsupervised learning approaches and the development of more complex and interpretable emotion cognition LLMs. We also discuss advanced methods such as contrastive learning used to improve LLMs' emotion cognition capabilities.
Authors: Yuyan Chen, Hao Wang, Songzhou Yan, Sijia Liu, Yueze Li, Yi Zhao, Yanghua Xiao
Abstract: Emotional intelligence in large language models (LLMs) is of great importance in Natural Language Processing. However, the previous research mainly focus on basic sentiment analysis tasks, such as emotion recognition, which is not enough to evaluate LLMs' overall emotional intelligence. Therefore, this paper presents a novel framework named EmotionQueen for evaluating the emotional intelligence of LLMs. The framework includes four distinctive tasks: Key Event Recognition, Mixed Event Recognition, Implicit Emotional Recognition, and Intention Recognition. LLMs are requested to recognize important event or implicit emotions and generate empathetic response. We also design two metrics to evaluate LLMs' capabilities in recognition and response for emotion-related statements. Experiments yield significant conclusions about LLMs' capabilities and limitations in emotion intelligence.
Authors: Sourav Verma
Abstract: Large Language Models (LLMs) showcase remarkable abilities, yet they struggle with limitations such as hallucinations, outdated knowledge, opacity, and inexplicable reasoning. To address these challenges, Retrieval-Augmented Generation (RAG) has proven to be a viable solution, leveraging external databases to improve the consistency and coherence of generated content, especially valuable for complex, knowledge-rich tasks, and facilitates continuous improvement by leveraging domain-specific insights. By combining the intrinsic knowledge of LLMs with the vast, dynamic repositories of external databases, RAG achieves a synergistic effect. However, RAG is not without its limitations, including a limited context window, irrelevant information, and the high processing overhead for extensive contextual data. In this comprehensive work, we explore the evolution of Contextual Compression paradigms, providing an in-depth examination of the field. Finally, we outline the current challenges and suggest potential research and development directions, paving the way for future advancements in this area.
Authors: Mohanna Hoveyda, Arjen P. de Vries, Harrie Oosterhuis, Maarten de Rijke, Faegheh Hasibi
Abstract: In question answering (QA), different questions can be effectively addressed with different answering strategies. Some require a simple lookup, while others need complex, multi-step reasoning to be answered adequately. This observation motivates the development of a dynamic method that adaptively selects the most suitable QA strategy for each question, enabling more efficient and effective systems capable of addressing a broader range of question types. To this aim, we build on recent advances in the orchestration of multiple large language models (LLMs) and formulate adaptive QA as a dynamic orchestration challenge. We define this as a contextual multi-armed bandit problem, where the context is defined by the characteristics of the incoming question and the action space consists of potential communication graph configurations among the LLM agents. We then train a linear upper confidence bound model to learn an optimal mapping between different question types and their corresponding optimal multi-LLM communication graph representation. Our experiments show that the proposed solution is viable for adaptive orchestration of a QA system with multiple modules, as it combines the superior performance of more complex strategies while avoiding their costs when simpler strategies suffice.%
Authors: Ming Wang, Yuanzhong Liu, Xiaoyu Liang, Yijie Huang, Daling Wang, Xiaocui Yang, Sijia Shen, Shi Feng, Xiaoming Zhang, Chaofeng Guan, Yifei Zhang
Abstract: LLMs have demonstrated commendable performance across diverse domains. Nevertheless, formulating high-quality prompts to assist them in their work poses a challenge for non-AI experts. Existing research in prompt engineering suggests somewhat scattered optimization principles and designs empirically dependent prompt optimizers. Unfortunately, these endeavors lack a structural design, incurring high learning costs and it is not conducive to the iterative updating of prompts, especially for non-AI experts. Inspired by structured reusable programming languages, we propose LangGPT, a structural prompt design framework. Furthermore, we introduce Minstrel, a multi-generative agent system with reflection to automate the generation of structural prompts. Experiments and the case study illustrate that structural prompts generated by Minstrel or written manually significantly enhance the performance of LLMs. Furthermore, we analyze the ease of use of structural prompts through a user survey in our online community.
Authors: Anmol Mekala, Vineeth Dorna, Shreya Dubey, Abhishek Lalwani, David Koleczek, Mukund Rungta, Sadid Hasan, Elita Lobo
Abstract: Machine unlearning aims to efficiently eliminate the influence of specific training data, known as the forget set, from the model. However, existing unlearning methods for Large Language Models (LLMs) face a critical challenge: they rely solely on negative feedback to suppress responses related to the forget set, which often results in nonsensical or inconsistent outputs, diminishing model utility and posing potential privacy risks. To address this limitation, we propose a novel approach called Alternate Preference Optimization (AltPO), which combines negative feedback with in-domain positive feedback on the forget set. Additionally, we introduce new evaluation metrics to assess the quality of responses related to the forget set. Extensive experiments show that our approach not only enables effective unlearning but also avoids undesirable model behaviors while maintaining overall model performance.
Authors: Georgios Sidiropoulos, Evangelos Kanoulas
Abstract: Speech-based open-domain question answering (QA over a large corpus of text passages with spoken questions) has emerged as an important task due to the increasing number of users interacting with QA systems via speech interfaces. Passage retrieval is a key task in speech-based open-domain QA. So far, previous works adopted pipelines consisting of an automatic speech recognition (ASR) model that transcribes the spoken question before feeding it to a dense text retriever. Such pipelines have several limitations. The need for an ASR model limits the applicability to low-resource languages and specialized domains with no annotated speech data. Furthermore, the ASR model propagates its errors to the retriever. In this work, we try to alleviate these limitations by proposing an ASR-free, end-to-end trained multimodal dense retriever that can work directly on spoken questions. Our experimental results showed that, on shorter questions, our retriever is a promising alternative to the \textit{ASR and Retriever} pipeline, achieving better retrieval performance in cases where ASR would have mistranscribed important words in the question or have produced a transcription with a high word error rate.
Authors: Ishika Joshi, Ishita Gupta, Adrita Dey, Tapan Parikh
Abstract: Large Language Models (LLMs) are increasingly being used to generate text across various languages, for tasks such as translation, customer support, and education. Despite these advancements, LLMs show notable gender biases in English, which become even more pronounced when generating content in relatively underrepresented languages like Hindi. This study explores implicit gender biases in Hindi text generation and compares them to those in English. We developed Hindi datasets inspired by WinoBias to examine stereotypical patterns in responses from models like GPT-4o and Claude-3 sonnet. Our results reveal a significant gender bias of 87.8% in Hindi, compared to 33.4% in English GPT-4o generation, with Hindi responses frequently relying on gender stereotypes related to occupations, power hierarchies, and social class. This research underscores the variation in gender biases across languages and provides considerations for navigating these biases in generative AI systems.
Authors: Zizheng Lin, Chunkit Chan, Yangqiu Song, Xin Liu
Abstract: Theory-of-Mind (ToM) ability possessed by Large Language Models (LLMs) has been shown to be limited. Most existing methods for improving ToM in LLMs adopt zero-shot prompting, and they face challenges including poor performance in complex ToM reasoning tasks and an inability to handle non-narrative contexts. We propose a zero-shot prompting method named Constrained Chain-of-ToM (CCoToM) that leverages domain knowledge and the causal relations between ToM dimensions to address these limitations. Specifically, CCoToM guides LLMs to construct explicit reasoning chains by first prompting LLMs to infer related ToM dimensions (e.g., belief). Afterward, CCoToM prompts LLMs to infer the queried ToM dimension based on the generated related ToM dimensions and corresponding causal relations. Additionally, CCoToM adaptively imposes constraints on prompts to introduce inductive biases and improve consistency between ToM dimensions. Besides narratives, CCoToM can also handle non-narrative contexts like conversations. Extensive experiments show that CCoToM consistently outperforms previous state-of-the-art methods by large margins across all LLMs and datasets used. We also conduct in-depth analyses to gain deeper insights into CCoToM. We have made our code publicly available.
Authors: Iuliia Thorbecke, Juan Zuluaga-Gomez, Esa\'u Villatoro-Tello, Shashi Kumar, Pradeep Rangappa, Sergio Burdisso, Petr Motlicek, Karthik Pandia, Aravind Ganapathiraju
Abstract: The training of automatic speech recognition (ASR) with little to no supervised data remains an open question. In this work, we demonstrate that streaming Transformer-Transducer (TT) models can be trained from scratch in consumer and accessible GPUs in their entirety with pseudo-labeled (PL) speech from foundational speech models (FSM). This allows training a robust ASR model just in one stage and does not require large data and computational budget compared to the two-step scenario with pre-training and fine-tuning. We perform a comprehensive ablation on different aspects of PL-based streaming TT models such as the impact of (1) shallow fusion of n-gram LMs, (2) contextual biasing with named entities, (3) chunk-wise decoding for low-latency streaming applications, and (4) TT overall performance as the function of the FSM size. Our results demonstrate that TT can be trained from scratch without supervised data, even with very noisy PLs. We validate the proposed framework on 6 languages from CommonVoice and propose multiple heuristics to filter out hallucinated PLs.
Authors: Geyuan Zhang, Xiaofei Zhou, Chuheng Chen
Abstract: Fine-tuning pre-trained language models for downstream tasks has achieved impressive results in NLP. However, fine-tuning all parameters becomes impractical due to the rapidly increasing size of model parameters. To address this, Parameter Efficient Fine-Tuning (PEFT) methods update only a subset of parameters. Most PEFT methods, such as LoRA, use incremental updates, which involve adding learned weight matrix increments to the original parameters. Although effective, these methods face limitations in capturing complex parameter dynamics and do not maintain a strong correlation between the original and updated parameters. To overcome these challenges, we propose the direct Updated Transformation (UT) paradigm, which constructs a transformation directly from the original to the updated parameters. This approach ensures that the correlation between the original and updated parameters is preserved, leveraging the semantic features learned during pre-training. Building on this paradigm, we present the Hadamard Updated Transformation (HUT) method. HUT efficiently updates the original weight matrix using the Hadamard transformation with two low-rank matrices, offering a more expressive and flexible update mechanism. This allows HUT to capture richer parameter features through functional transformations, reducing computational complexity while maintaining or improving model quality. Theoretical analysis and extensive experiments on RoBERTa and GPT-2 validate the effectiveness of HUT. Results show that HUT performs on par with or better than other PEFT methods in terms of model quality, while significantly reducing computational complexity.
Authors: Iuliia Thorbecke, Juan Zuluaga-Gomez, Esa\'u Villatoro-Tello, Andres Carofilis, Shashi Kumar, Petr Motlicek, Karthik Pandia, Aravind Ganapathiraju
Abstract: Despite the recent success of end-to-end models for automatic speech recognition, recognizing special rare and out-of-vocabulary words, as well as fast domain adaptation with text, are still challenging. It often happens that biasing to the special entities leads to a degradation in the overall performance. We propose a light on-the-fly method to improve automatic speech recognition performance by combining a bias list of named entities with a word-level n-gram language model with the shallow fusion approach based on the Aho-Corasick string matching algorithm. The Aho-Corasick algorithm has proved to be more efficient than other methods and allows fast context adaptation. An n-gram language model is introduced as a graph with fail and output arcs, where the arc weights are adapted from the n-gram probabilities. The language model is used as an additional support to keyword biasing when the language model is combined with bias entities in a single context graph to take care of the overall performance. We demonstrate our findings on 4 languages, 2 public and 1 private datasets including performance on named entities and out-of-vocabulary entities. We achieve up to 21.6% relative improvement in the general word error rate with no practical difference in the inverse real-time factor.
Authors: Lorenzo Zangari, Candida M. Greco, Davide Picca, Andrea Tagarelli
Abstract: Moral values have deep roots in early civilizations, codified within norms and laws that regulated societal order and the common good. They play a crucial role in understanding the psychological basis of human behavior and cultural orientation. The Moral Foundation Theory (MFT) is a well-established framework that identifies the core moral foundations underlying the manner in which different cultures shape individual and social lives. Recent advancements in natural language processing, particularly Pre-trained Language Models (PLMs), have enabled the extraction and analysis of moral dimensions from textual data. This survey presents a comprehensive review of MFT-informed PLMs, providing an analysis of moral tendencies in PLMs and their application in the context of the MFT. We also review relevant datasets and lexicons and discuss trends, limitations, and future directions. By providing a structured overview of the intersection between PLMs and MFT, this work bridges moral psychology insights within the realm of PLMs, paving the way for further research and development in creating morally aware AI systems.
Authors: Piotr \.Zelasko, Zhehuai Chen, Mengru Wang, Daniel Galvez, Oleksii Hrinchuk, Shuoyang Ding, Ke Hu, Jagadeesh Balam, Vitaly Lavrukhin, Boris Ginsburg
Abstract: A rising interest in the modality extension of foundation language models warrants discussion on the most effective, and efficient, multimodal training approach. This work focuses on neural machine translation (NMT) and proposes a joint multimodal training regime of Speech-LLM to include automatic speech translation (AST). We investigate two different foundation model architectures, decoder-only GPT and encoder-decoder T5, extended with Canary-1B's speech encoder. To handle joint multimodal training, we propose a novel training framework called EMMeTT. EMMeTT improves training efficiency with the following: balanced sampling across languages, datasets, and modalities; efficient sequential data iteration; and a novel 2D bucketing scheme for multimodal data, complemented by a batch size optimizer (OOMptimizer). We show that a multimodal training consistently helps with both architectures. Moreover, SALM-T5 trained with EMMeTT retains the original NMT capability while outperforming AST baselines on four-language subsets of FLORES and FLEURS. The resultant Multimodal Translation Model produces strong text and speech translation results at the same time.
Authors: Shuting Yang, Zehui Liu, Wolfgang Mayer
Abstract: Recent developments in large language models (LLMs) have led to significant improvements in intelligent dialogue systems'ability to handle complex inquiries. However, current LLMs still exhibit limitations in specialized domain knowledge, particularly in technical fields such as agriculture. To address this problem, we propose ShizishanGPT, an intelligent question answering system for agriculture based on the Retrieval Augmented Generation (RAG) framework and agent architecture. ShizishanGPT consists of five key modules: including a generic GPT-4 based module for answering general questions; a search engine module that compensates for the problem that the large language model's own knowledge cannot be updated in a timely manner; an agricultural knowledge graph module for providing domain facts; a retrieval module which uses RAG to supplement domain knowledge; and an agricultural agent module, which invokes specialized models for crop phenotype prediction, gene expression analysis, and so on. We evaluated the ShizishanGPT using a dataset containing 100 agricultural questions specially designed for this study. The experimental results show that the tool significantly outperforms general LLMs as it provides more accurate and detailed answers due to its modular design and integration of different domain knowledge sources. Our source code, dataset, and model weights are publicly available at https://github.com/Zaiwen/CropGPT.
Authors: Danyang Liu, Mirella Lapata, Frank Keller
Abstract: Characters are important in narratives. They move the plot forward, create emotional connections, and embody the story's themes. Visual storytelling methods focus more on the plot and events relating to it, without building the narrative around specific characters. As a result, the generated stories feel generic, with character mentions being absent, vague, or incorrect. To mitigate these issues, we introduce the new task of character-centric story generation and present the first model capable of predicting visual stories with consistently grounded and coreferent character mentions. Our model is finetuned on a new dataset which we build on top of the widely used VIST benchmark. Specifically, we develop an automated pipeline to enrich VIST with visual and textual character coreference chains. We also propose new evaluation metrics to measure the richness of characters and coreference in stories. Experimental results show that our model generates stories with recurring characters which are consistent and coreferent to larger extent compared to baselines and state-of-the-art systems.
Authors: Parisa Jamadi Khiabani, Arkaitz Zubiaga
Abstract: Stance detection is the task of determining the viewpoint expressed in a text towards a given target. A specific direction within the task focuses on cross-target stance detection, where a model trained on samples pertaining to certain targets is then applied to a new, unseen target. With the increasing need to analyze and mining viewpoints and opinions online, the task has recently seen a significant surge in interest. This review paper examines the advancements in cross-target stance detection over the last decade, highlighting the evolution from basic statistical methods to contemporary neural and LLM-based models. These advancements have led to notable improvements in accuracy and adaptability. Innovative approaches include the use of topic-grouped attention and adversarial learning for zero-shot detection, as well as fine-tuning techniques that enhance model robustness. Additionally, prompt-tuning methods and the integration of external knowledge have further refined model performance. A comprehensive overview of the datasets used for evaluating these models is also provided, offering valuable insights into the progress and challenges in the field. We conclude by highlighting emerging directions of research and by suggesting avenues for future work in the task.
Authors: Haoran Li, Qiang Gao, Hongmei Wu, Li Huang
Abstract: Event Causality Identification (ECI) focuses on extracting causal relations between events in texts. Existing methods for ECI primarily rely on causal features and external knowledge. However, these approaches fall short in two dimensions: (1) causal features between events in a text often lack explicit clues, and (2) external knowledge may introduce bias, while specific problems require tailored analyses. To address these issues, we propose SemDI - a simple and effective Semantic Dependency Inquiry Network for ECI. SemDI captures semantic dependencies within the context using a unified encoder. Then, it utilizes a Cloze Analyzer to generate a fill-in token based on comprehensive context understanding. Finally, this fill-in token is used to inquire about the causal relation between two events. Extensive experiments demonstrate the effectiveness of SemDI, surpassing state-of-the-art methods on three widely used benchmarks. Code is available at https://github.com/hrlics/SemDI.
Authors: Daniele Rege Cambrin, Giuseppe Gallipoli, Irene Benedetto, Luca Cagliero, Paolo Garza
Abstract: Large Language Models (LLMs) have demonstrated impressive performance across various tasks. However, current training approaches combine standard cross-entropy loss with extensive data, human feedback, or ad hoc methods to enhance performance. These solutions are often not scalable or feasible due to their associated costs, complexity, or resource requirements. This study investigates the use of established semantic segmentation loss functions in natural language generation to create a versatile, practical, and scalable solution for fine-tuning different architectures. We evaluate their effectiveness in solving Math Word Problems and question answering across different models of varying sizes. For the analyzed tasks, we found that the traditional Cross-Entropy loss represents a sub-optimal choice, while models trained to minimize alternative (task-dependent) losses, such as Focal or Lov\'asz, achieve a mean improvement of +42% on exact match without requiring additional data or human feedback. These findings suggest a promising pathway for more efficient and accessible training processes.
Authors: Mingmeng Geng, Caixi Chen, Yanru Wu, Dongping Chen, Yao Wan, Pan Zhou
Abstract: Large language models (LLMs) are increasingly impacting human society, particularly in textual information. Based on more than 30,000 papers and 1,000 presentations from machine learning conferences, we examined and compared the words used in writing and speaking, representing the first large-scale investigating study of how LLMs influence the two main modes of verbal communication and expression within the same group of people. Our empirical results show that LLM-style words such as "significant" have been used more frequently in abstracts and oral presentations. The impact on speaking is beginning to emerge and is likely to grow in the future, calling attention to the implicit influence and ripple effect of LLMs on human society.
Authors: Moussa Koulako Bala Doumbouya, Ananjan Nandi, Gabriel Poesia, Davide Ghilardi, Anna Goldie, Federico Bianchi, Dan Jurafsky, Christopher D. Manning
Abstract: The safety of Large Language Models (LLMs) remains a critical concern due to a lack of adequate benchmarks for systematically evaluating their ability to resist generating harmful content. Previous efforts towards automated red teaming involve static or templated sets of illicit requests and adversarial prompts which have limited utility given jailbreak attacks' evolving and composable nature. We propose a novel dynamic benchmark of composable jailbreak attacks to move beyond static datasets and taxonomies of attacks and harms. Our approach consists of three components collectively called h4rm3l: (1) a domain-specific language that formally expresses jailbreak attacks as compositions of parameterized prompt transformation primitives, (2) bandit-based few-shot program synthesis algorithms that generate novel attacks optimized to penetrate the safety filters of a target black box LLM, and (3) open-source automated red-teaming software employing the previous two components. We use h4rm3l to generate a dataset of 2656 successful novel jailbreak attacks targeting 6 state-of-the-art (SOTA) open-source and proprietary LLMs. Several of our synthesized attacks are more effective than previously reported ones, with Attack Success Rates exceeding 90% on SOTA closed language models such as claude-3-haiku and GPT4-o. By generating datasets of jailbreak attacks in a unified formal representation, h4rm3l enables reproducible benchmarking and automated red-teaming, contributes to understanding LLM safety limitations, and supports the development of robust defenses in an increasingly LLM-integrated world. Warning: This paper and related research artifacts contain offensive and potentially disturbing prompts and model-generated content.
Authors: Mingjie Liu, Yun-Da Tsai, Wenfei Zhou, Haoxing Ren
Abstract: Despite the significant progress made in code generation with large language models, challenges persist, especially with hardware description languages such as Verilog. This paper first presents an analysis of fine-tuned LLMs on Verilog coding, with synthetic data from prior methods. We identify two main issues: difficulties in handling non-textual representations (Karnaugh maps, state-transition diagrams and waveforms) and significant variability during training with models randomly making "minor" mistakes. To address these limitations, we enhance data curation by creating correct-by-construction data targeting non-textual representations. Additionally, we introduce an automated framework that generates error reports from various model checkpoints and injects these errors into open-source code to create targeted code repair data. Our fine-tuned Starcoder2-15B outperforms prior state-of-the-art results by 3.8%, 10.9%, 6.6% for pass@1 on VerilogEval-Machine, VerilogEval-Human, and RTLLM.
Authors: James Michels, Ramya Bandarupalli, Amin Ahangar Akbari, Thai Le, Hong Xiao, Jing Li, Erik F. Y. Hom
Abstract: Recent advances in Natural Language Processing (NLP) have ignited interest in developing effective methods for predicting protein-ligand interactions (PLIs) given their relevance to drug discovery and protein engineering efforts and the ever-growing volume of biochemical sequence and structural data available. The parallels between human languages and the "languages" used to represent proteins and ligands have enabled the use of NLP machine learning approaches to advance PLI studies. In this review, we explain where and how such approaches have been applied in the recent literature and discuss useful mechanisms such as long short-term memory, transformers, and attention. We conclude with a discussion of the current limitations of NLP methods for the study of PLIs as well as key challenges that need to be addressed in future work.
Authors: Jason Chuan-Chih Chou, Nahid Alam
Abstract: Since the publication of CLIP, the approach of using InfoNCE loss for contrastive pre-training has become widely popular for bridging two or more modalities. Despite its wide adoption, CLIP's original design choices of L2 normalization and cosine similarity logit have rarely been revisited. We have systematically experimented with alternative geometries and softmax logits for language-image pre-training and identified that variants with intuitive Euclidean geometry, Euclidean CLIP (EuCLIP), match or exceed the performance of CLIP and support hierarchical relationships at least as well as more complicated hyperbolic alternative.
Authors: Zhonghao Shi, Harshvardhan Srivastava, Xuan Shi, Shrikanth Narayanan, Maja J. Matari\'c
Abstract: Accurate automatic speech recognition (ASR) for children is crucial for effective real-time child-AI interaction, especially in educational applications. However, off-the-shelf ASR models primarily pre-trained on adult data tend to generalize poorly to children's speech due to the data domain shift from adults to children. Recent studies have found that supervised fine-tuning on children's speech data can help bridge this domain shift, but human annotations may be impractical to obtain for real-world applications and adaptation at training time can overlook additional domain shifts occurring at test time. We devised a novel ASR pipeline to apply unsupervised test-time adaptation (TTA) methods for child speech recognition, so that ASR models pre-trained on adult speech can be continuously adapted to each child speaker at test time without further human annotations. Our results show that ASR models adapted with TTA methods significantly outperform the unadapted off-the-shelf ASR baselines both on average and statistically across individual child speakers. Our analysis also discovered significant data domain shifts both between child speakers and within each child speaker, which further motivates the need for test-time adaptation.
Authors: Zihan Zhao, Bo Chen, Jingpiao Li, Lu Chen, Liyang Wen, Pengyu Wang, Zichen Zhu, Danyang Zhang, Ziping Wan, Yansi Li, Zhongyang Dai, Xin Chen, Kai Yu
Abstract: Rapid developments of AI tools are expected to offer unprecedented assistance to the research of natural science including chemistry. However, neither existing unimodal task-specific specialist models nor emerging general large multimodal models (LMM) can cover the wide range of chemical data modality and task categories. To address the real demands of chemists, a cross-modal Chemical General Intelligence (CGI) system, which serves as a truly practical and useful research assistant utilizing the great potential of LMMs, is in great need. In this work, we introduce the first Cross-modal Dialogue Foundation Model for Chemistry (ChemDFM-X). Diverse multimodal data are generated from an initial modality by approximate calculations and task-specific model predictions. This strategy creates sufficient chemical training corpora, while significantly reducing excessive expense, resulting in an instruction-tuning dataset containing 7.6M data. After instruction finetuning, ChemDFM-X is evaluated on extensive experiments of different chemical tasks with various data modalities. The results demonstrate the capacity of ChemDFM-X for multimodal and inter-modal knowledge comprehension. ChemDFM-X marks a significant milestone toward aligning all modalities in chemistry, a step closer to CGI.
Authors: Yinmin Zhong, Zili Zhang, Bingyang Wu, Shengyu Liu, Yukun Chen, Changyi Wan, Hanpeng Hu, Lei Xia, Ranchen Ming, Yibo Zhu, Xin Jin
Abstract: Reinforcement Learning from Human Feedback (RLHF) stands as a pivotal post-training technique to enhance the alignment between LLMs and human preference. The workflow of RLHF typically involves several models and tasks in a series of distinct stages. Existing RLHF training systems view each task as the smallest execution unit thus overlooking the opportunities for subtask-level optimizations. Due to the intrinsic nature of RLHF training, i.e., the data skewness in the generation stage, and the pipeline bubbles in the training stage, existing RLHF systems suffer from low GPU utilization in production deployments. RLHFuse breaks the traditional view of RLHF workflow as a composition of individual tasks, splitting each task into finer-grained subtasks, and performing stage fusion to improve GPU utilization. RLHFuse contains two key ideas. First, for generation and inference tasks, RLHFuse splits them into sample-level subtasks, enabling efficient inter-stage fusion to mitigate the original generation bottleneck dominated by long-tailed samples. Second, for training tasks, RLHFuse breaks them into subtasks of micro-batches. By leveraging the intuition that pipeline execution can be essentially complemented by another pipeline, RLHFuse performs intra-stage fusion to concurrently execute these subtasks in the training stage with a fused pipeline schedule, resulting in fewer pipeline bubbles. In addition, RLHFuse incorporates a series of system optimizations tailored for each stage of RLHF, making it efficient and scalable for our internal product usage. We evaluate RLHFuse on various popular LLMs and the results show that RLHFuse increases the training throughput by up to 3.7x, compared to existing state-of-the-art systems.
Authors: Jinge Wu, Yunsoo Kim, Daqian Shi, David Cliffton, Fenglin Liu, Honghan Wu
Abstract: Inspired by the success of large language models (LLMs), there is growing research interest in developing LLMs in the medical domain to assist clinicians. However, for hospitals, using closed-source commercial LLMs involves privacy issues, and developing open-source public LLMs requires large-scale computational resources, which are usually limited, especially in resource-efficient regions and low-income countries. We propose an open-source Small Language and Vision Assistant (SLaVA-CXR) that can be used for Chest X-Ray report automation. To efficiently train a small assistant, we first propose the Re$^3$Training method, which simulates the cognitive development of radiologists and optimizes the model in the Recognition, Reasoning, and Reporting training manner. Then, we introduce a data synthesis method, RADEX, which can generate a high-quality and diverse training corpus with privacy regulation compliance. The extensive experiments show that our SLaVA-CXR built on a 2.7B backbone not only outperforms but also achieves 6 times faster inference efficiency than previous state-of-the-art larger models.
Authors: Karthik Valmeekam, Kaya Stechly, Subbarao Kambhampati
Abstract: The ability to plan a course of action that achieves a desired state of affairs has long been considered a core competence of intelligent agents and has been an integral part of AI research since its inception. With the advent of large language models (LLMs), there has been considerable interest in the question of whether or not they possess such planning abilities. PlanBench, an extensible benchmark we developed in 2022, soon after the release of GPT3, has remained an important tool for evaluating the planning abilities of LLMs. Despite the slew of new private and open source LLMs since GPT3, progress on this benchmark has been surprisingly slow. OpenAI claims that their recent o1 (Strawberry) model has been specifically constructed and trained to escape the normal limitations of autoregressive LLMs--making it a new kind of model: a Large Reasoning Model (LRM). Using this development as a catalyst, this paper takes a comprehensive look at how well current LLMs and new LRMs do on PlanBench. As we shall see, while o1's performance is a quantum improvement on the benchmark, outpacing the competition, it is still far from saturating it. This improvement also brings to the fore questions about accuracy, efficiency, and guarantees which must be considered before deploying such systems.
Authors: Dimitrios Panagopoulos, Adoldo Perrusquia, Weisi Guo
Abstract: In recent years, robots and autonomous systems have become increasingly integral to our daily lives, offering solutions to complex problems across various domains. Their application in search and rescue (SAR) operations, however, presents unique challenges. Comprehensively exploring the disaster-stricken area is often infeasible due to the vastness of the terrain, transformed environment, and the time constraints involved. Traditional robotic systems typically operate on predefined search patterns and lack the ability to incorporate and exploit ground truths provided by human stakeholders, which can be the key to speeding up the learning process and enhancing triage. Addressing this gap, we introduce a system that integrates social interaction via large language models (LLMs) with a hierarchical reinforcement learning (HRL) framework. The proposed system is designed to translate verbal inputs from human stakeholders into actionable RL insights and adjust its search strategy. By leveraging human-provided information through LLMs and structuring task execution through HRL, our approach not only bridges the gap between autonomous capabilities and human intelligence but also significantly improves the agent's learning efficiency and decision-making process in environments characterised by long horizons and sparse rewards.
Authors: Matthew Caren, Kartik Chandra, Joshua B. Tenenbaum, Jonathan Ragan-Kelley, Karima Ma
Abstract: We present a method for automatically producing human-like vocal imitations of sounds: the equivalent of "sketching," but for auditory rather than visual representation. Starting with a simulated model of the human vocal tract, we first try generating vocal imitations by tuning the model's control parameters to make the synthesized vocalization match the target sound in terms of perceptually-salient auditory features. Then, to better match human intuitions, we apply a cognitive theory of communication to take into account how human speakers reason strategically about their listeners. Finally, we show through several experiments and user studies that when we add this type of communicative reasoning to our method, it aligns with human intuitions better than matching auditory features alone does. This observation has broad implications for the study of depiction in computer graphics.
Authors: Junjie Huang, Daya Guo, Chenglong Wang, Jiazhen Gu, Shuai Lu, Jeevana Priya Inala, Cong Yan, Jianfeng Gao, Nan Duan, Michael R. Lyu
Abstract: Data wrangling, the process of preparing raw data for further analysis in computational notebooks, is a crucial yet time-consuming step in data science. Code generation has the potential to automate the data wrangling process to reduce analysts' overhead by translating user intents into executable code. Precisely generating data wrangling code necessitates a comprehensive consideration of the rich context present in notebooks, including textual context, code context and data context. However, notebooks often interleave multiple non-linear analysis tasks into linear sequence of code blocks, where the contextual dependencies are not clearly reflected. Directly training models with source code blocks fails to fully exploit the contexts for accurate wrangling code generation. To bridge the gap, we aim to construct a high quality datasets with clear and rich contexts to help training models for data wrangling code generation tasks. In this work, we first propose an automated approach, CoCoMine to mine data-wrangling code generation examples with clear multi-modal contextual dependency. It first adopts data flow analysis to identify the code blocks containing data wrangling codes. Then, CoCoMine extracts the contextualized datawrangling code examples through tracing and replaying notebooks. With CoCoMine, we construct CoCoNote, a dataset containing 58,221 examples for Contextualized Data-wrangling Code generation in Notebooks. To demonstrate the effectiveness of our dataset, we finetune a range of pretrained code models and prompt various large language models on our task. Furthermore, we also propose DataCoder, which encodes data context and code&textual contexts separately to enhance code generation. Experiment results demonstrate the significance of incorporating data context in data-wrangling code generation and the effectiveness of our model. We release code and data at url...
Authors: Junjie Huang, Zhihan Jiang, Jinyang Liu, Yintong Huo, Jiazhen Gu, Zhuangbin Chen, Cong Feng, Hui Dong, Zengyin Yang, Michael R. Lyu
Abstract: Logs are imperative in the maintenance of online service systems, which often encompass important information for effective failure mitigation. While existing anomaly detection methodologies facilitate the identification of anomalous logs within extensive runtime data, manual investigation of log messages by engineers remains essential to comprehend faults, which is labor-intensive and error-prone. Upon examining the log-based troubleshooting practices at CloudA, we find that engineers typically prioritize two categories of log information for diagnosis. These include fault-indicating descriptions, which record abnormal system events, and fault-indicating parameters, which specify the associated entities. Motivated by this finding, we propose an approach to automatically extract such faultindicating information from logs for fault diagnosis, named LoFI. LoFI comprises two key stages. In the first stage, LoFI performs coarse-grained filtering to collect logs related to the faults based on semantic similarity. In the second stage, LoFI leverages a pre-trained language model with a novel prompt-based tuning method to extract fine-grained information of interest from the collected logs. We evaluate LoFI on logs collected from Apache Spark and an industrial dataset from CloudA. The experimental results demonstrate that LoFI outperforms all baseline methods by a significant margin, achieving an absolute improvement of 25.8~37.9 in F1 over the best baseline method, ChatGPT. This highlights the effectiveness of LoFI in recognizing fault-indicating information. Furthermore, the successful deployment of LoFI at CloudA and user studies validate the utility of our method. The code and data are available at https://github.com/Jun-jie-Huang/LoFI.
Authors: Abhilash Nandy, Yash Agarwal, Ashish Patwa, Millon Madhur Das, Aman Bansal, Ankit Raj, Pawan Goyal, Niloy Ganguly
Abstract: Understanding satire and humor is a challenging task for even current Vision-Language models. In this paper, we propose the challenging tasks of Satirical Image Detection (detecting whether an image is satirical), Understanding (generating the reason behind the image being satirical), and Completion (given one half of the image, selecting the other half from 2 given options, such that the complete image is satirical) and release a high-quality dataset YesBut, consisting of 2547 images, 1084 satirical and 1463 non-satirical, containing different artistic styles, to evaluate those tasks. Each satirical image in the dataset depicts a normal scenario, along with a conflicting scenario which is funny or ironic. Despite the success of current Vision-Language Models on multimodal tasks such as Visual QA and Image Captioning, our benchmarking experiments show that such models perform poorly on the proposed tasks on the YesBut Dataset in Zero-Shot Settings w.r.t both automated as well as human evaluation. Additionally, we release a dataset of 119 real, satirical photographs for further research. The dataset and code are available at https://github.com/abhi1nandy2/yesbut_dataset.
Authors: Ting Liu, Zunnan Xu, Yue Hu, Liangtao Shi, Zhiqiang Wang, Quanjun Yin
Abstract: Referring Expression Comprehension (REC), which aims to ground a local visual region via natural language, is a task that heavily relies on multimodal alignment. Most existing methods utilize powerful pre-trained models to transfer visual/linguistic knowledge by full fine-tuning. However, full fine-tuning the entire backbone not only breaks the rich prior knowledge embedded in the pre-training, but also incurs significant computational costs. Motivated by the recent emergence of Parameter-Efficient Transfer Learning (PETL) methods, we aim to solve the REC task in an effective and efficient manner. Directly applying these PETL methods to the REC task is inappropriate, as they lack the specific-domain abilities for precise local visual perception and visual-language alignment. Therefore, we propose a novel framework of Multimodal Prior-guided Parameter Efficient Tuning, namely MaPPER. Specifically, MaPPER comprises Dynamic Prior Adapters guided by a aligned prior, and Local Convolution Adapters to extract precise local semantics for better visual perception. Moreover, the Prior-Guided Text module is proposed to further utilize the prior for facilitating the cross-modal alignment. Experimental results on three widely-used benchmarks demonstrate that MaPPER achieves the best accuracy compared to the full fine-tuning and other PETL methods with only 1.41% tunable backbone parameters.
Authors: Abrar Anwar, John Welsh, Joydeep Biswas, Soha Pouya, Yan Chang
Abstract: Navigating and understanding complex environments over extended periods of time is a significant challenge for robots. People interacting with the robot may want to ask questions like where something happened, when it occurred, or how long ago it took place, which would require the robot to reason over a long history of their deployment. To address this problem, we introduce a Retrieval-augmented Memory for Embodied Robots, or ReMEmbR, a system designed for long-horizon video question answering for robot navigation. To evaluate ReMEmbR, we introduce the NaVQA dataset where we annotate spatial, temporal, and descriptive questions to long-horizon robot navigation videos. ReMEmbR employs a structured approach involving a memory building and a querying phase, leveraging temporal information, spatial information, and images to efficiently handle continuously growing robot histories. Our experiments demonstrate that ReMEmbR outperforms LLM and VLM baselines, allowing ReMEmbR to achieve effective long-horizon reasoning with low latency. Additionally, we deploy ReMEmbR on a robot and show that our approach can handle diverse queries. The dataset, code, videos, and other material can be found at the following link: https://nvidia-ai-iot.github.io/remembr
Authors: Adir Rahamim, Yonatan Belinkov
Abstract: Recent work has compared neural network representations via similarity-based analyses to improve model interpretation. The quality of a similarity measure is typically evaluated by its success in assigning a high score to representations that are expected to be matched. However, existing similarity measures perform mediocrely on standard benchmarks. In this work, we develop a new similarity measure, dubbed ContraSim, based on contrastive learning. In contrast to common closed-form similarity measures, ContraSim learns a parameterized measure by using both similar and dissimilar examples. We perform an extensive experimental evaluation of our method, with both language and vision models, on the standard layer prediction benchmark and two new benchmarks that we introduce: the multilingual benchmark and the image-caption benchmark. In all cases, ContraSim achieves much higher accuracy than previous similarity measures, even when presented with challenging examples. Finally, ContraSim is more suitable for the analysis of neural networks, revealing new insights not captured by previous measures.
Authors: Jen-tse Huang, Wenxiang Jiao, Man Ho Lam, Eric John Li, Wenxuan Wang, Michael R. Lyu
Abstract: Recent research has focused on examining Large Language Models' (LLMs) characteristics from a psychological standpoint, acknowledging the necessity of understanding their behavioral characteristics. The administration of personality tests to LLMs has emerged as a noteworthy area in this context. However, the suitability of employing psychological scales, initially devised for humans, on LLMs is a matter of ongoing debate. Our study aims to determine the reliability of applying personality assessments to LLMs, explicitly investigating whether LLMs demonstrate consistent personality traits. Analysis of 2,500 settings per model, including GPT-3.5, GPT-4, Gemini-Pro, and LLaMA-3.1, reveals that various LLMs show consistency in responses to the Big Five Inventory, indicating a satisfactory level of reliability. Furthermore, our research explores the potential of GPT-3.5 to emulate diverse personalities and represent various groups-a capability increasingly sought after in social sciences for substituting human participants with LLMs to reduce costs. Our findings reveal that LLMs have the potential to represent different personalities with specific prompt instructions.
Authors: Weihe Zhai, Arkaitz Zubiaga
Abstract: The fusion of language models (LMs) and knowledge graphs (KGs) is widely used in commonsense question answering, but generating faithful explanations remains challenging. Current methods often overlook path decoding faithfulness, leading to divergence between graph encoder outputs and model predictions. We identify confounding effects and LM-KG misalignment as key factors causing spurious explanations. To address this, we introduce the LM-KG Fidelity metric to assess KG representation reliability and propose the LM-KG Distribution-aware Alignment (\textit{LKDA}) algorithm to improve explanation faithfulness. Without ground truth, we evaluate KG explanations using the proposed Fidelity-Sparsity Trade-off Curve. Experiments on CommonsenseQA and OpenBookQA show that LKDA significantly enhances explanation fidelity and model performance, highlighting the need to address distributional misalignment for reliable commonsense reasoning.
Authors: Alexander Brinkmann, Roee Shraga, Christian Bizer
Abstract: E-commerce platforms require structured product data in the form of attribute-value pairs to offer features such as faceted product search or attribute-based product comparison. However, vendors often provide unstructured product descriptions, necessitating the extraction of attribute-value pairs from these texts. BERT-based extraction methods require large amounts of task-specific training data and struggle with unseen attribute values. This paper explores using large language models (LLMs) as a more training-data efficient and robust alternative. We propose prompt templates for zero-shot and few-shot scenarios, comparing textual and JSON-based target schema representations. Our experiments show that GPT-4 achieves the highest average F1-score of 85% using detailed attribute descriptions and demonstrations. Llama-3-70B performs nearly as well, offering a competitive open-source alternative. GPT-4 surpasses the best PLM baseline by 5% in F1-score. Fine-tuning GPT-3.5 increases the performance to the level of GPT-4 but reduces the model's ability to generalize to unseen attribute values.
Authors: Song Wang, Yaochen Zhu, Haochen Liu, Zaiyi Zheng, Chen Chen, Jundong Li
Abstract: Large language models (LLMs) have recently transformed both the academic and industrial landscapes due to their remarkable capacity to understand, analyze, and generate texts based on their vast knowledge and reasoning ability. Nevertheless, one major drawback of LLMs is their substantial computational cost for pre-training due to their unprecedented amounts of parameters. The disadvantage is exacerbated when new knowledge frequently needs to be introduced into the pre-trained model. Therefore, it is imperative to develop effective and efficient techniques to update pre-trained LLMs. Traditional methods encode new knowledge in pre-trained LLMs through direct fine-tuning. However, naively re-training LLMs can be computationally intensive and risks degenerating valuable pre-trained knowledge irrelevant to the update in the model. Recently, Knowledge-based Model Editing (KME) has attracted increasing attention, which aims to precisely modify the LLMs to incorporate specific knowledge, without negatively influencing other irrelevant knowledge. In this survey, we aim to provide a comprehensive and in-depth overview of recent advances in the field of KME. We first introduce a general formulation of KME to encompass different KME strategies. Afterward, we provide an innovative taxonomy of KME techniques based on how the new knowledge is introduced into pre-trained LLMs, and investigate existing KME strategies while analyzing key insights, advantages, and limitations of methods from each category. Moreover, representative metrics, datasets, and applications of KME are introduced accordingly. Finally, we provide an in-depth analysis regarding the practicality and remaining challenges of KME and suggest promising research directions for further advancement in this field.
Authors: Lang Cao
Abstract: Large language models (LLMs) have demonstrated impressive language understanding and generation capabilities, enabling them to answer a wide range of questions across various domains. However, these models are not flawless and often produce responses that contain errors or misinformation. These inaccuracies, commonly referred to as hallucinations, render LLMs unreliable and even unusable in many scenarios. In this paper, our focus is on mitigating the issue of hallucination in LLMs, particularly in the context of question-answering. Instead of attempting to answer all questions, we explore a refusal mechanism that instructs LLMs to refuse to answer challenging questions in order to avoid errors. We then propose a simple yet effective solution called Learn to Refuse (L2R), which incorporates the refusal mechanism to enable LLMs to recognize and refuse to answer questions that they find difficult to address. To achieve this, we utilize a structured knowledge base to represent all the LLM's understanding of the world, enabling it to provide traceable gold knowledge. This knowledge base is separate from the LLM and initially empty. It can be filled with validated knowledge and progressively expanded. When an LLM encounters questions outside its domain, the system recognizes its knowledge scope and determines whether it can answer the question independently. Additionally, we introduce a method for automatically and efficiently expanding the knowledge base of LLMs. Through qualitative and quantitative analysis, we demonstrate that our approach enhances the controllability and reliability of LLMs.
Authors: Zihan Zhao, Da Ma, Lu Chen, Liangtai Sun, Zihao Li, Yi Xia, Bo Chen, Hongshen Xu, Zichen Zhu, Su Zhu, Shuai Fan, Guodong Shen, Kai Yu, Xin Chen
Abstract: Artificial intelligence (AI) has played an increasingly important role in chemical research. However, most models currently used in chemistry are specialist models that require training and tuning for specific tasks. A more generic and efficient solution would be an AI model that could address many tasks and support free-form dialogue in the broad field of chemistry. In its utmost form, such a generalist AI chemist could be referred to as Chemical General Intelligence. Large language models (LLMs) have recently logged tremendous success in the general domain of natural language processing, showing emerging task generalization and free-form dialogue capabilities. However, domain knowledge of chemistry is largely missing when training general-domain LLMs. The lack of such knowledge greatly hinders the performance of generalist LLMs in the field of chemistry. To this end, we develop ChemDFM, a pioneering LLM for chemistry trained on 34B tokens from chemical literature and textbooks, and fine-tuned using 2.7M instructions. As a result, it can understand and reason with chemical knowledge in free-form dialogue. Quantitative evaluations show that ChemDFM significantly surpasses most representative open-source LLMs. It outperforms GPT-4 on a great portion of chemical tasks, despite the substantial size difference. We have open-sourced the inference codes, evaluation datasets, and model weights of ChemDFM on Huggingface (https://huggingface.co/AI4Chem/ChemLLM-7B-Chat).
Authors: Spyridon Mouselinos, Henryk Michalewski, Mateusz Malinowski
Abstract: Large Language Models (LLMs) demonstrate ever-increasing abilities in mathematical and algorithmic tasks, yet their geometric reasoning skills are underexplored. We investigate LLMs' abilities in constructive geometric problem-solving one of the most fundamental steps in the development of human mathematical reasoning. Our work reveals notable challenges that the state-of-the-art LLMs face in this domain despite many successes in similar areas. LLMs exhibit biases in target variable selection and struggle with 2D spatial relationships, often misrepresenting and hallucinating objects and their placements. To this end, we introduce a framework that formulates an LLMs-based multi-agents system that enhances their existing reasoning potential by conducting an internal dialogue. This work underscores LLMs' current limitations in geometric reasoning and improves geometric reasoning capabilities through self-correction, collaboration, and diverse role specializations.
Authors: Xianzhen Luo, Qingfu Zhu, Zhiming Zhang, Libo Qin, Xuanyu Zhang, Qing Yang, Dongliang Xu, Wanxiang Che
Abstract: Program of Thoughts (PoT) is an approach characterized by its executable intermediate steps, which ensure the accuracy of the logical calculations in the reasoning process. Currently, PoT primarily uses Python. However, relying solely on a single language may result in suboptimal solutions and overlook the potential benefits of other programming languages. In this paper, we conduct comprehensive experiments on the programming languages used in PoT and find that no single language consistently delivers optimal performance across all tasks and models. The effectiveness of each language varies depending on the specific scenarios. Inspired by this, we propose a task and model agnostic approach called MultiPoT, which harnesses strength and diversity from various languages. Experimental results reveal that it significantly outperforms Python Self-Consistency. Furthermore, it achieves comparable or superior performance compared to the best monolingual PoT in almost all tasks across all models. In particular, MultiPoT achieves more than 4.6% improvement on average on ChatGPT (gpt-3.5-turbo-0701).
Authors: Mohamed Taher Alrefaie, Nour Eldin Morsy, Nada Samir
Abstract: This paper presents a comprehensive examination of the impact of tokenization strategies and vocabulary sizes on the performance of Arabic language models in downstream natural language processing tasks. Our investigation focused on the effectiveness of four tokenizers across various tasks, including News Classification, Hate Speech Detection, Sentiment Analysis, and Natural Language Inference. Leveraging a diverse set of vocabulary sizes, we scrutinize the intricate interplay between tokenization approaches and model performance. The results reveal that Byte Pair Encoding (BPE) with Farasa outperforms other strategies in multiple tasks, underscoring the significance of morphological analysis in capturing the nuances of the Arabic language. However, challenges arise in sentiment analysis, where dialect specific segmentation issues impact model efficiency. Computational efficiency analysis demonstrates the stability of BPE with Farasa, suggesting its practical viability. Our study uncovers limited impacts of vocabulary size on model performance while keeping the model size unchanged. This is challenging the established beliefs about the relationship between vocabulary, model size, and downstream tasks, emphasizing the need for the study of models' size and their corresponding vocabulary size to generalize across domains and mitigate biases, particularly in dialect based datasets. Paper's recommendations include refining tokenization strategies to address dialect challenges, enhancing model robustness across diverse linguistic contexts, and expanding datasets to encompass the rich dialect based Arabic. This work not only advances our understanding of Arabic language models but also lays the foundation for responsible and ethical developments in natural language processing technologies tailored to the intricacies of the Arabic language.
Authors: Huanxuan Liao, Shizhu He, Yao Xu, Yuanzhe Zhang, Kang Liu, Shengping Liu, Jun Zhao
Abstract: Retrieval-Augmented-Generation and Generation-Augmented-Generation have been proposed to enhance the knowledge required for question answering with Large Language Models (LLMs) by leveraging richer context. However, the former relies on external resources, and both require incorporating explicit documents into the context, which increases execution costs and susceptibility to noise data during inference. Recent works indicate that LLMs model rich knowledge, but it is often not effectively activated and awakened. Inspired by this, we propose a novel knowledge-augmented framework, $\textbf{Awakening-Augmented-Generation}$ (AAG), which mimics the human ability to answer questions using only thinking and recalling to compensate for knowledge gaps, thereby awaking relevant knowledge in LLMs without relying on external resources. AAG consists of two key components for awakening richer context. Explicit awakening fine-tunes a context generator to create a synthetic, compressed document that functions as symbolic context. Implicit awakening utilizes a hypernetwork to generate adapters based on the question and synthetic document, which are inserted into LLMs to serve as parameter context. Experimental results on three datasets demonstrate that AAG exhibits significant advantages in both open-domain and closed-book settings, as well as in out-of-distribution generalization. Our code will be available at \url{https://github.com/Xnhyacinth/IAG}.
Authors: Julen Etxaniz, Oscar Sainz, Naiara Perez, Itziar Aldabe, German Rigau, Eneko Agirre, Aitor Ormazabal, Mikel Artetxe, Aitor Soroa
Abstract: We introduce Latxa, a family of large language models for Basque ranging from 7 to 70 billion parameters. Latxa is based on Llama 2, which we continue pretraining on a new Basque corpus comprising 4.3M documents and 4.2B tokens. Addressing the scarcity of high-quality benchmarks for Basque, we further introduce 4 multiple choice evaluation datasets: EusProficiency, comprising 5,169 questions from official language proficiency exams; EusReading, comprising 352 reading comprehension questions; EusTrivia, comprising 1,715 trivia questions from 5 knowledge areas; and EusExams, comprising 16,774 questions from public examinations. In our extensive evaluation, Latxa outperforms all previous open models we compare to by a large margin. In addition, it is competitive with GPT-4 Turbo in language proficiency and understanding, despite lagging behind in reading comprehension and knowledge-intensive tasks. Both the Latxa family of models, as well as our new pretraining corpora and evaluation datasets, are publicly available under open licenses. Our suite enables reproducible research on methods to build LLMs for low-resource languages.
Authors: Jiayin Wang, Fengran Mo, Weizhi Ma, Peijie Sun, Min Zhang, Jian-Yun Nie
Abstract: Large language models (LLMs) are essential tools that users employ across various scenarios, so evaluating their performance and guiding users in selecting the suitable service is important. Although many benchmarks exist, they mainly focus on specific predefined model abilities, such as world knowledge, reasoning, etc. Based on these ability scores, it is hard for users to determine which LLM best suits their particular needs. To address these issues, we propose to evaluate LLMs from a user-centric perspective and design this benchmark to measure their efficacy in satisfying user needs under distinct intents. Firstly, we collect 1,846 real-world use cases from a user study with 712 participants from 23 countries. This first-hand data helps us understand actual user intents and needs in LLM interactions, forming the User Reported Scenarios (URS) dataset, which is categorized with six types of user intents. Secondly, based on this authentic dataset, we benchmark 10 LLM services with GPT-4-as-Judge. Thirdly, we show that benchmark scores align well with human preference in both real-world experience and pair-wise annotations, achieving Pearson correlations of 0.95 and 0.94, respectively. This alignment confirms that the URS dataset and our evaluation method establish an effective user-centric benchmark. The dataset, code, and process data are available at https://github.com/Alice1998/URS.
Authors: Yu Xia, Rui Wang, Xu Liu, Mingyan Li, Tong Yu, Xiang Chen, Julian McAuley, Shuai Li
Abstract: Chain-of-Thought (CoT) has been a widely adopted prompting method, eliciting impressive reasoning abilities of Large Language Models (LLMs). Inspired by the sequential thought structure of CoT, a number of Chain-of-X (CoX) methods have been developed to address various challenges across diverse domains and tasks involving LLMs. In this paper, we provide a comprehensive survey of Chain-of-X methods for LLMs in different contexts. Specifically, we categorize them by taxonomies of nodes, i.e., the X in CoX, and application tasks. We also discuss the findings and implications of existing CoX methods, as well as potential future directions. Our survey aims to serve as a detailed and up-to-date resource for researchers seeking to apply the idea of CoT to broader scenarios.
Authors: Lianming Huang, Shangyu Wu, Yufei Cui, Ying Xiong, Xue Liu, Tei-Wei Kuo, Nan Guan, Chun Jason Xue
Abstract: Deploying large language model inference remains challenging due to their high computational overhead. Early exiting optimizes model inference by adaptively reducing the number of inference layers. Existing methods typically train internal classifiers to determine whether to exit at intermediate layers. However, such classifier-based early exiting frameworks require significant effort to train the classifiers while can only achieve comparable performance at best. To address these limitations, this paper proposes RAEE, a robust Retrieval-Augmented Early Exiting framework for efficient inference. First, this paper demonstrates that the early exiting problem can be modeled as a distribution prediction problem, where the distribution is approximated using similar data's exiting information. Then, this paper details the process of collecting exiting information to build the retrieval database. Finally, based on the pre-built retrieval database, RAEE leverages the retrieved similar data's exiting information to guide the backbone model to exit at the layer, which is predicted by the approximated distribution. Experimental results demonstrate that the proposed RAEE can significantly accelerate inference. More importantly, RAEE can also achieve a robust zero-shot performance on 8 downstream tasks.
Authors: Shruthi Bannur, Kenza Bouzid, Daniel C. Castro, Anton Schwaighofer, Anja Thieme, Sam Bond-Taylor, Maximilian Ilse, Fernando P\'erez-Garc\'ia, Valentina Salvatelli, Harshita Sharma, Felix Meissen, Mercy Ranjit, Shaury Srivastav, Julia Gong, Noel C. F. Codella, Fabian Falck, Ozan Oktay, Matthew P. Lungren, Maria Teodora Wetscherek, Javier Alvarez-Valle, Stephanie L. Hyland
Abstract: Radiology reporting is a complex task requiring detailed medical image understanding and precise language generation, for which generative multimodal models offer a promising solution. However, to impact clinical practice, models must achieve a high level of both verifiable performance and utility. We augment the utility of automated report generation by incorporating localisation of individual findings on the image - a task we call grounded report generation - and enhance performance by incorporating realistic reporting context as inputs. We design a novel evaluation framework (RadFact) leveraging the logical inference capabilities of large language models (LLMs) to quantify report correctness and completeness at the level of individual sentences, while supporting the new task of grounded reporting. We develop MAIRA-2, a large radiology-specific multimodal model designed to generate chest X-ray reports with and without grounding. MAIRA-2 achieves state of the art on existing report generation benchmarks and establishes the novel task of grounded report generation.
Authors: Qianli Wang, Tatiana Anikina, Nils Feldhus, Simon Ostermann, Sebastian M\"oller
Abstract: Conversational explainable artificial intelligence (ConvXAI) systems based on large language models (LLMs) have garnered significant interest from the research community in natural language processing (NLP) and human-computer interaction (HCI). Such systems can provide answers to user questions about explanations in dialogues, have the potential to enhance users' comprehension and offer more information about the decision-making and generation processes of LLMs. Currently available ConvXAI systems are based on intent recognition rather than free chat, as this has been found to be more precise and reliable in identifying users' intentions. However, the recognition of intents still presents a challenge in the case of ConvXAI, since little training data exist and the domain is highly specific, as there is a broad range of XAI methods to map requests onto. In order to bridge this gap, we present CoXQL, the first dataset in the NLP domain for user intent recognition in ConvXAI, covering 31 intents, seven of which require filling multiple slots. Subsequently, we enhance an existing parsing approach by incorporating template validations, and conduct an evaluation of several LLMs on CoXQL using different parsing strategies. We conclude that the improved parsing approach (MP+) surpasses the performance of previous approaches. We also discover that intents with multiple slots remain highly challenging for LLMs.
Authors: Huaizhi Ge, Frank Rudzicz, Zining Zhu
Abstract: In the field of language model interpretability, circuit discovery is gaining popularity. Despite this, the true meaning of these circuits remains largely unanswered. We introduce a novel method to learn their meanings as a holistic object through the lens of knowledge editing. We extract circuits in the GPT-2 base model for classification tasks related to syntax and model safety, and study their knowledge property via a model edit dataset containing hierarchical entities. We find that these circuits contain entity knowledge but resist new knowledge, demonstrating a "confirmation bias" behavior. Additionally, we examine the impact of circuit size, discovering that an ideal "theoretical circuit" where essential knowledge is concentrated likely incorporates more than 5% but less than 50% of the model's parameters. We also assess the overlap between circuits from different datasets, finding moderate similarities. We proceed with analyzing the modular components of the circuits, finding that up to 60% of the circuits consist of layer normalization modules rather than attention or MLP modules, adding evidence to the ongoing debates regarding knowledge localization. In summary, our findings offer novel insights into the meanings of the circuits, and introduce directions for further interpretability and safety research of language models.
Authors: Juli Bakagianni, Kanella Pouli, Maria Gavriilidou, John Pavlopoulos
Abstract: Natural Language Processing (NLP) research has traditionally been predominantly focused on English, driven by the availability of resources, the size of the research community, and market demands. Recently, there has been a noticeable shift towards multilingualism in NLP, recognizing the need for inclusivity and effectiveness across diverse languages and cultures. Monolingual surveys have the potential to complement the broader trend towards multilingualism in NLP by providing foundational insights and resources necessary for effectively addressing the linguistic diversity of global communication. However, monolingual NLP surveys are extremely rare in literature. This study fills the gap by introducing a method for creating systematic and comprehensive monolingual NLP surveys. Characterized by a structured search protocol, it can be used to select publications and organize them through a taxonomy of NLP tasks. We include a classification of Language Resources (LRs), according to their availability, and datasets, according to their annotation, to highlight publicly-available and machine-actionable LRs. By applying our method, we conducted a systematic literature review of Greek NLP from 2012 to 2022, providing a comprehensive overview of the current state and challenges of Greek NLP research. We discuss the progress of Greek NLP and outline encountered Greek LRs, classified by availability and usability. As we show, our proposed method helps avoid common pitfalls, such as data leakage and contamination, and to assess language support per NLP task. We consider this systematic literature review of Greek NLP an application of our method that showcases the benefits of a monolingual NLP survey. Similar applications could be regard the myriads of languages whose progress in NLP lags behind that of well-supported languages.
Authors: Jeonghyun Park, Hwanhee Lee
Abstract: Conversational search seeks to retrieve relevant passages for the given questions in conversational question answering. Conversational Query Reformulation (CQR) improves conversational search by refining the original queries into de-contextualized forms to resolve the issues in the original queries, such as omissions and coreferences. Previous CQR methods focus on imitating human written queries which may not always yield meaningful search results for the retriever. In this paper, we introduce GuideCQR, a framework that refines queries for CQR by leveraging key information from the initially retrieved documents. Specifically, GuideCQR extracts keywords and generates expected answers from the retrieved documents, then unifies them with the queries after filtering to add useful information that enhances the search process. Experimental results demonstrate that our proposed method achieves state-of-the-art performance across multiple datasets, outperforming previous CQR methods. Additionally, we show that GuideCQR can get additional performance gains in conversational search using various types of queries, even for queries written by humans.
Authors: Yanjun Gao, Skatje Myers, Shan Chen, Dmitriy Dligach, Timothy A Miller, Danielle Bitterman, Matthew Churpek, Majid Afshar
Abstract: The introduction of Large Language Models (LLMs) has advanced data representation and analysis, bringing significant progress in their use for medical questions and answering. Despite these advancements, integrating tabular data, especially numerical data pivotal in clinical contexts, into LLM paradigms has not been thoroughly explored. In this study, we examine the effectiveness of vector representations from last hidden states of LLMs for medical diagnostics and prognostics using electronic health record (EHR) data. We compare the performance of these embeddings with that of raw numerical EHR data when used as feature inputs to traditional machine learning (ML) algorithms that excel at tabular data learning, such as eXtreme Gradient Boosting. We focus on instruction-tuned LLMs in a zero-shot setting to represent abnormal physiological data and evaluating their utilities as feature extractors to enhance ML classifiers for predicting diagnoses, length of stay, and mortality. Furthermore, we examine prompt engineering techniques on zero-shot and few-shot LLM embeddings to measure their impact comprehensively. Although findings suggest the raw data features still prevails in medical ML tasks, zero-shot LLM embeddings demonstrate competitive results, suggesting a promising avenue for future research in medical applications.
Authors: Yang Liu, Xichou Zhu, Zhou Shen, Yi Liu, Min Li, Yujun Chen, Benzi John, Zhenzhen Ma, Zhi Li, Tao Hu, Zhiyang Xu, Wei Luo, Junhui Wang
Abstract: Large Language Models (LLMs) have recently displayed their extraordinary capabilities in language understanding. However, how to comprehensively assess the sentiment capabilities of LLMs continues to be a challenge. This paper investigates the ability of LLMs to detect and react to sentiment in text modal. As the integration of LLMs into diverse applications is on the rise, it becomes highly critical to comprehend their sensitivity to emotional tone, as it can influence the user experience and the efficacy of sentiment-driven tasks. We conduct a series of experiments to evaluate the performance of several prominent LLMs in identifying and responding appropriately to sentiments like positive, negative, and neutral emotions. The models' outputs are analyzed across various sentiment benchmarks, and their responses are compared with human evaluations. Our discoveries indicate that although LLMs show a basic sensitivity to sentiment, there are substantial variations in their accuracy and consistency, emphasizing the requirement for further enhancements in their training processes to better capture subtle emotional cues. Take an example in our findings, in some cases, the models might wrongly classify a strongly positive sentiment as neutral, or fail to recognize sarcasm or irony in the text. Such misclassifications highlight the complexity of sentiment analysis and the areas where the models need to be refined. Another aspect is that different LLMs might perform differently on the same set of data, depending on their architecture and training datasets. This variance calls for a more in-depth study of the factors that contribute to the performance differences and how they can be optimized.
Authors: Xichou Zhu, Yang Liu, Zhou Shen, Yi Liu, Min Li, Yujun Chen, Benzi John, Zhenzhen Ma, Zhi Li, Tao Hu, Bolong Yang, Manman Wang, Zongxing Xie, Peng Liu, Dan Cai, Junhui Wang
Abstract: The recent advances in large language models (LLMs) have significantly expanded their applications across various fields such as language generation, summarization, and complex question answering. However, their application to privacy compliance and technical privacy reviews remains under-explored, raising critical concerns about their ability to adhere to global privacy standards and protect sensitive user data. This paper seeks to address this gap by providing a comprehensive case study evaluating LLMs' performance in privacy-related tasks such as privacy information extraction (PIE), legal and regulatory key point detection (KPD), and question answering (QA) with respect to privacy policies and data protection regulations. We introduce a Privacy Technical Review (PTR) framework, highlighting its role in mitigating privacy risks during the software development life-cycle. Through an empirical assessment, we investigate the capacity of several prominent LLMs, including BERT, GPT-3.5, GPT-4, and custom models, in executing privacy compliance checks and technical privacy reviews. Our experiments benchmark the models across multiple dimensions, focusing on their precision, recall, and F1-scores in extracting privacy-sensitive information and detecting key regulatory compliance points. While LLMs show promise in automating privacy reviews and identifying regulatory discrepancies, significant gaps persist in their ability to fully comply with evolving legal standards. We provide actionable recommendations for enhancing LLMs' capabilities in privacy compliance, emphasizing the need for robust model improvements and better integration with legal and regulatory requirements. This study underscores the growing importance of developing privacy-aware LLMs that can both support businesses in compliance efforts and safeguard user privacy rights.
Authors: Gene-Ping Yang, Hao Tang
Abstract: Despite the recent advance in self-supervised representations, unsupervised phonetic segmentation remains challenging. Most approaches focus on improving phonetic representations with self-supervised learning, with the hope that the improvement can transfer to phonetic segmentation. In this paper, contrary to recent approaches, we show that peak detection on Mel spectrograms is a strong baseline, better than many self-supervised approaches. Based on this finding, we propose a simple hidden Markov model that uses self-supervised representations and features at the boundaries for phone segmentation. Our results demonstrate consistent improvements over previous approaches, with a generalized formulation allowing versatile design adaptations.
Authors: Asir Saadat, Nahian Ibn Asad, Md Farhan Ishmam
Abstract: Contextual question-answering models are susceptible to adversarial perturbations to input context, commonly observed in real-world scenarios. These adversarial noises are designed to degrade the performance of the model by distorting the textual input. We introduce a unique dataset that incorporates seven distinct types of adversarial noise into the context, each applied at five different intensity levels on the SQuAD dataset. To quantify the robustness, we utilize robustness metrics providing a standardized measure for assessing model performance across varying noise types and levels. Experiments on transformer-based question-answering models reveal robustness vulnerabilities and important insights into the model's performance in realistic textual input.
Authors: Adrian Cosma, Ana-Maria Bucur, Emilian Radoi
Abstract: Mathematics has long been conveyed through natural language, primarily for human understanding. With the rise of mechanized mathematics and proof assistants, there is a growing need to understand informal mathematical text, yet most existing benchmarks focus solely on English, overlooking other languages. This paper introduces RoMath, a Romanian mathematical reasoning benchmark suite comprising three datasets: RoMath-Baccalaureate, RoMath-Competitions and RoMath-Synthetic, which cover a range of mathematical domains and difficulty levels, aiming to improve non-English language models and promote multilingual AI development. By focusing on Romanian, a low-resource language with unique linguistic features, RoMath addresses the limitations of Anglo-centric models and emphasizes the need for dedicated resources beyond simple automatic translation. We benchmark several open-weight language models, highlighting the importance of creating resources for underrepresented languages. We make the code and dataset available.
Authors: Bo Liu, Liming Zhan, Yujie Feng, Zexin Lu, Chengqiang Xie, Lei Xue, Albert Y. S. Lam, Xiao-Ming Wu
Abstract: In the realm of task-oriented dialogue systems, a robust intent detection mechanism must effectively handle malformed utterances encountered in real-world scenarios. This study presents a novel fine-tuning framework for large language models (LLMs) aimed at enhancing in-distribution (ID) intent classification and out-of-distribution (OOD) intent detection, which utilizes semantic matching with prototypes derived from ID class names. By harnessing the highly distinguishable representations of LLMs, we construct semantic prototypes for each ID class using a diversity-grounded prompt tuning approach. We rigorously test our framework in a challenging OOD context, where ID and OOD classes are semantically close yet distinct, referred to as \emph{near} OOD detection. For a thorough assessment, we benchmark our method against the prevalent fine-tuning approaches. The experimental findings reveal that our method demonstrates superior performance in both few-shot ID intent classification and near-OOD intent detection tasks.
Authors: Raffaele Marino
Abstract: In this manuscript, I present an analysis on the performance of OpenAI O1-preview model in solving random K-SAT instances for K$\in {2,3,4}$ as a function of $\alpha=M/N$ where $M$ is the number of clauses and $N$ is the number of variables of the satisfiable problem. I show that the model can call an external SAT solver to solve the instances, rather than solving them directly. Despite using external solvers, the model reports incorrect assignments as output. Moreover, I propose and present an analysis to quantify whether the OpenAI O1-preview model demonstrates a spark of intelligence or merely makes random guesses when outputting an assignment for a Boolean satisfiability problem.
Authors: Issey Sukeda
Abstract: The recent success of large language models (LLMs) and the scaling law has led to a widespread adoption of larger models. Particularly in the healthcare industry, there is an increasing demand for locally operated LLMs due to security concerns. However, the majority of high quality open-source LLMs have a size of 70B parameters, imposing significant financial burdens on users for GPU preparation and operation. To overcome these issues, we present a medical adaptation based on the recent 7B models, which enables the operation in low computational resources. We compare the performance on medical question-answering benchmarks in two languages (Japanese and English), demonstrating that its scores reach parity with or surpass those of currently existing medical LLMs that are ten times larger. We find that fine-tuning an English-centric base model on Japanese medical dataset improves the score in both language, supporting the effect of cross-lingual knowledge transfer. We hope that this study will alleviate financial challenges, serving as a stepping stone for clinical institutions to practically utilize LLMs locally. Our evaluation code is available at https://github.com/stardust-coder/japanese-lm-med-harness.
URLs: https://github.com/stardust-coder/japanese-lm-med-harness.
Authors: Somonnoy Banerjee, Sujan Dutta, Soumyajit Datta, Ashiqur R. KhudaBukhsh
Abstract: This paper makes three key contributions. First, via a substantial corpus of 51,278 interview questions sourced from 888 YouTube videos of mock interviews of Indian civil service candidates, we demonstrate stark gender bias in the broad nature of questions asked to male and female candidates. Second, our experiments with large language models show a strong presence of gender bias in explanations provided by the LLMs on the gender inference task. Finally, we present a novel dataset of 51,278 interview questions that can inform future social science studies.
Authors: Sujan Dutta, Deepak Pandita, Tharindu Cyril Weerasooriya, Marcos Zampieri, Christopher M. Homan, Ashiqur R. KhudaBukhsh
Abstract: Ensuring annotator quality in training and evaluation data is a key piece of machine learning in NLP. Tasks such as sentiment analysis and offensive speech detection are intrinsically subjective, creating a challenging scenario for traditional quality assessment approaches because it is hard to distinguish disagreement due to poor work from that due to differences of opinions between sincere annotators. With the goal of increasing diverse perspectives in annotation while ensuring consistency, we propose \texttt{ARTICLE}, an in-context learning (ICL) framework to estimate annotation quality through self-consistency. We evaluate this framework on two offensive speech datasets using multiple LLMs and compare its performance with traditional methods. Our findings indicate that \texttt{ARTICLE} can be used as a robust method for identifying reliable annotators, hence improving data quality.
Authors: Jun Rao, Xuebo Liu, Zepeng Lin, Liang Ding, Jing Li, Dacheng Tao, Min Zhang
Abstract: Knowledge distillation (KD) is a technique that compresses large teacher models by training smaller student models to mimic them. The success of KD in auto-regressive language models mainly relies on Reverse KL for mode-seeking and student-generated output (SGO) to combat exposure bias. Our theoretical analyses and experimental validation reveal that while Reverse KL effectively mimics certain features of the teacher distribution, it fails to capture most of its behaviors. Conversely, SGO incurs higher computational costs and presents challenges in optimization, particularly when the student model is significantly smaller than the teacher model. These constraints are primarily due to the immutable distribution of the teacher model, which fails to adjust adaptively to models of varying sizes. We introduce Online Knowledge Distillation (OKD), where the teacher network integrates small online modules to concurrently train with the student model. This strategy abolishes the necessity for on-policy sampling and merely requires minimal updates to the parameters of the teacher's online module during training, thereby allowing dynamic adaptation to the student's distribution to make distillation better. Extensive results across multiple generation datasets show that OKD achieves or exceeds the performance of leading methods in various model architectures and sizes, reducing training time by up to fourfold.
Authors: Kiran Vodrahalli, Santiago Ontanon, Nilesh Tripuraneni, Kelvin Xu, Sanil Jain, Rakesh Shivanna, Jeffrey Hui, Nishanth Dikkala, Mehran Kazemi, Bahare Fatemi, Rohan Anil, Ethan Dyer, Siamak Shakeri, Roopali Vij, Harsh Mehta, Vinay Ramasesh, Quoc Le, Ed Chi, Yifeng Lu, Orhan Firat, Angeliki Lazaridou, Jean-Baptiste Lespiau, Nithya Attaluri, Kate Olszewska
Abstract: We introduce Michelangelo: a minimal, synthetic, and unleaked long-context reasoning evaluation for large language models which is also easy to automatically score. This evaluation is derived via a novel, unifying framework for evaluations over arbitrarily long contexts which measure the model's ability to do more than retrieve a single piece of information from its context. The central idea of the Latent Structure Queries framework (LSQ) is to construct tasks which require a model to ``chisel away'' the irrelevant information in the context, revealing a latent structure in the context. To verify a model's understanding of this latent structure, we query the model for details of the structure. Using LSQ, we produce three diagnostic long-context evaluations across code and natural-language domains intended to provide a stronger signal of long-context language model capabilities. We perform evaluations on several state-of-the-art models and demonstrate both that a) the proposed evaluations are high-signal and b) that there is significant room for improvement in synthesizing long-context information.
Authors: Mohammad Samragh, Iman Mirzadeh, Keivan Alizadeh Vahid, Fartash Faghri, Minsik Cho, Moin Nabi, Devang Naik, Mehrdad Farajtabar
Abstract: The pre-training phase of language models often begins with randomly initialized parameters. With the current trends in scaling models, training their large number of parameters can be extremely slow and costly. In contrast, small language models are less expensive to train, but they often cannot achieve the accuracy of large models. In this paper, we explore an intriguing idea to connect these two different regimes: Can we develop a method to initialize large language models using smaller pre-trained models? Will such initialization bring any benefits in terms of training time and final accuracy? In this paper, we introduce HyperCloning, a method that can expand the parameters of a pre-trained language model to those of a larger model with increased hidden dimensions. Our method ensures that the larger model retains the functionality of the smaller model. As a result, the larger model already inherits the predictive power and accuracy of the smaller model before the training starts. We demonstrate that training such an initialized model results in significant savings in terms of GPU hours required for pre-training large language models.
Authors: Junzhe Zhang, Huixuan Zhang, Xunjian Yin, Xiaojun Wan
Abstract: News image captioning requires model to generate an informative caption rich in entities, with the news image and the associated news article. Current MLLMs still bear limitations in handling entity information in news image captioning tasks. Besides, generating high-quality news image captions requires a trade-off between sufficiency and conciseness of textual input information. To explore the potential of MLLMs and address problems we discovered, we propose EAMA: an Entity-Aware Multimodal Alignment based approach for News Image Captioning. Our approach first aligns the MLLM with two extra alignment tasks: Entity-Aware Sentence Selection task and Entity Selection task, together with News Image Captioning task. The aligned MLLM will utilize the additional entity-related information extracted by itself to supplement the textual input while generating news image captions. Our approach achieves better results than all previous models on two mainstream news image captioning datasets.
Authors: Junfeng Jiao, Saleh Afroogh, Yiming Xu, Connor Phillips
Abstract: This study addresses ethical issues surrounding Large Language Models (LLMs) within the field of artificial intelligence. It explores the common ethical challenges posed by both LLMs and other AI systems, such as privacy and fairness, as well as ethical challenges uniquely arising from LLMs. It highlights challenges such as hallucination, verifiable accountability, and decoding censorship complexity, which are unique to LLMs and distinct from those encountered in traditional AI systems. The study underscores the need to tackle these complexities to ensure accountability, reduce biases, and enhance transparency in the influential role that LLMs play in shaping information dissemination. It proposes mitigation strategies and future directions for LLM ethics, advocating for interdisciplinary collaboration. It recommends ethical frameworks tailored to specific domains and dynamic auditing systems adapted to diverse contexts. This roadmap aims to guide responsible development and integration of LLMs, envisioning a future where ethical considerations govern AI advancements in society.
Authors: Gengze Zhou, Yicong Hong, Zun Wang, Xin Eric Wang, Qi Wu
Abstract: Capitalizing on the remarkable advancements in Large Language Models (LLMs), there is a burgeoning initiative to harness LLMs for instruction following robotic navigation. Such a trend underscores the potential of LLMs to generalize navigational reasoning and diverse language understanding. However, a significant discrepancy in agent performance is observed when integrating LLMs in the Vision-and-Language navigation (VLN) tasks compared to previous downstream specialist models. Furthermore, the inherent capacity of language to interpret and facilitate communication in agent interactions is often underutilized in these integrations. In this work, we strive to bridge the divide between VLN-specialized models and LLM-based navigation paradigms, while maintaining the interpretative prowess of LLMs in generating linguistic navigational reasoning. By aligning visual content in a frozen LLM, we encompass visual observation comprehension for LLMs and exploit a way to incorporate LLMs and navigation policy networks for effective action predictions and navigational reasoning. We demonstrate the data efficiency of the proposed methods and eliminate the gap between LM-based agents and state-of-the-art VLN specialists.
Authors: Ting-Chih Chen, Chia-Wei Tang, Chris Thomas
Abstract: Fact-checking real-world claims often requires reviewing multiple multimodal documents to assess a claim's truthfulness, which is a highly laborious and time-consuming task. In this paper, we present a summarization model designed to generate claim-specific summaries useful for fact-checking from multimodal, multi-document datasets. The model takes inputs in the form of documents, images, and a claim, with the objective of assisting in fact-checking tasks. We introduce a dynamic perceiver-based model that can handle inputs from multiple modalities of arbitrary lengths. To train our model, we leverage a novel reinforcement learning-based entailment objective to generate summaries that provide evidence distinguishing between different truthfulness labels. To assess the efficacy of our approach, we conduct experiments on both an existing benchmark and a new dataset of multi-document claims that we contribute. Our approach outperforms the SOTA approach by 4.6% in the claim verification task on the MOCHEG dataset and demonstrates strong performance on our new Multi-News-Fact-Checking dataset.
Authors: Minxuan Zhou, Hao Liang, Tianpeng Li, Zhiyu Wu, Mingan Lin, Linzhuang Sun, Yaqi Zhou, Yan Zhang, Xiaoqin Huang, Yicong Chen, Yujing Qiao, Weipeng Chen, Bin Cui, Wentao Zhang, Zenan Zhou
Abstract: With the development of Multimodal Large Language Models (MLLMs), the evaluation of multimodal models in the context of mathematical problems has become a valuable research field. Multimodal visual-textual mathematical reasoning serves as a critical indicator for evaluating the comprehension and complex multi-step quantitative reasoning abilities of MLLMs. However, previous multimodal math benchmarks have not sufficiently integrated visual and textual information. To address this gap, we proposed MathScape, a new benchmark that emphasizes the understanding and application of combined visual and textual information. MathScape is designed to evaluate photo-based math problem scenarios, assessing the theoretical understanding and application ability of MLLMs through a categorical hierarchical approach. We conduct a multi-dimensional evaluation on 11 advanced MLLMs, revealing that our benchmark is challenging even for the most sophisticated models. By analyzing the evaluation results, we identify the limitations of MLLMs, offering valuable insights for enhancing model performance.
Authors: Gueter Josmy Faure, Jia-Fong Yeh, Min-Hung Chen, Hung-Ting Su, Winston H. Hsu, Shang-Hong Lai
Abstract: Existing research often treats long-form videos as extended short videos, leading to several limitations: inadequate capture of long-range dependencies, inefficient processing of redundant information, and failure to extract high-level semantic concepts. To address these issues, we propose a novel approach that more accurately reflects human cognition. This paper introduces HERMES: temporal-coHERent long-forM understanding with Episodes and Semantics, a model that simulates episodic memory accumulation to capture action sequences and reinforces them with semantic knowledge dispersed throughout the video. Our work makes two key contributions: First, we develop an Episodic COmpressor (ECO) that efficiently aggregates crucial representations from micro to semi-macro levels, overcoming the challenge of long-range dependencies. Second, we propose a Semantics ReTRiever (SeTR) that enhances these aggregated representations with semantic information by focusing on the broader context, dramatically reducing feature dimensionality while preserving relevant macro-level information. This addresses the issues of redundancy and lack of high-level concept extraction. Extensive experiments demonstrate that HERMES achieves state-of-the-art performance across multiple long-video understanding benchmarks in both zero-shot and fully-supervised settings.
Authors: Tzu-Quan Lin, Guan-Ting Lin, Hung-yi Lee, Hao Tang
Abstract: There have been many studies on analyzing self-supervised speech Transformers, in particular, with layer-wise analysis. It is, however, desirable to have an approach that can pinpoint exactly a subset of neurons that is responsible for a particular property of speech, being amenable to model pruning and model editing. In this work, we identify a set of property neurons in the feedforward layers of Transformers to study how speech-related properties, such as phones, gender, and pitch, are stored. When removing neurons of a particular property (a simple form of model editing), the respective downstream performance significantly degrades, showing the importance of the property neurons. We apply this approach to pruning the feedforward layers in Transformers, where most of the model parameters are. We show that protecting property neurons during pruning is significantly more effective than norm-based pruning. The code for identifying property neurons is available at https://github.com/nervjack2/PropertyNeurons.