Authors: Mahendra Gupta, Maitreyee Dutta, Chandresh Kumar Maurya
Abstract: Speech-to-Speech Translation (S2ST) models transform speech from one language to another target language with the same linguistic information. S2ST is important for bridging the communication gap among communities and has diverse applications. In recent years, researchers have introduced direct S2ST models, which have the potential to translate speech without relying on intermediate text generation, have better decoding latency, and the ability to preserve paralinguistic and non-linguistic features. However, direct S2ST has yet to achieve quality performance for seamless communication and still lags behind the cascade models in terms of performance, especially in real-world translation. To the best of our knowledge, no comprehensive survey is available on the direct S2ST system, which beginners and advanced researchers can look upon for a quick survey. The present work provides a comprehensive review of direct S2ST models, data and application issues, and performance metrics. We critically analyze the models' performance over the benchmark datasets and provide research challenges and future directions.
Authors: Jiyeong Kim, Michael L. Chen, Shawheen J. Rezaei, Mariana Ramirez-Posada, Jennifer L. Caswell-Jin, Allison W. Kurian, Fauzia Riaz, Kavita Y. Sarin, Jean Y. Tang, Steven M. Asch, Eleni Linos
Abstract: Patient-centered research is increasingly important in narrowing the gap between research and patient care, yet incorporating patient perspectives into health research has been inconsistent. We propose an automated framework leveraging innovative natural language processing (NLP) and artificial intelligence (AI) with patient portal messages to generate research ideas that prioritize important patient issues. We further quantified the quality of AI-generated research topics. To define patient clinical concerns, we analyzed 614,464 patient messages from 25,549 individuals with breast or skin cancer obtained from a large academic hospital (2013 to 2024), constructing a 2-staged unsupervised NLP topic model. Then, we generated research topics to resolve the defined issues using a widely used AI (ChatGPT-4o, OpenAI Inc, April 2024 version) with prompt-engineering strategies. We guided AI to perform multi-level tasks: 1) knowledge interpretation and summarization (e.g., interpreting and summarizing the NLP-defined topics), 2) knowledge generation (e.g., generating research ideas corresponding to patients issues), 3) self-reflection and correction (e.g., ensuring and revising the research ideas after searching for scientific articles), and 4) self-reassurance (e.g., confirming and finalizing the research ideas). Six highly experienced breast oncologists and dermatologists assessed the significance and novelty of AI-generated research topics using a 5-point Likert scale (1-exceptional, 5-poor). One-third of the AI-suggested research topics were highly significant and novel when both scores were lower than the average. Two-thirds of the AI-suggested topics were novel in both cancers. Our findings demonstrate that AI-generated research topics reflecting patient perspectives via a large volume of patient messages can meaningfully guide future directions in patient-centered health research.
Authors: Zhangchi Qiu, Linhao Luo, Shirui Pan, Alan Wee-Chung Liew
Abstract: Conversational Recommender Systems (CRSs) aim to provide personalized recommendations through dynamically capturing user preferences in interactive conversations. Conventional CRSs often extract user preferences as hidden representations, which are criticized for their lack of interpretability. This diminishes the transparency and trustworthiness of the recommendation process. Recent works have explored combining the impressive capabilities of Large Language Models (LLMs) with the domain-specific knowledge of Knowledge Graphs (KGs) to generate human-understandable recommendation explanations. Despite these efforts, the integration of LLMs and KGs for CRSs remains challenging due to the modality gap between unstructured dialogues and structured KGs. Moreover, LLMs pre-trained on large-scale corpora may not be well-suited for analyzing user preferences, which require domain-specific knowledge. In this paper, we propose COMPASS, a plug-and-play framework that synergizes LLMs and KGs to unveil user preferences, enhancing the performance and explainability of existing CRSs. To address integration challenges, COMPASS employs a two-stage training approach: first, it bridges the gap between the structured KG and natural language through an innovative graph entity captioning pre-training mechanism. This enables the LLM to transform KG entities into concise natural language descriptions, allowing them to comprehend domain-specific knowledge. Following, COMPASS optimizes user preference modeling via knowledge-aware instruction fine-tuning, where the LLM learns to reason and summarize user preferences from both dialogue histories and KG-augmented context. This enables COMPASS to perform knowledge-aware reasoning and generate comprehensive and interpretable user preferences that can seamlessly integrate with existing CRS models for improving recommendation performance and explainability.
Authors: Yao Xu, Shizhu He, Zeng Xiangrong, Jiabei Chen, Guang Liu, Bingning Wang, Jun Zhao, Kang Liu
Abstract: Structured data, such as tables, graphs, and databases, play a critical role in plentiful NLP tasks such as question answering and dialogue system. Recently, inspired by Vision-Language Models, Graph Neutral Networks (GNNs) have been introduced as an additional modality into the input of Large Language Models (LLMs) to improve their performance on Structured Knowledge Grounding (SKG) tasks. However, those GNN-enhanced LLMs have the following limitations: (1) They employ diverse GNNs to model varying types of structured data, rendering them unable to uniformly process various forms of structured data. (2) The pretraining of GNNs is coupled with specific LLMs, which prevents GNNs from fully aligning with the textual space and limits their adaptability to other LLMs. To address these issues, we propose \textbf{L}arge \textbf{L}anguage and \textbf{S}tructured Data \textbf{A}ssistant (LLaSA), a general framework for enhancing LLMs' ability to handle structured data. Specifically, we represent various types of structured data in a unified hypergraph format, and use self-supervised learning to pretrain a hypergraph encoder, and a G-Former compressing encoded hypergraph representations with cross-attention. The compressed hypergraph representations are appended to the serialized inputs during training and inference stages of LLMs. Experimental results on multiple SKG tasks show that our pretrained hypergraph encoder can adapt to various LLMs and enhance their ability to process different types of structured data. Besides, LLaSA, with LoRA fine-tuning, outperforms previous SOTA method using full parameters tuning.
Authors: Shaochen Xu, Yifan Zhou, Zhengliang Liu, Zihao Wu, Tianyang Zhong, Huaqin Zhao, Yiwei Li, Hanqi Jiang, Yi Pan, Junhao Chen, Jin Lu, Wei Zhang, Tuo Zhang, Lu Zhang, Dajiang Zhu, Xiang Li, Wei Liu, Quanzheng Li, Andrea Sikora, Xiaoming Zhai, Zhen Xiang, Tianming Liu
Abstract: Artificial Intelligence (AI) has become essential in modern healthcare, with large language models (LLMs) offering promising advances in clinical decision-making. Traditional model-based approaches, including those leveraging in-context demonstrations and those with specialized medical fine-tuning, have demonstrated strong performance in medical language processing but struggle with real-time adaptability, multi-step reasoning, and handling complex medical tasks. Agent-based AI systems address these limitations by incorporating reasoning traces, tool selection based on context, knowledge retrieval, and both short- and long-term memory. These additional features enable the medical AI agent to handle complex medical scenarios where decision-making should be built on real-time interaction with the environment. Therefore, unlike conventional model-based approaches that treat medical queries as isolated questions, medical AI agents approach them as complex tasks and behave more like human doctors. In this paper, we study the choice of the backbone LLM for medical AI agents, which is the foundation for the agent's overall reasoning and action generation. In particular, we consider the emergent o1 model and examine its impact on agents' reasoning, tool-use adaptability, and real-time information retrieval across diverse clinical scenarios, including high-stakes settings such as intensive care units (ICUs). Our findings demonstrate o1's ability to enhance diagnostic accuracy and consistency, paving the way for smarter, more responsive AI tools that support better patient outcomes and decision-making efficacy in clinical practice.
Authors: Christopher Gerling, Stefan Lessmann
Abstract: This paper explores the growing impact of AI and NLP in bank marketing, highlighting their evolving roles in enhancing marketing strategies, improving customer engagement, and creating value within this sector. While AI and NLP have been widely studied in general marketing, there is a notable gap in understanding their specific applications and potential within the banking sector. This research addresses this specific gap by providing a systematic review and strategic analysis of AI and NLP applications in bank marketing, focusing on their integration across the customer journey and operational excellence. Employing the PRISMA methodology, this study systematically reviews existing literature to assess the current landscape of AI and NLP in bank marketing. Additionally, it incorporates semantic mapping using Sentence Transformers and UMAP for strategic gap analysis to identify underexplored areas and opportunities for future research. The systematic review reveals limited research specifically focused on NLP applications in bank marketing. The strategic gap analysis identifies key areas where NLP can further enhance marketing strategies, including customer-centric applications like acquisition, retention, and personalized engagement, offering valuable insights for both academic research and practical implementation. This research contributes to the field of bank marketing by mapping the current state of AI and NLP applications and identifying strategic gaps. The findings provide actionable insights for developing NLP-driven growth and innovation frameworks and highlight the role of NLP in improving operational efficiency and regulatory compliance. This work has broader implications for enhancing customer experience, profitability, and innovation in the banking industry.
Authors: Elizaveta Reganova, Peter Steinbach
Abstract: Large Language Models (LLMs) have gained significant popularity in recent years for their ability to answer questions in various fields. However, these models have a tendency to "hallucinate" their responses, making it challenging to evaluate their performance. A major challenge is determining how to assess the certainty of a model's predictions and how it correlates with accuracy. In this work, we introduce an analysis for evaluating the performance of popular open-source LLMs, as well as gpt-3.5 Turbo, on multiple choice physics questionnaires. We focus on the relationship between answer accuracy and variability in topics related to physics. Our findings suggest that most models provide accurate replies in cases where they are certain, but this is by far not a general behavior. The relationship between accuracy and uncertainty exposes a broad horizontal bell-shaped distribution. We report how the asymmetry between accuracy and uncertainty intensifies as the questions demand more logical reasoning of the LLM agent, while the same relationship remains sharp for knowledge retrieval tasks.
Authors: Jie Zou, Jimmy Xiangji Huang, Zhaochun Ren, Evangelos Kanoulas
Abstract: Online shopping platforms, such as Amazon and AliExpress, are increasingly prevalent in society, helping customers purchase products conveniently. With recent progress in natural language processing, researchers and practitioners shift their focus from traditional product search to conversational product search. Conversational product search enables user-machine conversations and through them collects explicit user feedback that allows to actively clarify the users' product preferences. Therefore, prospective research on an intelligent shopping assistant via conversations is indispensable. Existing publications on conversational product search either model conversations independently from users, queries, and products or lead to a vocabulary mismatch. In this work, we propose a new conversational product search model, ConvPS, to assist users in locating desirable items. The model is first trained to jointly learn the semantic representations of user, query, item, and conversation via a unified generative framework. After learning these representations, they are integrated to retrieve the target items in the latent semantic space. Meanwhile, we propose a set of greedy and explore-exploit strategies to learn to ask the user a sequence of high-performance questions for conversations. Our proposed ConvPS model can naturally integrate the representation learning of the user, query, item, and conversation into a unified generative framework, which provides a promising avenue for constructing accurate and robust conversational product search systems that are flexible and adaptive. Experimental results demonstrate that our ConvPS model significantly outperforms state-of-the-art baselines.
Authors: Xinhua Wu, Qi R. Wang
Abstract: As large language models (LLMs) are increasingly applied in areas influencing societal outcomes, it is critical to understand their tendency to perpetuate and amplify biases. This study investigates whether LLMs exhibit biases in predicting human mobility -- a fundamental human behavior -- based on race and gender. Using three prominent LLMs -- GPT-4, Gemini, and Claude -- we analyzed their predictions of visitations to points of interest (POIs) for individuals, relying on prompts that included names with and without explicit demographic details. We find that LLMs frequently reflect and amplify existing societal biases. Specifically, predictions for minority groups were disproportionately skewed, with these individuals being significantly less likely to be associated with wealth-related points of interest (POIs). Gender biases were also evident, as female individuals were consistently linked to fewer career-related POIs compared to their male counterparts. These biased associations suggest that LLMs not only mirror but also exacerbate societal stereotypes, particularly in contexts involving race and gender.
Authors: Erica Ann Metheney, Lauren Yehle
Abstract: This paper explores and assesses in what ways generative AI can assist in translating survey instruments. Writing effective survey questions is a challenging and complex task, made even more difficult for surveys that will be translated and deployed in multiple linguistic and cultural settings. Translation errors can be detrimental, with known errors rendering data unusable for its intended purpose and undetected errors leading to incorrect conclusions. A growing number of institutions face this problem as surveys deployed by private and academic organizations globalize, and the success of their current efforts depends heavily on researchers' and translators' expertise and the amount of time each party has to contribute to the task. Thus, multilinguistic and multicultural surveys produced by teams with limited expertise, budgets, or time are at significant risk for translation-based errors in their data. We implement a zero-shot prompt experiment using ChatGPT to explore generative AI's ability to identify features of questions that might be difficult to translate to a linguistic audience other than the source language. We find that ChatGPT can provide meaningful feedback on translation issues, including common source survey language, inconsistent conceptualization, sensitivity and formality issues, and nonexistent concepts. In addition, we provide detailed information on the practicality of the approach, including accessing the necessary software, associated costs, and computational run times. Lastly, based on our findings, we propose avenues for future research that integrate AI into survey translation practices.
Authors: Cau\~a Ferreira Barros, Bruna Borges Azevedo, Valdemar Vicente Graciano Neto, Mohamad Kassab, Marcos Kalinowski, Hugo Alexandre D. do Nascimento, Michelle C. G. S. P. Bandeira
Abstract: The exponential growth of text-based data in domains such as healthcare, education, and social sciences has outpaced the capacity of traditional qualitative analysis methods, which are time-intensive and prone to subjectivity. Large Language Models (LLMs), powered by advanced generative AI, have emerged as transformative tools capable of automating and enhancing qualitative analysis. This study systematically maps the literature on the use of LLMs for qualitative research, exploring their application contexts, configurations, methodologies, and evaluation metrics. Findings reveal that LLMs are utilized across diverse fields, demonstrating the potential to automate processes traditionally requiring extensive human input. However, challenges such as reliance on prompt engineering, occasional inaccuracies, and contextual limitations remain significant barriers. This research highlights opportunities for integrating LLMs with human expertise, improving model robustness, and refining evaluation methodologies. By synthesizing trends and identifying research gaps, this study aims to guide future innovations in the application of LLMs for qualitative analysis.
Authors: Zongrong Li, Junhao Xu, Siqin Wang, Yifan Wu, Haiyang Li
Abstract: Geospatial predictions are crucial for diverse fields such as disaster management, urban planning, and public health. Traditional machine learning methods often face limitations when handling unstructured or multi-modal data like street view imagery. To address these challenges, we propose StreetViewLLM, a novel framework that integrates a large language model with the chain-of-thought reasoning and multimodal data sources. By combining street view imagery with geographic coordinates and textual data, StreetViewLLM improves the precision and granularity of geospatial predictions. Using retrieval-augmented generation techniques, our approach enhances geographic information extraction, enabling a detailed analysis of urban environments. The model has been applied to seven global cities, including Hong Kong, Tokyo, Singapore, Los Angeles, New York, London, and Paris, demonstrating superior performance in predicting urban indicators, including population density, accessibility to healthcare, normalized difference vegetation index, building height, and impervious surface. The results show that StreetViewLLM consistently outperforms baseline models, offering improved predictive accuracy and deeper insights into the built environment. This research opens new opportunities for integrating the large language model into urban analytics, decision-making in urban planning, infrastructure management, and environmental monitoring.
Authors: Yuze Liu, Tingjie Liu, Tiehua Zhang, Youhua Xia, Jinze Wang, Zhishu Shen, Jiong Jin, Fei Richard Yu
Abstract: Large language models (LLMs) have demonstrated impressive success in a wide range of natural language processing (NLP) tasks due to their extensive general knowledge of the world. Recent works discovered that the performance of LLMs is heavily dependent on the input prompt. However, prompt engineering is usually done manually in a trial-and-error fashion, which can be labor-intensive and challenging in order to find the optimal prompts. To address these problems and unleash the utmost potential of LLMs, we propose a novel LLMs-agnostic framework for prompt optimization, namely GRL-Prompt, which aims to automatically construct optimal prompts via reinforcement learning (RL) in an end-to-end manner. To provide structured action/state representation for optimizing prompts, we construct a knowledge graph (KG) that better encodes the correlation between the user query and candidate in-context examples. Furthermore, a policy network is formulated to generate the optimal action by selecting a set of in-context examples in a rewardable order to construct the prompt. Additionally, the embedding-based reward shaping is utilized to stabilize the RL training process. The experimental results show that GRL-Prompt outperforms recent state-of-the-art methods, achieving an average increase of 0.10 in ROUGE-1, 0.07 in ROUGE-2, 0.07 in ROUGE-L, and 0.05 in BLEU.
Authors: Roland Daynauth, Christopher Clarke, Krisztian Flautner, Lingjia Tang, Jason Mars
Abstract: Deciding which large language model (LLM) to use is a complex challenge. Pairwise ranking has emerged as a new method for evaluating human preferences for LLMs. This approach entails humans evaluating pairs of model outputs based on a predefined criterion. By collecting these comparisons, a ranking can be constructed using methods such as Elo. However, applying these algorithms as constructed in the context of LLM evaluation introduces several challenges. In this paper, we explore the effectiveness of ranking systems for head-to-head comparisons of LLMs. We formally define a set of fundamental principles for effective ranking and conduct a series of extensive evaluations on the robustness of several ranking algorithms in the context of LLMs. Our analysis uncovers key insights into the factors that affect ranking accuracy and efficiency, offering guidelines for selecting the most appropriate methods based on specific evaluation contexts and resource constraints.
Authors: Atharva Gundawar, Karthik Valmeekam, Mudit Verma, Subbarao Kambhampati
Abstract: Previous work has attempted to boost Large Language Model (LLM) performance on planning and scheduling tasks through a variety of prompt engineering techniques. While these methods can work within the distributions tested, they are neither robust nor predictable. This limitation can be addressed through compound LLM architectures where LLMs work in conjunction with other components to ensure reliability. In this paper, we present a technical evaluation of a compound LLM architecture--the LLM-Modulo framework. In this framework, an LLM is paired with a complete set of sound verifiers that validate its output, re-prompting it if it fails. This approach ensures that the system can never output any fallacious output, and therefore that every output generated is guaranteed correct--something previous techniques have not been able to claim. Our results, evaluated across four scheduling domains, demonstrate significant performance gains with the LLM-Modulo framework using various models. Additionally, we explore modifications to the base configuration of the framework and assess their impact on overall system performance.
Authors: Moritz Rietschel, Fang Guo, Kyle Steinfeld
Abstract: Here is an updated version of your abstract, cleaned for submission to arXiv with potential "bad characters" corrected to conform to ASCII standards: Architects adopt visual scripting and parametric design tools to explore more expansive design spaces (Coates, 2010), refine their thinking about the geometric logic of their design (Woodbury, 2010), and overcome conventional software limitations (Burry, 2011). Despite two decades of effort to make design scripting more accessible, a disconnect between a designer's free ways of thinking and the rigidity of algorithms remains (Burry, 2011). Recent developments in Large Language Models (LLMs) suggest this might soon change, as LLMs encode a general understanding of human context and exhibit the capacity to produce geometric logic. This project speculates that if LLMs can effectively mediate between user intent and algorithms, they become a powerful tool to make scripting in design more widespread and fun. We explore if such systems can interpret natural language prompts to assemble geometric operations relevant to computational design scripting. In the system, multiple layers of LLM agents are configured with specific context to infer the user intent and construct a sequential logic. Given a user's high-level text prompt, a geometric description is created, distilled into a sequence of logic operations, and mapped to software-specific commands. The completed script is constructed in the user's visual programming interface. The system succeeds in generating complete visual scripts up to a certain complexity but fails beyond this complexity threshold. It shows how LLMs can make design scripting much more aligned with human creativity and thought. Future research should explore conversational interactions, expand to multimodal inputs and outputs, and assess the performance of these tools.
Authors: David Noever, Forrest McKee
Abstract: This research introduces a novel evaluation framework designed to assess large language models' (LLMs) ability to acknowledge uncertainty on 675 fundamentally unsolvable problems. Using a curated dataset of graduate-level grand challenge questions with intentionally unknowable answers, we evaluated twelve state-of-the-art LLMs, including both open and closed-source models, on their propensity to admit ignorance rather than generate plausible but incorrect responses. The best models scored in 62-68% accuracy ranges for admitting the problem solution was unknown in fields ranging from biology to philosophy and mathematics. We observed an inverse relationship between problem difficulty and model accuracy, with GPT-4 demonstrating higher rates of uncertainty acknowledgment on more challenging problems (35.8%) compared to simpler ones (20.0%). This pattern indicates that models may be more prone to generate speculative answers when problems appear more tractable. The study also revealed significant variations across problem categories, with models showing difficulty in acknowledging uncertainty in invention and NP-hard problems while performing relatively better on philosophical and psychological challenges. These results contribute to the growing body of research on artificial general intelligence (AGI) assessment by highlighting the importance of uncertainty recognition as a critical component of future machine intelligence evaluation. This impossibility test thus extends previous theoretical frameworks for universal intelligence testing by providing empirical evidence of current limitations in LLMs' ability to recognize their own knowledge boundaries, suggesting new directions for improving model training architectures and evaluation approaches.
Authors: Yifan Yang, Qiao Jin, Robert Leaman, Xiaoyu Liu, Guangzhi Xiong, Maame Sarfo-Gyamfi, Changlin Gong, Santiago Ferri\`ere-Steinert, W. John Wilbur, Xiaojun Li, Jiaxin Yuan, Bang An, Kelvin S. Castro, Francisco Erramuspe \'Alvarez, Mat\'ias Stockle, Aidong Zhang, Furong Huang, Zhiyong Lu
Abstract: The remarkable capabilities of Large Language Models (LLMs) make them increasingly compelling for adoption in real-world healthcare applications. However, the risks associated with using LLMs in medical applications have not been systematically characterized. We propose using five key principles for safe and trustworthy medical AI: Truthfulness, Resilience, Fairness, Robustness, and Privacy, along with ten specific aspects. Under this comprehensive framework, we introduce a novel MedGuard benchmark with 1,000 expert-verified questions. Our evaluation of 11 commonly used LLMs shows that the current language models, regardless of their safety alignment mechanisms, generally perform poorly on most of our benchmarks, particularly when compared to the high performance of human physicians. Despite recent reports indicate that advanced LLMs like ChatGPT can match or even exceed human performance in various medical tasks, this study underscores a significant safety gap, highlighting the crucial need for human oversight and the implementation of AI safety guardrails.
Authors: Hang Zhou, Xiaoxu Zheng, Yunhe Wang, Michael Bi Mi, Deyi Xiong, Kai Han
Abstract: Recurrent neural network (RNNs) that are capable of modeling long-distance dependencies are widely used in various speech tasks, eg., keyword spotting (KWS) and speech enhancement (SE). Due to the limitation of power and memory in low-resource devices, efficient RNN models are urgently required for real-world applications. In this paper, we propose an efficient RNN architecture, GhostRNN, which reduces hidden state redundancy with cheap operations. In particular, we observe that partial dimensions of hidden states are similar to the others in trained RNN models, suggesting that redundancy exists in specific RNNs. To reduce the redundancy and hence computational cost, we propose to first generate a few intrinsic states, and then apply cheap operations to produce ghost states based on the intrinsic states. Experiments on KWS and SE tasks demonstrate that the proposed GhostRNN significantly reduces the memory usage (~40%) and computation cost while keeping performance similar.
Authors: Jing Yi Wang, Nicholas Sukiennik, Tong Li, Weikang Su, Qianyue Hao, Jingbo Xu, Zihan Huang, Fengli Xu, Yong Li
Abstract: The rapid evolution of large language models (LLMs) and their capacity to simulate human cognition and behavior has given rise to LLM-based frameworks and tools that are evaluated and applied based on their ability to perform tasks traditionally performed by humans, namely those involving cognition, decision-making, and social interaction. This survey provides a comprehensive examination of such human-centric LLM capabilities, focusing on their performance in both individual tasks (where an LLM acts as a stand-in for a single human) and collective tasks (where multiple LLMs coordinate to mimic group dynamics). We first evaluate LLM competencies across key areas including reasoning, perception, and social cognition, comparing their abilities to human-like skills. Then, we explore real-world applications of LLMs in human-centric domains such as behavioral science, political science, and sociology, assessing their effectiveness in replicating human behaviors and interactions. Finally, we identify challenges and future research directions, such as improving LLM adaptability, emotional intelligence, and cultural sensitivity, while addressing inherent biases and enhancing frameworks for human-AI collaboration. This survey aims to provide a foundational understanding of LLMs from a human-centric perspective, offering insights into their current capabilities and potential for future development.
Authors: Muhammad Sharif, Zeeshan Abbas, Jiangyan Yi, Chenglin Liu
Abstract: Automatic Speech Recognition (ASR) technology has witnessed significant advancements in recent years, revolutionizing human-computer interactions. While major languages have benefited from these developments, lesser-resourced languages like Urdu face unique challenges. This paper provides an extensive exploration of the dynamic landscape of ASR research, focusing particularly on the resource-constrained Urdu language, which is widely spoken across South Asian nations. It outlines current research trends, technological advancements, and potential directions for future studies in Urdu ASR, aiming to pave the way for forthcoming researchers interested in this domain. By leveraging contemporary technologies, analyzing existing datasets, and evaluating effective algorithms and tools, the paper seeks to shed light on the unique challenges and opportunities associated with Urdu language processing and its integration into the broader field of speech research.
Authors: Hang Zhou, Yehui Tang, Haochen Qin, Yujie Yang, Renren Jin, Deyi Xiong, Kai Han, Yunhe Wang
Abstract: The efficacy of large language models (LLMs) on downstream tasks usually hinges on instruction tuning, which relies critically on the quality of training data. Unfortunately, collecting high-quality and diverse data is both expensive and time-consuming. To mitigate this issue, we propose a novel Star-Agents framework, which automates the enhancement of data quality across datasets through multi-agent collaboration and assessment. The framework adopts a three-pronged strategy. It initially generates diverse instruction data with multiple LLM agents through a bespoke sampling method. Subsequently, the generated data undergo a rigorous evaluation using a dual-model method that assesses both difficulty and quality. Finaly, the above process evolves in a dynamic refinement phase, where more effective LLMs are prioritized, enhancing the overall data quality. Our empirical studies, including instruction tuning experiments with models such as Pythia and LLaMA, demonstrate the effectiveness of the proposed framework. Optimized datasets have achieved substantial improvements, with an average increase of 12% and notable gains in specific metrics, such as a 40% improvement in Fermi, as evidenced by benchmarks like MT-bench, Vicuna bench, and WizardLM testset.
Authors: Jingtao Ding, Yunke Zhang, Yu Shang, Yuheng Zhang, Zefang Zong, Jie Feng, Yuan Yuan, Hongyuan Su, Nian Li, Nicholas Sukiennik, Fengli Xu, Yong Li
Abstract: The concept of world models has garnered significant attention due to advancements in multimodal large language models such as GPT-4 and video generation models such as Sora, which are central to the pursuit of artificial general intelligence. This survey offers a comprehensive review of the literature on world models. Generally, world models are regarded as tools for either understanding the present state of the world or predicting its future dynamics. This review presents a systematic categorization of world models, emphasizing two primary functions: (1) constructing internal representations to understand the mechanisms of the world, and (2) predicting future states to simulate and guide decision-making. Initially, we examine the current progress in these two categories. We then explore the application of world models in key domains, including autonomous driving, robotics, and social simulacra, with a focus on how each domain utilizes these aspects. Finally, we outline key challenges and provide insights into potential future research directions.
Authors: Qingquan Zhang, Qiqi Duan, Bo Yuan, Yuhui Shi, Jialin Liu
Abstract: Large Language Models (LLMs) have made significant strides in the field of artificial intelligence, showcasing their ability to interact with humans and influence human cognition through information dissemination. However, recent studies have brought to light instances of bias inherent within these LLMs, presenting a critical issue that demands attention. In our research, we delve deeper into the intricate challenge of harmonising accuracy and fairness in the enhancement of LLMs. While improving accuracy can indeed enhance overall LLM performance, it often occurs at the expense of fairness. Overemphasising optimisation of one metric invariably leads to a significant degradation of the other. This underscores the necessity of taking into account multiple considerations during the design and optimisation phases of LLMs. Therefore, we advocate for reformulating the LLM training process as a multi-objective learning task. Our investigation reveals that multi-objective evolutionary learning (MOEL) methodologies offer promising avenues for tackling this challenge. Our MOEL framework enables the simultaneous optimisation of both accuracy and fairness metrics, resulting in a Pareto-optimal set of LLMs. In summary, our study sheds valuable lights on the delicate equilibrium between accuracy and fairness within LLMs, which is increasingly significant for their real-world applications. By harnessing MOEL, we present a promising pathway towards fairer and more efficacious AI technologies.
Authors: Arthur Elwing Torres, Edleno Silva de Moura, Altigran Soares da Silva, Mario A. Nascimento, Filipe Mesquita
Abstract: Named Entity Recognition (NER) is a machine learning task that traditionally relies on supervised learning and annotated data. Acquiring such data is often a challenge, particularly in specialized fields like medical, legal, and financial sectors. Those are commonly referred to as low-resource domains, which comprise long-tail entities, due to the scarcity of available data. To address this, data augmentation techniques are increasingly being employed to generate additional training instances from the original dataset. In this study, we evaluate the effectiveness of two prominent text augmentation techniques, Mention Replacement and Contextual Word Replacement, on two widely-used NER models, Bi-LSTM+CRF and BERT. We conduct experiments on four datasets from low-resource domains, and we explore the impact of various combinations of training subset sizes and number of augmented examples. We not only confirm that data augmentation is particularly beneficial for smaller datasets, but we also demonstrate that there is no universally optimal number of augmented examples, i.e., NER practitioners must experiment with different quantities in order to fine-tune their projects.
Authors: Yurii Paniv, Artur Kiulian, Dmytro Chaplynskyi, Mykola Khandoga, Anton Polishko, Tetiana Bas, Guillermo Gabrielli
Abstract: While the evaluation of multimodal English-centric models is an active area of research with numerous benchmarks, there is a profound lack of benchmarks or evaluation suites for low- and mid-resource languages. We introduce ZNO-Vision, a comprehensive multimodal Ukrainian-centric benchmark derived from standardized university entrance examination (ZNO). The benchmark consists of over 4,300 expert-crafted questions spanning 12 academic disciplines, including mathematics, physics, chemistry, and humanities. We evaluated the performance of both open-source models and API providers, finding that only a handful of models performed above baseline. Alongside the new benchmark, we performed the first evaluation study of multimodal text generation for the Ukrainian language: we measured caption generation quality on the Multi30K-UK dataset, translated the VQA benchmark into Ukrainian, and measured performance degradation relative to original English versions. Lastly, we tested a few models from a cultural perspective on knowledge of national cuisine. We believe our work will advance multimodal generation capabilities for the Ukrainian language and our approach could be useful for other low-resource languages.
Authors: Jinming Xing, Ruilin Xing, Yan Sun
Abstract: Large Language Models (LLMs) have revolutionized natural language processing (NLP) by delivering state-of-the-art performance across a variety of tasks. Among these, Transformer-based models like BERT and GPT rely on pooling layers to aggregate token-level embeddings into sentence-level representations. Common pooling mechanisms such as Mean, Max, and Weighted Sum play a pivotal role in this aggregation process. Despite their widespread use, the comparative performance of these strategies on different LLM architectures remains underexplored. To address this gap, this paper investigates the effects of these pooling mechanisms on two prominent LLM families -- BERT and GPT, in the context of sentence-level sentiment analysis. Comprehensive experiments reveal that each pooling mechanism exhibits unique strengths and weaknesses depending on the task's specific requirements. Our findings underline the importance of selecting pooling methods tailored to the demands of particular applications, prompting a re-evaluation of common assumptions regarding pooling operations. By offering actionable insights, this study contributes to the optimization of LLM-based models for downstream tasks.
Authors: Pittawat Taveekitworachai, Chollakorn Nimpattanavong, Mustafa Can Gursesli, Antonio Lanata, Andrea Guazzini, Ruck Thawonmas
Abstract: This paper presents Dynamic Context Prompting/Programming (DCP/P), a novel framework for interacting with LLMs to generate graph-based content with a dynamic context window history. While there is an existing study utilizing LLMs to generate a visual novel game, the previous study involved a manual process of output extraction and did not provide flexibility in generating a longer, coherent story. We evaluate DCP/P against our baseline, which does not provide context history to an LLM and only relies on the initial story data. Through objective evaluation, we show that simply providing the LLM with a summary leads to a subpar story compared to additionally providing the LLM with the proper context of the story. We also provide an extensive qualitative analysis and discussion. We qualitatively examine the quality of the objectively best-performing generated game from each approach. In addition, we examine biases in word choices and word sentiment of the generated content. We find a consistent observation with previous studies that LLMs are biased towards certain words, even with a different LLM family. Finally, we provide a comprehensive discussion on opportunities for future studies.
Authors: Xunyu Zhu, Jian Li, Can Ma, Weiping Wang
Abstract: Large Language Models (LLMs) demonstrate exceptional reasoning capabilities, often achieving state-of-the-art performance in various tasks. However, their substantial computational and memory demands, due to billions of parameters, hinder deployment in resource-constrained environments. A promising solution is knowledge distillation, where LLMs transfer reasoning capabilities to Small Language Models (SLMs, $\le$ 1B parameters), enabling wider deployment on low-resource devices. Existing methods primarily focus on generating high-quality reasoning rationales for distillation datasets but often neglect the critical role of data quantity and quality. To address these challenges, we propose a Feedback-Driven Distillation (FDD) framework to enhance SLMs' mathematical reasoning capabilities. In the initialization stage, a distillation dataset is constructed by prompting LLMs to pair mathematical problems with corresponding reasoning rationales. We classify problems into easy and hard categories based on SLM performance. For easy problems, LLMs generate more complex variations, while for hard problems, new questions of similar complexity are synthesized. In addition, we propose a multi-round distillation paradigm to iteratively enrich the distillation datasets, thereby progressively improving the mathematical reasoning abilities of SLMs. Experimental results demonstrate that our method can make SLMs achieve SOTA mathematical reasoning performance.
Authors: Luhang Sun, Varsha Pendyala, Yun-Shiuan Chuang, Shanglin Yang, Jonathan Feldman, Andrew Zhao, Munmun De Choudhury, Sijia Yang, Dhavan Shah
Abstract: This paper leverages large-language models (LLMs) to experimentally determine optimal strategies for scaling up social media content annotation for stance detection on HPV vaccine-related tweets. We examine both conventional fine-tuning and emergent in-context learning methods, systematically varying strategies of prompt engineering across widely used LLMs and their variants (e.g., GPT4, Mistral, and Llama3, etc.). Specifically, we varied prompt template design, shot sampling methods, and shot quantity to detect stance on HPV vaccination. Our findings reveal that 1) in general, in-context learning outperforms fine-tuning in stance detection for HPV vaccine social media content; 2) increasing shot quantity does not necessarily enhance performance across models; and 3) different LLMs and their variants present differing sensitivity to in-context learning conditions. We uncovered that the optimal in-context learning configuration for stance detection on HPV vaccine tweets involves six stratified shots paired with detailed contextual prompts. This study highlights the potential and provides an applicable approach for applying LLMs to research on social media stance and skepticism detection.
Authors: Jiatong Li, Yunqing Liu, Wei Liu, Jingdi Le, Di Zhang, Wenqi Fan, Dongzhan Zhou, Yuqiang Li, Qing Li
Abstract: Molecule discovery is a pivotal research field, impacting everything from the medicines we take to the materials we use. Recently, Large Language Models (LLMs) have been widely adopted in molecule understanding and generation, yet the alignments between molecules and their corresponding captions remain a significant challenge. Previous endeavours often treat the molecule as a general SMILES string or molecular graph, neglecting the fine-grained alignments between the molecular sub-structures and the descriptive textual phrases, which are crucial for accurate and explainable predictions. In this case, we introduce MolReFlect, a novel teacher-student framework designed to contextually perform the molecule-caption alignments in a fine-grained way. Our approach initially leverages a larger teacher LLM to label the detailed alignments by directly extracting critical phrases from molecule captions or SMILES strings and implying them to corresponding sub-structures or characteristics. To refine these alignments, we propose In-Context Selective Reflection, which retrieves previous extraction results as context examples for teacher LLM to reflect and lets a smaller student LLM select from in-context reflection and previous extraction results. Finally, we enhance the learning process of the student LLM through Chain-of-Thought In-Context Molecule Tuning, integrating the fine-grained alignments and the reasoning processes within the Chain-of-Thought format. Our experimental results demonstrate that MolReFlect enables LLMs like Mistral-7B to significantly outperform the previous baselines, achieving SOTA performance on the ChEBI-20 dataset. This advancement not only enhances the generative capabilities of LLMs in the molecule-caption translation task, but also contributes to a more explainable framework.
Authors: Jiashuo Liang, Guancheng Li, Yang Yu
Abstract: Large language models (LLMs) have been widely adopted in applications such as automated content generation and even critical decision-making systems. However, the risk of prompt injection allows for potential manipulation of LLM outputs. While numerous attack methods have been documented, achieving full control over these outputs remains challenging, often requiring experienced attackers to make multiple attempts and depending heavily on the prompt context. Recent advancements in gradient-based white-box attack techniques have shown promise in tasks like jailbreaks and system prompt leaks. Our research generalizes gradient-based attacks to find a trigger that is (1) Universal: effective irrespective of the target output; (2) Context-Independent: robust across diverse prompt contexts; and (3) Precise Output: capable of manipulating LLM inputs to yield any specified output with high accuracy. We propose a novel method to efficiently discover such triggers and assess the effectiveness of the proposed attack. Furthermore, we discuss the substantial threats posed by such attacks to LLM-based applications, highlighting the potential for adversaries to taking over the decisions and actions made by AI agents.
Authors: Zheni Zeng, Yuxuan Chen, Shi Yu, Yukun Yan, Zhenghao Liu, Shuo Wang, Xu Han, Zhiyuan Liu, Maosong Sun
Abstract: Humans can utilize techniques to quickly acquire knowledge from specific materials in advance, such as creating self-assessment questions, enabling us to achieving related tasks more efficiently. In contrast, large language models (LLMs) usually relies on retrieval-augmented generation to exploit knowledge materials in an instant manner, or requires external signals such as human preference data and stronger LLM annotations to conduct knowledge adaptation. To unleash the self-learning potential of LLMs, we propose KBAda, an approach designed for efficient adaptation to downstream tasks involving knowledge bases. Our method utilizes iterative training with self-annotated data such as Q&A pairs and revision suggestions, enabling the model to grasp the knowledge content efficiently. Experimental results on multiple datasets demonstrate the effectiveness of our approach, significantly boosting model performance in downstream tasks that require specific knowledge at a low cost. Notably, our approach achieves over 90% of the performance improvement that can be obtained by using GPT-4-turbo annotation, while relying entirely on self-supervision. We release our experimental data, models, and process analyses to the community for further exploration (https://github.com/thunlp/KBAda).
Authors: Haitao Li, Ziyu Li, Yiheng Mao, Ziyi Liu, Zhoujian Sun, Zhengxing Huang
Abstract: Multimodal large language models (MLLMs) are increasingly being applied in the medical field, particularly in medical imaging. However, developing MLLMs for ECG signals, which are crucial in clinical settings, has been a significant challenge beyond medical imaging. Previous studies have attempted to address this by converting ECGs into several text tags using an external classifier in a training-free manner. However, this approach significantly compresses the information in ECGs and underutilizes the reasoning capabilities of LLMs. In this work, we directly feed the embeddings of ECGs into the LLM through a projection layer, retaining more information about ECGs and better leveraging the reasoning abilities of LLMs. Our method can also effectively handle a common situation in clinical practice where it is necessary to compare two ECGs taken at different times. Recent studies found that MLLMs may rely solely on text input to provide answers, ignoring inputs from other modalities. We analyzed this phenomenon from a causal perspective in the context of ECG MLLMs and discovered that the confounder, severity of illness, introduces a spurious correlation between the question and answer, leading the model to rely on this spurious correlation and ignore the ECG input. Such models do not comprehend the ECG input and perform poorly in adversarial tests where different expressions of the same question are used in the training and testing sets. We designed a de-biased pre-training method to eliminate the confounder's effect according to the theory of backdoor adjustment. Our model performed well on the ECG-QA task under adversarial testing and demonstrated zero-shot capabilities. An interesting random ECG test further validated that our model effectively understands and utilizes the input ECG signal.
Authors: Arno Simons
Abstract: I present Astro-HEP-BERT, a transformer-based language model specifically designed for generating contextualized word embeddings (CWEs) to study the meanings of concepts in astrophysics and high-energy physics. Built on a general pretrained BERT model, Astro-HEP-BERT underwent further training over three epochs using the Astro-HEP Corpus, a dataset I curated from 21.84 million paragraphs extracted from more than 600,000 scholarly articles on arXiv, all belonging to at least one of these two scientific domains. The project demonstrates both the effectiveness and feasibility of adapting a bidirectional transformer for applications in the history, philosophy, and sociology of science (HPSS). The entire training process was conducted using freely available code, pretrained weights, and text inputs, completed on a single MacBook Pro Laptop (M2/96GB). Preliminary evaluations indicate that Astro-HEP-BERT's CWEs perform comparably to domain-adapted BERT models trained from scratch on larger datasets for domain-specific word sense disambiguation and induction and related semantic change analyses. This suggests that retraining general language models for specific scientific domains can be a cost-effective and efficient strategy for HPSS researchers, enabling high performance without the need for extensive training from scratch.
Authors: Wanqiu Long, Bonnie Webber
Abstract: Implicit discourse relation recognition involves determining relationships that hold between spans of text that are not linked by an explicit discourse connective. In recent years, the pre-train, prompt, and predict paradigm has emerged as a promising approach for tackling this task. However, previous work solely relied on manual verbalizers for implicit discourse relation recognition, which suffer from issues of ambiguity and even incorrectness. To overcome these limitations, we leverage the prototypes that capture certain class-level semantic features and the hierarchical label structure for different classes as the verbalizer. We show that our method improves on competitive baselines. Besides, our proposed approach can be extended to enable zero-shot cross-lingual learning, facilitating the recognition of discourse relations in languages with scarce resources. These advancement validate the practicality and versatility of our approach in addressing the issues of implicit discourse relation recognition across different languages.
Authors: Anna Glazkova, Olga Zakharova
Abstract: Large language models (LLMs) play a crucial role in natural language processing (NLP) tasks, improving the understanding, generation, and manipulation of human language across domains such as translating, summarizing, and classifying text. Previous studies have demonstrated that instruction-based LLMs can be effectively utilized for data augmentation to generate diverse and realistic text samples. This study applied prompt-based data augmentation to detect mentions of green practices in Russian social media. Detecting green practices in social media aids in understanding their prevalence and helps formulate recommendations for scaling eco-friendly actions to mitigate environmental issues. We evaluated several prompts for augmenting texts in a multi-label classification task, either by rewriting existing datasets using LLMs, generating new data, or combining both approaches. Our results revealed that all strategies improved classification performance compared to the models fine-tuned only on the original dataset, outperforming baselines in most cases. The best results were obtained with the prompt that paraphrased the original text while clearly indicating the relevant categories.
Authors: Aniket Bhattacharyya, Anurag Tripathi
Abstract: Invoices and receipts submitted by employees are visually rich documents (VRDs) with textual, visual and layout information. To protect against the risk of fraud and abuse, it is crucial for organizations to efficiently extract desired information from submitted receipts. This helps in the assessment of key factors such as appropriateness of the expense claim, adherence to spending and transaction policies, the validity of the receipt, as well as downstream anomaly detection at various levels. These documents are heterogenous, with multiple formats and languages, uploaded with different image qualities, and often do not contain ground truth labels for the efficient training of models. In this paper we propose Task Aware Instruction-based Labelling (TAIL), a method for synthetic label generation in VRD corpuses without labels, and fine-tune a multimodal Visually Rich Document Understanding Model (VRDU) on TAIL labels using response-based knowledge distillation without using the teacher model's weights or training dataset to conditionally generate annotations in the appropriate format. Using a benchmark external dataset where ground truth labels are available, we demonstrate conditions under which our approach performs at par with Claude 3 Sonnet through empirical studies. We then show that the resulting model performs at par or better on the internal expense documents of a large multinational organization than state-of-the-art LMM (large multimodal model) Claude 3 Sonnet while being 85% less costly and ~5X faster, and outperforms layout-aware baselines by more than 10% in Average Normalized Levenshtein Similarity (ANLS) scores due to its ability to reason and extract information from rare formats. Finally, we illustrate the usage of our approach in overpayment prevention.
Authors: Hitesh Laxmichand Patel, Amit Agarwal, Bhargava Kumar, Karan Gupta, Priyaranjan Pattnayak
Abstract: Accurate barcode detection and decoding in Identity documents is crucial for applications like security, healthcare, and education, where reliable data extraction and verification are essential. However, building robust detection models is challenging due to the lack of diverse, realistic datasets an issue often tied to privacy concerns and the wide variety of document formats. Traditional tools like Faker rely on predefined templates, making them less effective for capturing the complexity of real-world identity documents. In this paper, we introduce a new approach to synthetic data generation that uses LLMs to create contextually rich and realistic data without relying on predefined field. Using the vast knowledge LLMs have about different documents and content, our method creates data that reflects the variety found in real identity documents. This data is then encoded into barcode and overlayed on templates for documents such as Driver's licenses, Insurance cards, Student IDs. Our approach simplifies the process of dataset creation, eliminating the need for extensive domain knowledge or predefined fields. Compared to traditional methods like Faker, data generated by LLM demonstrates greater diversity and contextual relevance, leading to improved performance in barcode detection models. This scalable, privacy-first solution is a big step forward in advancing machine learning for automated document processing and identity verification.
Authors: Lukas Fischer, Yingqiang Gao, Alexa Lintner, Sarah Ebling
Abstract: Audio description (AD) is a crucial accessibility service provided to blind persons and persons with visual impairment, designed to convey visual information in acoustic form. Despite recent advancements in multilingual machine translation research, the lack of well-crafted and time-synchronized AD data impedes the development of audio description translation (ADT) systems that address the needs of multilingual countries such as Switzerland. Furthermore, since the majority of ADT systems rely solely on text, uncertainty exists as to whether incorporating visual information from the corresponding video clips can enhance the quality of ADT outputs. In this work, we present SwissADT, the first ADT system implemented for three main Swiss languages and English. By collecting well-crafted AD data augmented with video clips in German, French, Italian, and English, and leveraging the power of Large Language Models (LLMs), we aim to enhance information accessibility for diverse language populations in Switzerland by automatically translating AD scripts to the desired Swiss language. Our extensive experimental ADT results, composed of both automatic and human evaluations of ADT quality, demonstrate the promising capability of SwissADT for the ADT task. We believe that combining human expertise with the generation power of LLMs can further enhance the performance of ADT systems, ultimately benefiting a larger multilingual target population.
Authors: Junhong Shen, Atishay Jain, Zedian Xiao, Ishan Amlekar, Mouad Hadji, Aaron Podolny, Ameet Talwalkar
Abstract: Large Language Model (LLM) agents are rapidly improving to handle increasingly complex web-based tasks. Most of these agents rely on general-purpose, proprietary models like GPT-4 and focus on designing better prompts to improve their planning abilities. However, general-purpose LLMs are not specifically trained to understand specialized web contexts such as HTML, and they often struggle with long-horizon planning. We explore an alternative approach that fine-tunes open-source LLMs using production-scale workflow data collected from over 250 domains corresponding to 6 billion tokens. This simple yet effective approach shows substantial gains over prompting-based agents on existing benchmarks -- ScribeAgent achieves state-of-the-art direct generation performance on Mind2Web and improves the task success rate by 14.1% over the previous best text-only web agents on WebArena. We further perform detailed ablation studies on various fine-tuning design choices and provide insights into LLM selection, training recipes, context window optimization, and effect of dataset sizes.
Authors: Valentin Barriere
Abstract: Deep Learning models tend to learn correlations of patterns on huge datasets. The bigger these systems are, the more complex are the phenomena they can detect, and the more data they need for this. The use of Artificial Intelligence (AI) is becoming increasingly ubiquitous in our society, and its impact is growing everyday. The promises it holds strongly depend on their fair and universal use, such as access to information or education for all. In a world of inequalities, they can help to reach the most disadvantaged areas. However, such a universal systems must be able to represent society, without benefiting some at the expense of others. We must not reproduce the inequalities observed throughout the world, but educate these IAs to go beyond them. We have seen cases where these systems use gender, race, or even class information in ways that are not appropriate for resolving their tasks. Instead of real causal reasoning, they rely on spurious correlations, which is what we usually call a bias. In this paper, we first attempt to define what is a bias in general terms. It helps us to demystify the concept of bias, to understand why we can find them everywhere and why they are sometimes useful. Second, we focus over the notion of what is generally seen as negative bias, the one we want to avoid in machine learning, before presenting a general zoology containing the most common of these biases. We finally conclude by looking at classical methods to detect them, by means of specially crafted datasets of templates and specific algorithms, and also classical methods to mitigate them.
Authors: Sarah Griebel, Becca Cohen, Lucian Li, Jaihyun Park, Jiayu Liu, Jana Perkins, Ted Underwood
Abstract: Measures of textual similarity and divergence are increasingly used to study cultural change. But which measures align, in practice, with social evidence about change? We apply three different representations of text (topic models, document embeddings, and word-level perplexity) to three different corpora (literary studies, economics, and fiction). In every case, works by highly-cited authors and younger authors are textually ahead of the curve. We don't find clear evidence that one representation of text is to be preferred over the others. But alignment with social evidence is strongest when texts are represented through the top quartile of passages, suggesting that a text's impact may depend more on its most forward-looking moments than on sustaining a high level of innovation throughout.
Authors: Yixin Dong, Charlie F. Ruan, Yaxing Cai, Ruihang Lai, Ziyi Xu, Yilong Zhao, Tianqi Chen
Abstract: The applications of LLM Agents are becoming increasingly complex and diverse, leading to a high demand for structured outputs that can be parsed into code, structured function calls, and embodied agent commands. These developments bring significant demands for structured generation in LLM inference. Context-free grammar is a flexible approach to enable structured generation via constrained decoding. However, executing context-free grammar requires going through several stack states over all tokens in vocabulary during runtime, bringing non-negligible overhead for structured generation. In this paper, we propose XGrammar, a flexible and efficient structure generation engine for large language models. XGrammar accelerates context-free grammar execution by dividing the vocabulary into context-independent tokens that can be prechecked and context-dependent tokens that need to be interpreted during runtime. We further build transformations to expand the grammar context and reduce the number of context-independent tokens. Additionally, we build an efficient persistent stack to accelerate the context-dependent token checks. Finally, we co-design the grammar engine with LLM inference engine to overlap grammar computation with GPU executions. Evaluation results show that XGrammar can achieve up to 100x speedup over existing solutions. Combined with an LLM inference engine, it can generate near-zero overhead structure generation in end-to-end low-LLM serving.
Authors: Nathan Lambert, Jacob Morrison, Valentina Pyatkin, Shengyi Huang, Hamish Ivison, Faeze Brahman, Lester James V. Miranda, Alisa Liu, Nouha Dziri, Shane Lyu, Yuling Gu, Saumya Malik, Victoria Graf, Jena D. Hwang, Jiangjiang Yang, Ronan Le Bras, Oyvind Tafjord, Chris Wilhelm, Luca Soldaini, Noah A. Smith, Yizhong Wang, Pradeep Dasigi, Hannaneh Hajishirzi
Abstract: Language model post-training is applied to refine behaviors and unlock new skills across a wide range of recent language models, but open recipes for applying these techniques lag behind proprietary ones. The underlying training data and recipes for post-training are simultaneously the most important pieces of the puzzle and the portion with the least transparency. To bridge this gap, we introduce T\"ULU 3, a family of fully-open state-of-the-art post-trained models, alongside its data, code, and training recipes, serving as a comprehensive guide for modern post-training techniques. T\"ULU 3, which builds on Llama 3.1 base models, achieves results surpassing the instruct versions of Llama 3.1, Qwen 2.5, Mistral, and even closed models such as GPT-4o-mini and Claude 3.5-Haiku. The training algorithms for our models include supervised finetuning (SFT), Direct Preference Optimization (DPO), and a novel method we call Reinforcement Learning with Verifiable Rewards (RLVR). With T\"ULU 3, we introduce a multi-task evaluation scheme for post-training recipes with development and unseen evaluations, standard benchmark implementations, and substantial decontamination of existing open datasets on said benchmarks. We conclude with analysis and discussion of training methods that did not reliably improve performance. In addition to the T\"ULU 3 model weights and demo, we release the complete recipe -- including datasets for diverse core skills, a robust toolkit for data curation and evaluation, the training code and infrastructure, and, most importantly, a detailed report for reproducing and further adapting the T\"ULU 3 approach to more domains.
Authors: Alessandro Trevisan, Harry Giddens, Sarah Dillon, Alan F. Blackwell
Abstract: Generative large language models (LLMs), which create text without direct correspondence to truth value, are widely understood to resemble the uses of language described in Frankfurt's popular monograph On Bullshit. In this paper, we offer a rigorous investigation of this topic, identifying how the phenomenon has arisen, and how it might be analysed. In this paper, we elaborate on this argument to propose that LLM-based chatbots play the 'language game of bullshit'. We use statistical text analysis to investigate the features of this Wittgensteinian language game, based on a dataset constructed to contrast the language of 1,000 scientific publications with typical pseudo-scientific text generated by ChatGPT. We then explore whether the same language features can be detected in two well-known contexts of social dysfunction: George Orwell's critique of politics and language, and David Graeber's characterisation of bullshit jobs. Using simple hypothesis-testing methods, we demonstrate that a statistical model of the language of bullshit can reliably relate the Frankfurtian artificial bullshit of ChatGPT to the political and workplace functions of bullshit as observed in natural human language.
Authors: Kristina \v{S}ekrst, Jeremy McHugh, Jonathan Rodriguez Cefalu
Abstract: This paper explores the development of an ethical guardrail framework for AI systems, emphasizing the importance of customizable guardrails that align with diverse user values and underlying ethics. We address the challenges of AI ethics by proposing a structure that integrates rules, policies, and AI assistants to ensure responsible AI behavior, while comparing the proposed framework to the existing state-of-the-art guardrails. By focusing on practical mechanisms for implementing ethical standards, we aim to enhance transparency, user autonomy, and continuous improvement in AI systems. Our approach accommodates ethical pluralism, offering a flexible and adaptable solution for the evolving landscape of AI governance. The paper concludes with strategies for resolving conflicts between ethical directives, underscoring the present and future need for robust, nuanced and context-aware AI systems.
Authors: Zehua Pei, Hui-Ling Zhen, Xianzhi Yu, Sinno Jialin Pan, Mingxuan Yuan, Bei Yu
Abstract: Generative Pre-trained Transformers (GPTs) have demonstrated remarkable performance across diverse domains through the extensive scaling of model parameters. Recent works observe the redundancy across the transformer blocks and develop compression methods by structured pruning of the unimportant blocks. However, such straightforward elimination will always provide irreversible performance degradation. In this paper, we propose FuseGPT, a novel methodology to recycle the pruned transformer blocks to further recover the model performance. Firstly we introduce a new importance detection metric, Macro Influence (MI), to detect the long-term influence of each transformer block by calculating their loss of information after removal. Then we propose group-level layers fusion, which adopts the parameters in layers of the unimportant blocks and injects them into the corresponding layers inside the neighboring blocks. The fusion is not one-off but through iterative parameter updates by lightweight group-level fine-tuning. Specifically, these injected parameters are frozen but weighted with learnable rank decomposition matrices to reduce the overhead during fine-tuning. Our approach not only works well on large language models but also on large multimodal models. The experiments have shown that, by using modest amounts of data, FuseGPT can outperform previous works in both perplexity and zero-shot task performance.
Authors: Narcisa Guran, Florian Knauf, Man Ngo, Stefan Petrescu, Jan S. Rellermeyer
Abstract: Large language models have gained widespread popularity for their ability to process natural language inputs and generate insights derived from their training data, nearing the qualities of true artificial intelligence. This advancement has prompted enterprises worldwide to integrate LLMs into their services. So far, this effort is dominated by commercial cloud-based solutions like OpenAI's ChatGPT and Microsoft Azure. As the technology matures, however, there is a strong incentive for independence from major cloud providers through self-hosting "LLM as a Service", driven by privacy, cost, and customization needs. In practice, hosting LLMs independently presents significant challenges due to their complexity and integration issues with existing systems. In this paper, we discuss our vision for a forward-looking middleware system architecture that facilitates the deployment and adoption of LLMs in enterprises, even for advanced use cases in which we foresee LLMs to serve as gateways to a complete application ecosystem and, to some degree, absorb functionality traditionally attributed to the middleware.
Authors: Syed Mujtaba Hassan, Shahid Hussain, Abdul Samad
Abstract: Many NP-hard graph problems become easy for some classes of graphs, such as coloring is easy for bipartite graphs, but NP-hard in general. So we can ask question like when does a hard problem become easy? What is the minimum substructure for which the problem remains hard? We use the notion of boundary classes to study such questions. In this paper, we introduce a method for transforming the boundary class of one NP-hard graph problem into a boundary class for another problem. If $\Pi$ and $\Gamma$ are two NP-hard graph problems where $\Pi$ is reducible to $\Gamma$, we transform a boundary class of $\Pi$ into a boundary class of $\Gamma$. More formally if $\Pi$ is reducible to $\Gamma$, where the reduction is bijective and it maps hereditary classes of graphs to hereditary classes of graphs, then $X$ is a boundary class of $\Pi$ if and only if the image of $X$ under the reduction is a boundary class of $\Gamma$. This gives us a relationship between boundary classes and reducibility among several NP-hard problems. To show the strength of our main result, we apply our theorem to obtain some previously unknown boundary classes for a few graph problems namely; vertex-cover, clique, traveling-salesperson, bounded-degree-spanning-tree, subgraph-isomorphism and clique-cover.
Authors: Vijay Prakash, Kevin Lee, Arkaprabha Bhattacharya, Danny Yuxing Huang, Jessica Staddon
Abstract: Answering end user security questions is challenging. While large language models (LLMs) like GPT, LLAMA, and Gemini are far from error-free, they have shown promise in answering a variety of questions outside of security. We studied LLM performance in the area of end user security by qualitatively evaluating 3 popular LLMs on 900 systematically collected end user security questions. While LLMs demonstrate broad generalist ``knowledge'' of end user security information, there are patterns of errors and limitations across LLMs consisting of stale and inaccurate answers, and indirect or unresponsive communication styles, all of which impacts the quality of information received. Based on these patterns, we suggest directions for model improvement and recommend user strategies for interacting with LLMs when seeking assistance with security.
Authors: Shenglai Zeng, Jiankun Zhang, Bingheng Li, Yuping Lin, Tianqi Zheng, Dante Everaert, Hanqing Lu, Hui Liu, Hui Liu, Yue Xing, Monica Xiao Cheng, Jiliang Tang
Abstract: Retrieval-Augmented Generation (RAG) systems have shown promise in enhancing the performance of Large Language Models (LLMs). However, these systems face challenges in effectively integrating external knowledge with the LLM's internal knowledge, often leading to issues with misleading or unhelpful information. This work aims to provide a systematic study on knowledge checking in RAG systems. We conduct a comprehensive analysis of LLM representation behaviors and demonstrate the significance of using representations in knowledge checking. Motivated by the findings, we further develop representation-based classifiers for knowledge filtering. We show substantial improvements in RAG performance, even when dealing with noisy knowledge databases. Our study provides new insights into leveraging LLM representations for enhancing the reliability and effectiveness of RAG systems.
Authors: AJ Piergiovanni, Dahun Kim, Michael S. Ryoo, Isaac Noble, Anelia Angelova
Abstract: Generating automatic dense captions for videos that accurately describe their contents remains a challenging area of research. Most current models require processing the entire video at once. Instead, we propose an efficient, online approach which outputs frequent, detailed and temporally aligned captions, without access to future frames. Our model uses a novel autoregressive factorized decoding architecture, which models the sequence of visual features for each time segment, outputting localized descriptions and efficiently leverages the context from the previous video segments. This allows the model to output frequent, detailed captions to more comprehensively describe the video, according to its actual local content, rather than mimic the training data. Second, we propose an optimization for efficient training and inference, which enables scaling to longer videos. Our approach shows excellent performance compared to both offline and online methods, and uses 20\% less compute. The annotations produced are much more comprehensive and frequent, and can further be utilized in automatic video tagging and in large-scale video data harvesting.
Authors: Eric Tang, Bangding Yang, Xingyou Song
Abstract: With the rise of large language models (LLMs) for flexibly processing information as strings, a natural application is regression, specifically by preprocessing string representations into LLM embeddings as downstream features for metric prediction. In this paper, we provide one of the first comprehensive investigations into embedding-based regression and demonstrate that LLM embeddings as features can be better for high-dimensional regression tasks than using traditional feature engineering. This regression performance can be explained in part due to LLM embeddings over numeric data inherently preserving Lipschitz continuity over the feature space. Furthermore, we quantify the contribution of different model effects, most notably model size and language understanding, which we find surprisingly do not always improve regression performance.
Authors: Binqian Xu, Xiangbo Shu, Haiyang Mei, Guosen Xie, Basura Fernando, Mike Zheng Shou, Jinhui Tang
Abstract: Multimodal Large Language Models (MLLMs) have made significant advancements, demonstrating powerful capabilities in processing and understanding multimodal data. Fine-tuning MLLMs with Federated Learning (FL) allows for expanding the training data scope by including private data sources, thereby enhancing their practical applicability in privacy-sensitive domains. However, current research remains in the early stage, particularly in addressing the \textbf{multimodal heterogeneities} in real-world applications. In this paper, we introduce a benchmark for evaluating various downstream tasks in the federated fine-tuning of MLLMs within multimodal heterogeneous scenarios, laying the groundwork for the research in the field. Our benchmark encompasses two datasets, five comparison baselines, and four multimodal scenarios, incorporating over ten types of modal heterogeneities. To address the challenges posed by modal heterogeneity, we develop a general FedMLLM framework that integrates four representative FL methods alongside two modality-agnostic strategies. Extensive experimental results show that our proposed FL paradigm improves the performance of MLLMs by broadening the range of training data and mitigating multimodal heterogeneity. Code is available at https://github.com/1xbq1/FedMLLM
Authors: Feng Chen, Chenhui Gou, Jing Liu, Yang Yang, Zhaoyang Li, Jiyuan Zhang, Zhenbang Sun, Bohan Zhuang, Qi Wu
Abstract: As multimodal large language models (MLLMs) advance rapidly, rigorous evaluation has become essential, providing further guidance for their development. In this work, we focus on a unified and robust evaluation of \textbf{vision perception} abilities, the foundational skill of MLLMs. We find that existing perception benchmarks, each focusing on different question types, domains, and evaluation metrics, introduce significant evaluation variance, complicating comprehensive assessments of perception abilities when relying on any single benchmark. To address this, we introduce \textbf{AbilityLens}, a unified benchmark designed to evaluate MLLMs across six key perception abilities, focusing on both accuracy and stability, with each ability encompassing diverse question types, domains, and metrics. With the assistance of AbilityLens, we: (1) identify the strengths and weaknesses of current models, highlighting stability patterns and revealing a notable performance gap between open-source and closed-source models; (2) introduce an online evaluation mode, which uncovers interesting ability conflict and early convergence phenomena during MLLM training; and (3) design a simple ability-specific model merging method that combines the best ability checkpoint from early training stages, effectively mitigating performance decline due to ability conflict. The benchmark and online leaderboard will be released soon.
Authors: Simon Lupart, Zahra Abbasiantaeb, Mohammad Aliannejadi
Abstract: The Interactive Knowledge Assistant Track (iKAT) 2024 focuses on advancing conversational assistants, able to adapt their interaction and responses from personalized user knowledge. The track incorporates a Personal Textual Knowledge Base (PTKB) alongside Conversational AI tasks, such as passage ranking and response generation. Query Rewrite being an effective approach for resolving conversational context, we explore Large Language Models (LLMs), as query rewriters. Specifically, our submitted runs explore multi-aspect query generation using the MQ4CS framework, which we further enhance with Learned Sparse Retrieval via the SPLADE architecture, coupled with robust cross-encoder models. We also propose an alternative to the previous interleaving strategy, aggregating multiple aspects during the reranking phase. Our findings indicate that multi-aspect query generation is effective in enhancing performance when integrated with advanced retrieval and reranking models. Our results also lead the way for better personalization in Conversational Search, relying on LLMs to integrate personalization within query rewrite, and outperforming human rewrite performance.
Authors: Songhao Han, Wei Huang, Hairong Shi, Le Zhuo, Xiu Su, Shifeng Zhang, Xu Zhou, Xiaojuan Qi, Yue Liao, Si Liu
Abstract: The advancement of Large Vision Language Models (LVLMs) has significantly improved multimodal understanding, yet challenges remain in video reasoning tasks due to the scarcity of high-quality, large-scale datasets. Existing video question-answering (VideoQA) datasets often rely on costly manual annotations with insufficient granularity or automatic construction methods with redundant frame-by-frame analysis, limiting their scalability and effectiveness for complex reasoning. To address these challenges, we introduce VideoEspresso, a novel dataset that features VideoQA pairs preserving essential spatial details and temporal coherence, along with multimodal annotations of intermediate reasoning steps. Our construction pipeline employs a semantic-aware method to reduce redundancy, followed by generating QA pairs using GPT-4o. We further develop video Chain-of-Thought (CoT) annotations to enrich reasoning processes, guiding GPT-4o in extracting logical relationships from QA pairs and video content. To exploit the potential of high-quality VideoQA pairs, we propose a Hybrid LVLMs Collaboration framework, featuring a Frame Selector and a two-stage instruction fine-tuned reasoning LVLM. This framework adaptively selects core frames and performs CoT reasoning using multimodal evidence. Evaluated on our proposed benchmark with 14 tasks against 9 popular LVLMs, our method outperforms existing baselines on most tasks, demonstrating superior video reasoning capabilities. Our code and dataset will be released at: https://github.com/hshjerry/VideoEspresso
Authors: Ke Zhu, Yu Wang, Yanpeng Sun, Qiang Chen, Jiangjiang Liu, Gang Zhang, Jingdong Wang
Abstract: Multimodal RLHF usually happens after supervised finetuning (SFT) stage to continually improve vision-language models' (VLMs) comprehension. Conventional wisdom holds its superiority over continual SFT during this preference alignment stage. In this paper, we observe that the inherent value of multimodal RLHF lies in its negative supervision, the logit of the rejected responses. We thus propose a novel negative supervised finetuning (nSFT) approach that fully excavates these information resided. Our nSFT disentangles this negative supervision in RLHF paradigm, and continually aligns VLMs with a simple SFT loss. This is more memory efficient than multimodal RLHF where 2 (e.g., DPO) or 4 (e.g., PPO) large VLMs are strictly required. The effectiveness of nSFT is rigorously proved by comparing it with various multimodal RLHF approaches, across different dataset sources, base VLMs and evaluation metrics. Besides, fruitful of ablations are provided to support our hypothesis. We hope this paper will stimulate further research to properly align large vision language models.
Authors: Luca Parolari, Elena Izzo, Lamberto Ballan
Abstract: Referring Expression Comprehension (REC) aims to identify a particular object in a scene by a natural language expression, and is an important topic in visual language understanding. State-of-the-art methods for this task are based on deep learning, which generally requires expensive and manually labeled annotations. Some works tackle the problem with limited-supervision learning or relying on Large Vision and Language Models. However, the development of techniques to synthesize labeled data is overlooked. In this paper, we propose a novel framework that generates artificial data for the REC task, taking into account both textual and visual modalities. At first, our pipeline processes existing data to create variations in the annotations. Then, it generates an image using altered annotations as guidance. The result of this pipeline is a new dataset, called Harlequin, made by more than 1M queries. This approach eliminates manual data collection and annotation, enabling scalability and facilitating arbitrary complexity. We pre-train three REC models on Harlequin, then fine-tuned and evaluated on human-annotated datasets. Our experiments show that the pre-training on artificial data is beneficial for performance.
Authors: Yuhang Song, Mario Gianni, Chenguang Yang, Kunyang Lin, Te-Chuan Chiu, Anh Nguyen, Chun-Yi Lee
Abstract: This paper addresses the challenge of fine-grained alignment in Vision-and-Language Navigation (VLN) tasks, where robots navigate realistic 3D environments based on natural language instructions. Current approaches use contrastive learning to align language with visual trajectory sequences. Nevertheless, they encounter difficulties with fine-grained vision negatives. To enhance cross-modal embeddings, we introduce a novel Bayesian Optimization-based adversarial optimization framework for creating fine-grained contrastive vision samples. To validate the proposed methodology, we conduct a series of experiments to assess the effectiveness of the enriched embeddings on fine-grained vision negatives. We conduct experiments on two common VLN benchmarks R2R and REVERIE, experiments on the them demonstrate that these embeddings benefit navigation, and can lead to a promising performance enhancement. Our source code and trained models are available at: https://anonymous.4open.science/r/FGVLN.
Authors: Camilo Chac\'on Sartori, Christian Blum, Filippo Bistaffa
Abstract: The fast advancement of Large Vision-Language Models (LVLMs) has shown immense potential. These models are increasingly capable of tackling abstract visual tasks. Geometric structures, particularly graphs with their inherent flexibility and complexity, serve as an excellent benchmark for evaluating these models' predictive capabilities. While human observers can readily identify subtle visual details and perform accurate analyses, our investigation reveals that state-of-the-art LVLMs exhibit consistent limitations in specific visual graph scenarios, especially when confronted with stylistic variations. In response to these challenges, we introduce VisGraphVar (Visual Graph Variability), a customizable benchmark generator able to produce graph images for seven distinct task categories (detection, classification, segmentation, pattern recognition, link prediction, reasoning, matching), designed to systematically evaluate the strengths and limitations of individual LVLMs. We use VisGraphVar to produce 990 graph images and evaluate six LVLMs, employing two distinct prompting strategies, namely zero-shot and chain-of-thought. The findings demonstrate that variations in visual attributes of images (e.g., node labeling and layout) and the deliberate inclusion of visual imperfections, such as overlapping nodes, significantly affect model performance. This research emphasizes the importance of a comprehensive evaluation across graph-related tasks, extending beyond reasoning alone. VisGraphVar offers valuable insights to guide the development of more reliable and robust systems capable of performing advanced visual graph analysis.
Authors: Dingyuan Shi, Yong Wang, Hangyu Li, Xiangxiang Chu
Abstract: Diffusion models have shown remarkable success in text-to-image generation, making alignment methods for these models increasingly important. A key challenge is the sparsity of preference labels, which are typically available only at the terminal of denoising trajectories. This raises the issue of how to assign credit across denoising steps based on these sparse labels. In this paper, we propose Denoised Distribution Estimation (DDE), a novel method for credit assignment. Unlike previous approaches that rely on auxiliary models or hand-crafted schemes, DDE derives its strategy more explicitly. The proposed DDE directly estimates the terminal denoised distribution from the perspective of each step. It is equipped with two estimation strategies and capable of representing the entire denoising trajectory with a single model inference. Theoretically and empirically, we show that DDE prioritizes optimizing the middle part of the denoising trajectory, resulting in a novel and effective credit assignment scheme. Extensive experiments demonstrate that our approach achieves superior performance, both quantitatively and qualitatively.
Authors: Tanveer Hannan, Md Mohaiminul Islam, Jindong Gu, Thomas Seidl, Gedas Bertasius
Abstract: Large language models (LLMs) excel at retrieving information from lengthy text, but their vision-language counterparts (VLMs) face difficulties with hour-long videos, especially for temporal grounding. Specifically, these VLMs are constrained by frame limitations, often losing essential temporal details needed for accurate event localization in extended video content. We propose ReVisionLLM, a recursive vision-language model designed to locate events in hour-long videos. Inspired by human search strategies, our model initially targets broad segments of interest, progressively revising its focus to pinpoint exact temporal boundaries. Our model can seamlessly handle videos of vastly different lengths, from minutes to hours. We also introduce a hierarchical training strategy that starts with short clips to capture distinct events and progressively extends to longer videos. To our knowledge, ReVisionLLM is the first VLM capable of temporal grounding in hour-long videos, outperforming previous state-of-the-art methods across multiple datasets by a significant margin (+2.6% R1@0.1 on MAD). The code is available at https://github.com/Tanveer81/ReVisionLLM.
Authors: Kaichen Zhang, Yifei Shen, Bo Li, Ziwei Liu
Abstract: Recent advances in Large Multimodal Models (LMMs) lead to significant breakthroughs in both academia and industry. One question that arises is how we, as humans, can understand their internal neural representations. This paper takes an initial step towards addressing this question by presenting a versatile framework to identify and interpret the semantics within LMMs. Specifically, 1) we first apply a Sparse Autoencoder(SAE) to disentangle the representations into human understandable features. 2) We then present an automatic interpretation framework to interpreted the open-semantic features learned in SAE by the LMMs themselves. We employ this framework to analyze the LLaVA-NeXT-8B model using the LLaVA-OV-72B model, demonstrating that these features can effectively steer the model's behavior. Our results contribute to a deeper understanding of why LMMs excel in specific tasks, including EQ tests, and illuminate the nature of their mistakes along with potential strategies for their rectification. These findings offer new insights into the internal mechanisms of LMMs and suggest parallels with the cognitive processes of the human brain.
Authors: Eugene Eberbach
Abstract: Evolution by natural selection, which is one of the most compelling themes of modern science, brought forth evolutionary algorithms and evolutionary computation, applying mechanisms of evolution in nature to various problems solved by computers. In this paper we concentrate on evolutionary automata that constitute an analogous model of evolutionary computation compared to well-known evolutionary algorithms. Evolutionary automata provide a more complete dual model of evolutionary computation, similar like abstract automata (e.g., Turing machines) form a more formal and precise model compared to recursive algorithms and their subset - evolutionary algorithms. An evolutionary automaton is an automaton that evolves performing evolutionary computation perhaps using an infinite number of generations. This model allows for a direct modeling evolution of evolution, and leads to tremendous expressiveness of evolutionary automata and evolutionary computation. This also gives the hint to the power of natural evolution that is self-evolving by interactive feedback with the environment.
Authors: Tao Zhang, Ziqi Zhang, Zongyang Ma, Yuxin Chen, Zhongang Qi, Chunfeng Yuan, Bing Li, Junfu Pu, Yuxuan Zhao, Zehua Xie, Jin Ma, Ying Shan, Weiming Hu
Abstract: Advanced Multimodal Large Language Models (MLLMs) struggle with recent Knowledge-based VQA tasks, such as INFOSEEK and Encyclopedic-VQA, due to their limited and frozen knowledge scope, often leading to ambiguous and inaccurate responses. Thus, multimodal Retrieval-Augmented Generation (mRAG) is naturally introduced to provide MLLMs with comprehensive and up-to-date knowledge, effectively expanding the knowledge scope. However, current mRAG methods have inherent drawbacks, including: 1) Performing retrieval even when external knowledge is not needed. 2) Lacking of identification of evidence that supports the query. 3) Increasing model complexity due to additional information filtering modules or rules. To address these shortcomings, we propose a novel generalized framework called \textbf{m}ultimodal \textbf{R}etrieval-\textbf{R}eflection-\textbf{A}ugmented \textbf{G}eneration (mR$^2$AG), which achieves adaptive retrieval and useful information localization to enable answers through two easy-to-implement reflection operations, preventing high model complexity. In mR$^2$AG, Retrieval-Reflection is designed to distinguish different user queries and avoids redundant retrieval calls, and Relevance-Reflection is introduced to guide the MLLM in locating beneficial evidence of the retrieved content and generating answers accordingly. In addition, mR$^2$AG can be integrated into any well-trained MLLM with efficient fine-tuning on the proposed mR$^2$AG Instruction-Tuning dataset (mR$^2$AG-IT). mR$^2$AG significantly outperforms state-of-the-art MLLMs (e.g., GPT-4v/o) and RAG-based MLLMs on INFOSEEK and Encyclopedic-VQA, while maintaining the exceptional capabilities of base MLLMs across a wide range of Visual-dependent tasks.
Authors: E-Ro Nguyen, Hieu Le, Dimitris Samaras, Michael Ryoo
Abstract: Recent works on Generalized Referring Expression Segmentation (GRES) struggle with handling complex expressions referring to multiple distinct objects. This is because these methods typically employ an end-to-end foreground-background segmentation and lack a mechanism to explicitly differentiate and associate different object instances to the text query. To this end, we propose InstAlign, a method that incorporates object-level reasoning into the segmentation process. Our model leverages both text and image inputs to extract a set of object-level tokens that capture both the semantic information in the input prompt and the objects within the image. By modeling the text-object alignment via instance-level supervision, each token uniquely represents an object segment in the image, while also aligning with relevant semantic information from the text. Extensive experiments on the gRefCOCO and Ref-ZOM benchmarks demonstrate that our method significantly advances state-of-the-art performance, setting a new standard for precise and flexible GRES.
Authors: Karsten Roth, Zeynep Akata, Dima Damen, Ivana Bala\v{z}evi\'c, Olivier J. H\'enaff
Abstract: Large-scale multimodal representation learning successfully optimizes for zero-shot transfer at test time. Yet the standard pretraining paradigm (contrastive learning on large amounts of image-text data) does not explicitly encourage representations to support few-shot adaptation. In this work, we propose a simple, but carefully designed extension to multimodal pretraining which enables representations to accommodate additional context. Using this objective, we show that vision-language models can be trained to exhibit significantly increased few-shot adaptation: across 21 downstream tasks, we find up to four-fold improvements in test-time sample efficiency, and average few-shot adaptation gains of over 5%, while retaining zero-shot generalization performance across model scales and training durations. In particular, equipped with simple, training-free, metric-based adaptation mechanisms, our representations easily surpass more complex and expensive optimization-based schemes, vastly simplifying generalization to new domains.
Authors: Samarth N Ramesh, Zhixue Zhao
Abstract: As text-to-image models grow increasingly powerful and complex, their burgeoning size presents a significant obstacle to widespread adoption, especially on resource-constrained devices. This paper presents a pioneering study on post-training pruning of Stable Diffusion 2, addressing the critical need for model compression in text-to-image domain. Our study tackles the pruning techniques for the previously unexplored multi-modal generation models, and particularly examines the pruning impact on the textual component and the image generation component separately. We conduct a comprehensive comparison on pruning the model or the single component of the model in various sparsities. Our results yield previously undocumented findings. For example, contrary to established trends in language model pruning, we discover that simple magnitude pruning outperforms more advanced techniques in text-to-image context. Furthermore, our results show that Stable Diffusion 2 can be pruned to 38.5% sparsity with minimal quality loss, achieving a significant reduction in model size. We propose an optimal pruning configuration that prunes the text encoder to 47.5% and the diffusion generator to 35%. This configuration maintains image generation quality while substantially reducing computational requirements. In addition, our work uncovers intriguing questions about information encoding in text-to-image models: we observe that pruning beyond certain thresholds leads to sudden performance drops (unreadable images), suggesting that specific weights encode critical semantics information. This finding opens new avenues for future research in model compression, interoperability, and bias identification in text-to-image models. By providing crucial insights into the pruning behavior of text-to-image models, our study lays the groundwork for developing more efficient and accessible AI-driven image generation systems
Authors: Daeun Lee, Jaehong Yoon, Jaemin Cho, Mohit Bansal
Abstract: Recent text-to-video (T2V) diffusion models have demonstrated impressive generation capabilities across various domains. However, these models often generate videos that have misalignments with text prompts, especially when the prompts describe complex scenes with multiple objects and attributes. To address this, we introduce VideoRepair, a novel model-agnostic, training-free video refinement framework that automatically identifies fine-grained text-video misalignments and generates explicit spatial and textual feedback, enabling a T2V diffusion model to perform targeted, localized refinements. VideoRepair consists of four stages: In (1) video evaluation, we detect misalignments by generating fine-grained evaluation questions and answering those questions with MLLM. In (2) refinement planning, we identify accurately generated objects and then create localized prompts to refine other areas in the video. Next, in (3) region decomposition, we segment the correctly generated area using a combined grounding module. We regenerate the video by adjusting the misaligned regions while preserving the correct regions in (4) localized refinement. On two popular video generation benchmarks (EvalCrafter and T2V-CompBench), VideoRepair substantially outperforms recent baselines across various text-video alignment metrics. We provide a comprehensive analysis of VideoRepair components and qualitative examples.
Authors: Xiaoman Zhang, Hong-Yu Zhou, Xiaoli Yang, Oishi Banerjee, Juli\'an N. Acosta, Josh Miller, Ouwen Huang, Pranav Rajpurkar
Abstract: AI-driven models have demonstrated significant potential in automating radiology report generation for chest X-rays. However, there is no standardized benchmark for objectively evaluating their performance. To address this, we present ReXrank, https://rexrank.ai, a public leaderboard and challenge for assessing AI-powered radiology report generation. Our framework incorporates ReXGradient, the largest test dataset consisting of 10,000 studies, and three public datasets (MIMIC-CXR, IU-Xray, CheXpert Plus) for report generation assessment. ReXrank employs 8 evaluation metrics and separately assesses models capable of generating only findings sections and those providing both findings and impressions sections. By providing this standardized evaluation framework, ReXrank enables meaningful comparisons of model performance and offers crucial insights into their robustness across diverse clinical settings. Beyond its current focus on chest X-rays, ReXrank's framework sets the stage for comprehensive evaluation of automated reporting across the full spectrum of medical imaging.
URLs: https://rexrank.ai,
Authors: Vladislav Lialin, Vijeta Deshpande, Xiaowei Yao, Anna Rumshisky
Abstract: This paper presents a systematic overview of parameter-efficient fine-tuning methods, covering over 50 papers published between early 2019 and mid-2024. These methods aim to address the challenges of fine-tuning large language models by training only a small subset of parameters. We provide a taxonomy that covers a broad range of methods and present a detailed method comparison with a specific focus on real-life efficiency in fine-tuning multibillion-scale language models. We also conduct an extensive head-to-head experimental comparison of 15 diverse PEFT methods, evaluating their performance and efficiency on models up to 11B parameters. Our findings reveal that methods previously shown to surpass a strong LoRA baseline face difficulties in resource-constrained settings, where hyperparameter optimization is limited and the network is fine-tuned only for a few epochs. Finally, we provide a set of practical recommendations for using PEFT methods and outline potential future research directions.
Authors: Johannes Schneider, Steffi Haag, Leona Chandra Kruse
Abstract: Large language models LLMs like ChatGPT have reached the 100 Mio user barrier in record time and might increasingly enter all areas of our life leading to a diverse set of interactions between those Artificial Intelligence models and humans. While many studies have discussed governance and regulations deductively from first-order principles, few studies provide an inductive, data-driven lens based on observing dialogues between humans and LLMs especially when it comes to non-collaborative, competitive situations that have the potential to pose a serious threat to people. In this work, we conduct a user study engaging over 40 individuals across all age groups in price negotiations with an LLM. We explore how people interact with an LLM, investigating differences in negotiation outcomes and strategies. Furthermore, we highlight shortcomings of LLMs with respect to their reasoning capabilities and, in turn, susceptiveness to prompt hacking, which intends to manipulate the LLM to make agreements that are against its instructions or beyond any rationality. We also show that the negotiated prices humans manage to achieve span a broad range, which points to a literacy gap in effectively interacting with LLMs.
Authors: Zhiyu Zoey Chen, Jing Ma, Xinlu Zhang, Nan Hao, An Yan, Armineh Nourbakhsh, Xianjun Yang, Julian McAuley, Linda Petzold, William Yang Wang
Abstract: In the fast-evolving domain of artificial intelligence, large language models (LLMs) such as GPT-3 and GPT-4 are revolutionizing the landscapes of finance, healthcare, and law: domains characterized by their reliance on professional expertise, challenging data acquisition, high-stakes, and stringent regulatory compliance. This survey offers a detailed exploration of the methodologies, applications, challenges, and forward-looking opportunities of LLMs within these high-stakes sectors. We highlight the instrumental role of LLMs in enhancing diagnostic and treatment methodologies in healthcare, innovating financial analytics, and refining legal interpretation and compliance strategies. Moreover, we critically examine the ethics for LLM applications in these fields, pointing out the existing ethical concerns and the need for transparent, fair, and robust AI systems that respect regulatory norms. By presenting a thorough review of current literature and practical applications, we showcase the transformative impact of LLMs, and outline the imperative for interdisciplinary cooperation, methodological advancements, and ethical vigilance. Through this lens, we aim to spark dialogue and inspire future research dedicated to maximizing the benefits of LLMs while mitigating their risks in these precision-dependent sectors. To facilitate future research on LLMs in these critical societal domains, we also initiate a reading list that tracks the latest advancements under this topic, which will be continually updated: \url{https://github.com/czyssrs/LLM_X_papers}.
Authors: Juri Opitz, Shira Wein, Nathan Schneider
Abstract: Large Language Models (LLMs) have become capable of generating highly fluent text in certain languages, without modules specially designed to capture grammar or semantic coherence. What does this mean for the future of linguistic expertise in NLP? We highlight several aspects in which NLP (still) relies on linguistics, or where linguistic thinking can illuminate new directions. We argue our case around the acronym RELIES that encapsulates six major facets where linguistics contributes to NLP: Resources, Evaluation, Low-resource settings, Interpretability, Explanation, and the Study of language. This list is not exhaustive, nor is linguistics the main point of reference for every effort under these themes; but at a macro level, these facets highlight the enduring importance of studying machine systems vis-\`a-vis systems of human language.
Authors: Xun Liu, Zhengwei Ni
Abstract: Large language models (LLMs) exhibit impressive proficiency in natural language generation, understanding user instructions, and emulating human-like language use, which has led to significant interest in their application to role-playing scenarios. However, the manual collection of role-specific script data and the evaluation of model performance are resource-intensive processes. This project introduces a prompt-based framework designed to leverage GPT's capabilities for the generation of role-playing dialogue datasets and the evaluation of role-playing performance. To validate the effectiveness of the GPT-based generation and evaluation, we further incorporate the recall-oriented Rouge-L metric, providing an additional quantitative measure of performance.
Authors: Erik Derner, Sara Sansalvador de la Fuente, Yoan Guti\'errez, Paloma Moreda, Nuria Oliver
Abstract: Gender bias in text corpora that are used for a variety of natural language processing (NLP) tasks, such as for training large language models (LLMs), can lead to the perpetuation and amplification of societal inequalities. This phenomenon is particularly pronounced in gendered languages like Spanish or French, where grammatical structures inherently encode gender, making the bias analysis more challenging. A first step in quantifying gender bias in text entails computing biases in gender representation, i.e., differences in the prevalence of words referring to males vs. females. Existing methods to measure gender representation bias in text corpora have mainly been proposed for English and do not generalize to gendered languages due to the intrinsic linguistic differences between English and gendered languages. This paper introduces a novel methodology that leverages the contextual understanding capabilities of LLMs to quantitatively measure gender representation bias in Spanish corpora. By utilizing LLMs to identify and classify gendered nouns and pronouns in relation to their reference to human entities, our approach provides a robust analysis of gender representation bias in gendered languages. We empirically validate our method on four widely-used benchmark datasets, uncovering significant gender prevalence disparities with a male-to-female ratio ranging from 4:1 to 6:1. These findings demonstrate the value of our methodology for bias quantification in gendered language corpora and suggest its application in NLP, contributing to the development of more equitable language technologies.
Authors: Yifan Yang, Kai Zhen, Ershad Banijamal, Athanasios Mouchtaris, Zheng Zhang
Abstract: Fine-tuning large language models (LLMs) has achieved remarkable performance across various natural language processing tasks, yet it demands more and more memory as model sizes keep growing. To address this issue, the recently proposed Memory-efficient Zeroth-order (MeZO) methods attempt to fine-tune LLMs using only forward passes, thereby avoiding the need for a backpropagation graph. However, significant performance drops and a high risk of divergence have limited their widespread adoption. In this paper, we propose the Adaptive Zeroth-order Tensor-Train Adaption (AdaZeta) framework, specifically designed to improve the performance and convergence of the ZO methods. To enhance dimension-dependent ZO estimation accuracy, we introduce a fast-forward, low-parameter tensorized adapter. To tackle the frequently observed divergence issue in large-scale ZO fine-tuning tasks, we propose an adaptive query number schedule that guarantees convergence. Detailed theoretical analysis and extensive experimental results on Roberta-Large and Llama-2-7B models substantiate the efficacy of our AdaZeta framework in terms of accuracy, memory efficiency, and convergence speed.
Authors: Jize Wang, Zerun Ma, Yining Li, Songyang Zhang, Cailian Chen, Kai Chen, Xinyi Le
Abstract: Significant focus has been placed on integrating large language models (LLMs) with various tools in developing general-purpose agents. This poses a challenge to LLMs' tool-use capabilities. However, there are evident gaps between existing tool-use evaluations and real-world scenarios. Current evaluations often use AI-generated queries, single-step tasks, dummy tools, and text-only interactions, failing to reveal the agents' real-world problem-solving abilities effectively. To address this, we propose GTA, a benchmark for General Tool Agents, featuring three main aspects: (i) Real user queries: human-written queries with simple real-world objectives but implicit tool-use, requiring the LLM to reason the suitable tools and plan the solution steps. (ii) Real deployed tools: an evaluation platform equipped with tools across perception, operation, logic, and creativity categories to evaluate the agents' actual task execution performance. (iii) Real multimodal inputs: authentic image files, such as spatial scenes, web page screenshots, tables, code snippets, and printed/handwritten materials, used as the query contexts to align with real-world scenarios closely. We design 229 real-world tasks and executable tool chains to evaluate mainstream LLMs. Our findings show that real-world user queries are challenging for existing LLMs, with GPT-4 completing less than 50% of the tasks and most LLMs achieving below 25%. This evaluation reveals the bottlenecks in the tool-use capabilities of current LLMs in real-world scenarios, which provides future direction for advancing general-purpose tool agents. The code and dataset are available at https://github.com/open-compass/GTA.
Authors: Faeze Brahman, Sachin Kumar, Vidhisha Balachandran, Pradeep Dasigi, Valentina Pyatkin, Abhilasha Ravichander, Sarah Wiegreffe, Nouha Dziri, Khyathi Chandu, Jack Hessel, Yulia Tsvetkov, Noah A. Smith, Yejin Choi, Hannaneh Hajishirzi
Abstract: Chat-based language models are designed to be helpful, yet they should not comply with every user request. While most existing work primarily focuses on refusal of "unsafe" queries, we posit that the scope of noncompliance should be broadened. We introduce a comprehensive taxonomy of contextual noncompliance describing when and how models should not comply with user requests. Our taxonomy spans a wide range of categories including incomplete, unsupported, indeterminate, and humanizing requests (in addition to unsafe requests). To test noncompliance capabilities of language models, we use this taxonomy to develop a new evaluation suite of 1000 noncompliance prompts. We find that most existing models show significantly high compliance rates in certain previously understudied categories with models like GPT-4 incorrectly complying with as many as 30% of requests. To address these gaps, we explore different training strategies using a synthetically-generated training set of requests and expected noncompliant responses. Our experiments demonstrate that while direct finetuning of instruction-tuned models can lead to both over-refusal and a decline in general capabilities, using parameter efficient methods like low rank adapters helps to strike a good balance between appropriate noncompliance and other capabilities.
Authors: Jiale Kang
Abstract: Low-Rank Adaptation (LoRA) has achieved remarkable training results by freezing the original weights and training only low-rank matrices, establishing itself as the predominant fine-tuning method for LLMs. In pursuit of performance closer to full-parameter training, a series of LoRA variants have emerged, such as LoRA+, PISSA, Olora, and LoRA-GA. This paper introduces a novel PEFT technique distinct from LoRA, called Block-Affine Adaptation (Bone). By dividing the original weights into multiple subspaces that share a single matrix for weight updates, Bone simplifies the process by requiring the trainable matrix to be initialized to zero, eliminating the need for complex initialization as in some LoRA variants. Compared to LoRA, Bone significantly reduces memory usage and achieves faster computation. Evaluation of both NLU and NLG tasks demonstrates that Bone substantially outperforms LoRA and its variants. Inspired by Pissa, we further proposed the ``Weight Guide'' theory to better utilize the information from the original weights. By integrating ``Weight Guide'' with Bone, we developed a new structure called Block-Affine Transformation (Bat), and ablation experiments confirmed the effectiveness of ``Weight Guide''.
Authors: Zhengkai Lin, Zhihang Fu, Kai Liu, Liang Xie, Binbin Lin, Wenxiao Wang, Deng Cai, Yue Wu, Jieping Ye
Abstract: While large language models (LLMs) showcase unprecedented capabilities, they also exhibit certain inherent limitations when facing seemingly trivial tasks. A prime example is the recently debated "reversal curse", which surfaces when models, having been trained on the fact "A is B", struggle to generalize this knowledge to infer that "B is A". In this paper, we examine the manifestation of the reversal curse across various tasks and delve into both the generalization abilities and the problem-solving mechanisms of LLMs. This investigation leads to a series of significant insights: (1) LLMs are able to generalize to "B is A" when both A and B are presented in the context as in the case of a multiple-choice question. (2) This generalization ability is highly correlated to the structure of the fact "A is B" in the training documents. For example, this generalization only applies to biographies structured in "[Name] is [Description]" but not to "[Description] is [Name]". (3) We propose and verify the hypothesis that LLMs possess an inherent bias in fact recalling during knowledge application, which explains and underscores the importance of the document structure to successful learning. (4) The negative impact of this bias on the downstream performance of LLMs can hardly be mitigated through training alone. These findings offer a novel perspective on interpreting LLMs' generalization through their intrinsic mechanisms and provide insights for developing more effective learning methods. Our code and data are available at https://github.com/alibaba/thinking_bias.git.
Authors: Rohan Sanghera, Arun James Thirunavukarasu, Marc El Khoury, Jessica O'Logbon, Yuqing Chen, Archie Watt, Mustafa Mahmood, Hamid Butt, George Nishimura, Andrew Soltan
Abstract: Large language models (LLMs) excel in tasks requiring processing and interpretation of input text. Abstract screening is a labour-intensive component of systematic review involving repetitive application of inclusion and exclusion criteria on a large volume of studies identified by a literature search. Here, LLMs (GPT-3.5 Turbo, GPT-4 Turbo, GPT-4o, Llama 3 70B, Gemini 1.5 Pro, and Claude Sonnet 3.5) were trialled on systematic reviews in a full issue of the Cochrane Library to evaluate their accuracy in zero-shot binary classification for abstract screening. Trials over a subset of 800 records identified optimal prompting strategies and demonstrated superior performance of LLMs to human researchers in terms of sensitivity (LLM-max = 1.000, human-max = 0.775), precision (LLM-max = 0.927, human-max = 0.911), and balanced accuracy (LLM-max = 0.904, human-max = 0.865). The best performing LLM-prompt combinations were trialled across every replicated search result (n = 119,691), and exhibited consistent sensitivity (range 0.756-1.000) but diminished precision (range 0.004-0.096). 66 LLM-human and LLM-LLM ensembles exhibited perfect sensitivity with a maximal precision of 0.458, with less observed performance drop in larger trials. Significant variation in performance was observed between reviews, highlighting the importance of domain-specific validation before deployment. LLMs may reduce the human labour cost of systematic review with maintained or improved accuracy and sensitivity. Systematic review is the foundation of evidence synthesis across academic disciplines, including evidence-based medicine, and LLMs may increase the efficiency and quality of this mode of research.
Authors: Thang M. Pham, Phat T. Nguyen, Seunghyun Yoon, Viet Dac Lai, Franck Dernoncourt, Trung Bui
Abstract: While small language models (SLMs) show promises for mobile deployment, their real-world performance and applications on smartphones remains underexplored. We present SlimLM, a series of SLMs optimized for document assistance tasks on mobile devices. Through extensive experiments on a Samsung Galaxy S24, we identify the optimal trade-offs between model size (ranging from 125M to 7B parameters), context length, and inference time for efficient on-device processing. SlimLM is pre-trained on SlimPajama-627B and fine-tuned on DocAssist, our constructed dataset for summarization, question answering and suggestion tasks. Our smallest model demonstrates efficient performance on S24, while larger variants offer enhanced capabilities within mobile constraints. We evaluate SlimLM against existing SLMs, showing comparable or superior performance and offering a benchmark for future research in on-device language models. We also provide an Android application, offering practical insights into SLM deployment. Our findings provide valuable insights and illuminate the capabilities of running advanced language models on high-end smartphones, potentially reducing server costs and enhancing privacy through on-device processing.
Authors: Wang Qun, Liu Yang, Lin Qingquan, Jiang Ling
Abstract: We introduce Xmodel-1.5, a 1-billion-parameter multilingual large language model pretrained on 2 trillion tokens, designed for balanced performance and scalability. Unlike most large models that use the BPE tokenizer, Xmodel-1.5 employs a custom unigram tokenizer with 65,280 tokens, optimizing both efficiency and accuracy. The model delivers competitive results across multiple languages, including Thai, Arabic, French, Chinese, and English, outperforming Alibaba's PolyLM-1.7B on respective evaluation datasets. Xmodel-1.5 excels in benchmarks like mMMLU and PIQA, and achieves state-of-the-art results in Thai. To support low-resource language research, we release Xdata_Thai, a Thai-specific evaluation dataset featuring unique linguistic challenges such as gendered particles and idioms. While the model demonstrates strong performance, there is still room for improvement in handling culturally specific nuances. We hope this work contributes to advancements in multilingual AI research. Models and code are publicly available on GitHub at https://github.com/XiaoduoAILab/XmodelLM-1.5
Authors: John D. Hastings, Sherri Weitl-Harms, Joseph Doty, Zachary J. Myers, Warren Thompson
Abstract: This research explores the application of large language models (LLMs) to generate synthetic datasets for Product Desirability Toolkit (PDT) testing, a key component in evaluating user sentiment and product experience. Utilizing gpt-4o-mini, a cost-effective alternative to larger commercial LLMs, three methods, Word+Review, Review+Word, and Supply-Word, were each used to synthesize 1000 product reviews. The generated datasets were assessed for sentiment alignment, textual diversity, and data generation cost. Results demonstrated high sentiment alignment across all methods, with Pearson correlations ranging from 0.93 to 0.97. Supply-Word exhibited the highest diversity and coverage of PDT terms, although with increased generation costs. Despite minor biases toward positive sentiments, in situations with limited test data, LLM-generated synthetic data offers significant advantages, including scalability, cost savings, and flexibility in dataset production.
Authors: Christopher Nguyen, William Nguyen, Atsushi Suzuki, Daisuke Oku, Hong An Phan, Sang Dinh, Zooey Nguyen, Anh Ha, Shruti Raghavan, Huy Vo, Thang Nguyen, Lan Nguyen, Yoshikuni Hirayama
Abstract: Large Language Models (LLMs) have demonstrated the potential to address some issues within the semiconductor industry. However, they are often general-purpose models that lack the specialized knowledge needed to tackle the unique challenges of this sector, such as the intricate physics and chemistry of semiconductor devices and processes. SemiKong, the first industry-specific LLM for the semiconductor domain, provides a foundation that can be used to develop tailored proprietary models. With SemiKong 1.0, we aim to develop a foundational model capable of understanding etching problems at an expert level. Our key contributions include (a) curating a comprehensive corpus of semiconductor-related texts, (b) creating a foundational model with in-depth semiconductor knowledge, and (c) introducing a framework for integrating expert knowledge, thereby advancing the evaluation process of domain-specific AI models. Through fine-tuning a pre-trained LLM using our curated dataset, we have shown that SemiKong outperforms larger, general-purpose LLMs in various semiconductor manufacturing and design tasks. Our extensive experiments underscore the importance of developing domain-specific LLMs as a foundation for company- or tool-specific proprietary models, paving the way for further research and applications in the semiconductor domain. Code and dataset will be available at https://github.com/aitomatic/semikong
Authors: Chenggang Zhao, Genghan Zhang, Ao Shen, Mingyu Gao
Abstract: The demands for higher performance and accuracy in neural networks (NNs) never end. Existing tensor compilation and Neural Architecture Search (NAS) techniques orthogonally optimize the two goals but actually share many similarities in their concrete strategies. We exploit such opportunities by combining the two into one and make a case for Kernel Architecture Search (KAS). KAS reviews NAS from a system perspective and zooms into a more fine-grained level to generate neural kernels with both high performance and good accuracy. To demonstrate the potential of KAS, we build an end-to-end framework, Canvas, to find high-quality kernels as convolution replacements. Canvas samples from a rich set of fine-grained primitives to stochastically and iteratively construct new kernels and evaluate them according to user-specified constraints. Canvas supports freely adjustable tensor dimension sizes inside the kernel and uses two levels of solvers to satisfy structural legality and fully utilize model budgets. The evaluation shows that by replacing standard convolutions with generated new kernels in common NNs, Canvas achieves average 1.5x speedups compared to the previous state-of-the-art with acceptable accuracy loss and search efficiency. Canvas verifies the practicability of KAS by rediscovering many manually designed kernels in the past and producing new structures that may inspire future machine learning innovations. For source code and implementation, we open-sourced Canvas at https://github.com/tsinghua-ideal/Canvas.
Authors: Michael Ogezi, Bradley Hauer, Grzegorz Kondrak
Abstract: Language-vision models like CLIP have made significant strides in vision tasks, such as zero-shot image classification (ZSIC). However, generating specific and expressive visual descriptions remains challenging; descriptions produced by current methods are often ambiguous and lacking in granularity. To tackle these issues, we propose V-GLOSS: Visual Glosses, a novel method built upon two key ideas. The first is Semantic Prompting, which conditions a language model on structured semantic knowledge. The second is a new contrastive algorithm that elicits fine-grained distinctions between similar concepts. With both ideas, we demonstrate that V-GLOSS improves visual descriptions and achieves strong results in the zero-shot setting on general and fine-grained image-classification datasets, including ImageNet, STL-10, FGVC Aircraft, and Flowers 102. Moreover, these descriptive capabilities contribute to enhancing image-generation performance. Finally, we introduce a quality-tested silver dataset with descriptions generated with V-GLOSS for all ImageNet classes.
Authors: Zitao Shuai, Chenwei Wu, Zhengxu Tang, Liyue Shen
Abstract: Vision-language pre-training (VLP) has emerged as an effective scheme for multimodal representation learning, but its reliance on large-scale multimodal data poses significant challenges for medical applications. Federated learning (FL) offers a promising solution to scale up the dataset for medical VLP while preserving data privacy. However, we observe that client data heterogeneity in real-world scenarios could cause models to learn biased cross-modal alignment during local pre-training. This would limit the transferability of the federally learned representation model on downstream tasks. To address this challenge, we propose Federated Distributionally Robust Alignment (FedDRA), a framework for federated VLP that achieves robust vision-language alignment under heterogeneous conditions. Based on client datasets, we construct a distribution family that encompasses potential test-time domains, and apply a distributionally robust framework to optimize the pre-trained model's performance across this distribution space. This approach bridges the gap between pre-training samples and downstream applications. To avoid over-fitting on client-specific information, we use anchor representation from the global model to guide the local training, and adopt a two-stage approach to first tune deeper layers before updating the entire network. Extensive experiments on real-world datasets demonstrate FedDRA's effectiveness in enhancing medical federated VLP under data heterogeneity. Our method also adapts well to various medical pre-training methods.
Authors: Aadesh Salecha, Molly E. Ireland, Shashanka Subrahmanya, Jo\~ao Sedoc, Lyle H. Ungar, Johannes C. Eichstaedt
Abstract: As Large Language Models (LLMs) become widely used to model and simulate human behavior, understanding their biases becomes critical. We developed an experimental framework using Big Five personality surveys and uncovered a previously undetected social desirability bias in a wide range of LLMs. By systematically varying the number of questions LLMs were exposed to, we demonstrate their ability to infer when they are being evaluated. When personality evaluation is inferred, LLMs skew their scores towards the desirable ends of trait dimensions (i.e., increased extraversion, decreased neuroticism, etc). This bias exists in all tested models, including GPT-4/3.5, Claude 3, Llama 3, and PaLM-2. Bias levels appear to increase in more recent models, with GPT-4's survey responses changing by 1.20 (human) standard deviations and Llama 3's by 0.98 standard deviations-very large effects. This bias is robust to randomization of question order and paraphrasing. Reverse-coding all the questions decreases bias levels but does not eliminate them, suggesting that this effect cannot be attributed to acquiescence bias. Our findings reveal an emergent social desirability bias and suggest constraints on profiling LLMs with psychometric tests and on using LLMs as proxies for human participants.
Authors: Xinmeng Huang, Shuo Li, Edgar Dobriban, Osbert Bastani, Hamed Hassani, Dongsheng Ding
Abstract: The growing safety concerns surrounding large language models raise an urgent need to align them with diverse human preferences to simultaneously enhance their helpfulness and safety. A promising approach is to enforce safety constraints through Reinforcement Learning from Human Feedback (RLHF). For such constrained RLHF, typical Lagrangian-based primal-dual policy optimization methods are computationally expensive and often unstable. This paper presents a perspective of dualization that reduces constrained alignment to an equivalent unconstrained alignment problem. We do so by pre-optimizing a smooth and convex dual function that has a closed form. This shortcut eliminates the need for cumbersome primal-dual policy iterations, greatly reducing the computational burden and improving training stability. Our strategy leads to two practical algorithms in model-based and preference-based settings (MoCAN and PeCAN, respectively). A broad range of experiments demonstrate the effectiveness and merits of our algorithms.
Authors: Messi H. J. Lee, Jacob M. Montgomery, Calvin K. Lai
Abstract: Vision Language Models (VLMs), exemplified by GPT-4V, adeptly integrate text and vision modalities. This integration enhances Large Language Models' ability to mimic human perception, allowing them to process image inputs. Despite VLMs' advanced capabilities, however, there is a concern that VLMs inherit biases of both modalities in ways that make biases more pervasive and difficult to mitigate. Our study explores how VLMs perpetuate homogeneity bias and trait associations with regards to race and gender. When prompted to write stories based on images of human faces, GPT-4V describes subordinate racial and gender groups with greater homogeneity than dominant groups and relies on distinct, yet generally positive, stereotypes. Importantly, VLM stereotyping is driven by visual cues rather than group membership alone such that faces that are rated as more prototypically Black and feminine are subject to greater stereotyping. These findings suggest that VLMs may associate subtle visual cues related to racial and gender groups with stereotypes in ways that could be challenging to mitigate. We explore the underlying reasons behind this behavior and discuss its implications and emphasize the importance of addressing these biases as VLMs come to mirror human perception.
Authors: Andrea Failla, Salvatore Citraro, Giulio Rossetti, Francesco Cauteruccio
Abstract: In recent years, the proliferation of social platforms has drastically transformed the way individuals interact, organize, and share information. In this scenario, we experience an unprecedented increase in the scale and complexity of interactions and, at the same time, little to no research about some fringe social platforms. In this paper, we present a multi-dimensional framework for characterizing nodes and hyperedges in social hypernetworks, with a focus on the understudied alt-right platform Scored.co. Our approach integrates the possibility of studying higher-order interactions, thanks to the hypernetwork representation, and various node features such as user activity, sentiment, and toxicity, with the aim to define distinct user archetypes and understand their roles within the network. Utilizing a comprehensive dataset from Scored.co, we analyze the dynamics of these archetypes over time and explore their interactions and influence within the community. The framework's versatility allows for detailed analysis of both individual user behaviors and broader social structures. Our findings highlight the importance of higher-order interactions in understanding social dynamics, offering new insights into the roles and behaviors that emerge in complex online environments.
Authors: Kosuke Imai, Kentaro Nakamura
Abstract: In this paper, we demonstrate how to enhance the validity of causal inference with unstructured high-dimensional treatments like texts, by leveraging the power of generative Artificial Intelligence. Specifically, we propose to use a deep generative model such as large language models (LLMs) to efficiently generate treatments and use their internal representation for subsequent causal effect estimation. We show that the knowledge of this true internal representation helps disentangle the treatment features of interest, such as specific sentiments and certain topics, from other possibly unknown confounding features. Unlike the existing methods, our proposed approach eliminates the need to learn causal representation from the data and hence produces more accurate and efficient estimates. We formally establish the conditions required for the nonparametric identification of the average treatment effect, propose an estimation strategy that avoids the violation of the overlap assumption, and derive the asymptotic properties of the proposed estimator through the application of double machine learning. Finally, using an instrumental variables approach, we extend the proposed methodology to the settings, in which the treatment feature is based on human perception rather than is assumed to be fixed given the treatment object. The proposed methodology is also applicable to text reuse where an LLM is used to regenerate the existing texts. We conduct simulation and empirical studies, using the generated text data from an open-source LLM, Llama 3, to illustrate the advantages of our estimator over the state-of-the-art causal representation learning algorithms.
Authors: Pau Rodriguez, Arno Blaas, Michal Klein, Luca Zappella, Nicholas Apostoloff, Marco Cuturi, Xavier Suau
Abstract: The increasing capabilities of large generative models and their ever more widespread deployment have raised concerns about their reliability, safety, and potential misuse. To address these issues, recent works have proposed to control model generation by steering model activations in order to effectively induce or prevent the emergence of concepts or behaviors in the generated output. In this paper we introduce Activation Transport (AcT), a general framework to steer activations guided by optimal transport theory that generalizes many previous activation-steering works. AcT is modality-agnostic and provides fine-grained control over the model behavior with negligible computational overhead, while minimally impacting model abilities. We experimentally show the effectiveness and versatility of our approach by addressing key challenges in large language models (LLMs) and text-to-image diffusion models (T2Is). For LLMs, we show that AcT can effectively mitigate toxicity, induce arbitrary concepts, and increase their truthfulness. In T2Is, we show how AcT enables fine-grained style control and concept negation.
Authors: Zhirui Deng, Zhicheng Dou, Yutao Zhu, Ji-Rong Wen, Ruibin Xiong, Mang Wang, Weipeng Chen
Abstract: The outstanding capabilities of large language models (LLMs) render them a crucial component in various autonomous agent systems. While traditional methods depend on the inherent knowledge of LLMs without fine-tuning, more recent approaches have shifted toward the reinforcement learning strategy to further enhance agents' ability to solve complex interactive tasks with environments and tools. However, previous approaches are constrained by the sparse reward issue, where existing datasets solely provide a final scalar reward for each multi-step reasoning chain, potentially leading to ineffectiveness and inefficiency in policy learning. In this paper, we introduce StepAgent, which utilizes step-wise reward to optimize the agent's reinforcement learning process. Inheriting the spirit of novice-to-expert theory, we first compare the actions of the expert and the agent to automatically generate intermediate rewards for fine-grained optimization. Additionally, we propose implicit-reward and inverse reinforcement learning techniques to facilitate agent reflection and policy adjustment. Further theoretical analysis demonstrates that the action distribution of the agent can converge toward the expert action distribution over multiple training cycles. Experimental results across various datasets indicate that StepAgent outperforms existing baseline methods.
Authors: Haolin Chen, Yihao Feng, Zuxin Liu, Weiran Yao, Akshara Prabhakar, Shelby Heinecke, Ricky Ho, Phil Mui, Silvio Savarese, Caiming Xiong, Huan Wang
Abstract: Large language models (LLMs) have shown impressive capabilities, but still struggle with complex reasoning tasks requiring multiple steps. While prompt-based methods like Chain-of-Thought (CoT) can improve LLM reasoning at inference time, optimizing reasoning capabilities during training remains challenging. We introduce LaTent Reasoning Optimization (LaTRO), a principled framework that formulates reasoning as sampling from a latent distribution and optimizes it via variational approaches. LaTRO enables LLMs to concurrently improve both their reasoning process and ability to evaluate reasoning quality, without requiring external feedback or reward models. We validate LaTRO through experiments on GSM8K and ARC-Challenge datasets using multiple model architectures. On GSM8K, LaTRO improves zero-shot accuracy by an average of 12.5% over base models and 9.6% over supervised fine-tuning across Phi-3.5-mini, Mistral-7B, and Llama-3.1-8B. Our findings suggest that pre-trained LLMs possess latent reasoning capabilities that can be unlocked and enhanced through our proposed optimization approach in a self-improvement manner. The code of LaTRO is available at \url{https://github.com/SalesforceAIResearch/LaTRO}.
Authors: Jaehyeok Lee, Keisuke Sakaguchi, JinYeong Bak
Abstract: Self-training approach for large language models (LLMs) improves reasoning abilities by training the models on their self-generated rationales. Previous approaches have labeled rationales that produce correct answers for a given question as appropriate for training. However, a single measure risks misjudging rationale quality, leading the models to learn flawed reasoning patterns. To address this issue, we propose CREST (Consistency-driven Rationale Evaluation for Self-Training), a self-training framework that further evaluates each rationale through follow-up questions and leverages this evaluation to guide its training. Specifically, we introduce two methods: (1) filtering out rationales that frequently result in incorrect answers on follow-up questions and (2) preference learning based on mixed preferences from rationale evaluation results of both original and follow-up questions. Experiments on three question-answering datasets using open LLMs show that CREST not only improves the logical robustness and correctness of rationales but also improves reasoning abilities compared to previous self-training approaches.
Authors: Shayan Alipour, Indira Sen, Mattia Samory, Tanushree Mitra
Abstract: Large language models (LLMs) are known to exhibit demographic biases, yet few studies systematically evaluate these biases across multiple datasets or account for confounding factors. In this work, we examine LLM alignment with human annotations in five offensive language datasets, comprising approximately 220K annotations. Our findings reveal that while demographic traits, particularly race, influence alignment, these effects are inconsistent across datasets and often entangled with other factors. Confounders -- such as document difficulty, annotator sensitivity, and within-group agreement -- account for more variation in alignment patterns than demographic traits alone. Specifically, alignment increases with higher annotator sensitivity and group agreement, while greater document difficulty corresponds to reduced alignment. Our results underscore the importance of multi-dataset analyses and confounder-aware methodologies in developing robust measures of demographic bias in LLMs.