Authors: Xiongbin Gui, Hanlin Lv, Xiao Wang, Longting Lv, Yi Xiao, Lei Wang
Abstract: Background: Recruitment for cohorts involving complex liver diseases, such as hepatocellular carcinoma and liver cirrhosis, often requires interpreting semantically complex criteria. Traditional manual screening methods are time-consuming and prone to errors. While AI-powered pre-screening offers potential solutions, challenges remain regarding accuracy, efficiency, and data privacy. Methods: We developed a novel patient pre-screening pipeline that leverages clinical expertise to guide the precise, safe, and efficient application of large language models. The pipeline breaks down complex criteria into a series of composite questions and then employs two strategies to perform semantic question-answering through electronic health records - (1) Pathway A, Anthropomorphized Experts' Chain of Thought strategy, and (2) Pathway B, Preset Stances within an Agent Collaboration strategy, particularly in managing complex clinical reasoning scenarios. The pipeline is evaluated on three key metrics-precision, time consumption, and counterfactual inference - at both the question and criterion levels. Results: Our pipeline achieved high precision (0.921, in criteria level) and efficiency (0.44s per task). Pathway B excelled in complex reasoning, while Pathway A was effective in precise data extraction with faster processing times. Both pathways achieved comparable precision. The pipeline showed promising results in hepatocellular carcinoma (0.878) and cirrhosis trials (0.843). Conclusions: This data-secure and time-efficient pipeline shows high precision in hepatopathy trials, providing promising solutions for streamlining clinical trial workflows. Its efficiency and adaptability make it suitable for improving patient recruitment. And its capability to function in resource-constrained environments further enhances its utility in clinical settings.
Authors: Shuming Shi, Ruobing Zuo, Gaolei He, Jianlin Wang, Chenyang Xu, Zhengfeng Yang
Abstract: Automated theorem proving (ATP) is one of the most challenging mathematical reasoning tasks for Large Language Models (LLMs). Most existing LLM-based ATP methods rely on supervised fine-tuning, which results in a limited alignment between the theorem proving process and human preferences. Direct Preference Optimization (DPO), which aligns LLMs with human preferences, has shown positive effects for certain tasks. However, the lack of high-quality preference data for theorem proving presents a significant challenge. In this paper, we innovatively apply DPO to formal automated theorem proving and introduces a Curriculum Learning-based DPO Iterative Theorem Proving (CuDIP) method. Specifically, we propose a method for constructing preference data which utilizes LLMs and existing theorem proving data to enhance the diversity of the preference data while reducing the reliance on human preference annotations. We then integrate this preference data construction method with curriculum learning to iteratively fine-tune the theorem proving model through DPO. Experimental results on the MiniF2F and ProofNet datasets demonstrate the effectiveness of the proposed method.
Authors: Zhangqi Duan, Nigel Fernandez, Sri Kanakadandi, Bita Akram, Andrew Lan
Abstract: Knowledge components (KCs) mapped to problems help model student learning, tracking their mastery levels on fine-grained skills thereby facilitating personalized learning and feedback in online learning platforms. However, crafting and tagging KCs to problems, traditionally performed by human domain experts, is highly labor-intensive. We present a fully automated, LLM-based pipeline for KC generation and tagging for open-ended programming problems. We also develop an LLM-based knowledge tracing (KT) framework to leverage these LLM-generated KCs, which we refer to as KCGen-KT. We conduct extensive quantitative and qualitative evaluations validating the effectiveness of KCGen-KT. On a real-world dataset of student code submissions to open-ended programming problems, KCGen-KT outperforms existing KT methods. We investigate the learning curves of generated KCs and show that LLM-generated KCs have a comparable level-of-fit to human-written KCs under the performance factor analysis (PFA) model. We also conduct a human evaluation to show that the KC tagging accuracy of our pipeline is reasonably accurate when compared to that by human domain experts.
Authors: Fengze Yang (Dylan), Xiaoyue Cathy Liu (Dylan), Lingjiu Lu (Dylan), Bingzhang Wang (Dylan), Chenxi (Dylan), Liu
Abstract: With the urbanization process, an increasing number of sensors are being deployed in transportation systems, leading to an explosion of big data. To harness the power of this vast transportation data, various machine learning (ML) and artificial intelligence (AI) methods have been introduced to address numerous transportation challenges. However, these methods often require significant investment in data collection, processing, storage, and the employment of professionals with expertise in transportation and ML. Additionally, privacy issues are a major concern when processing data for real-world traffic control and management. To address these challenges, the research team proposes an innovative Multi-agent framework named Independent Mobility GPT (IDM-GPT) based on large language models (LLMs) for customized traffic analysis, management suggestions, and privacy preservation. IDM-GPT efficiently connects users, transportation databases, and ML models economically. IDM-GPT trains, customizes, and applies various LLM-based AI agents for multiple functions, including user query comprehension, prompts optimization, data analysis, model selection, and performance evaluation and enhancement. With IDM-GPT, users without any background in transportation or ML can efficiently and intuitively obtain data analysis and customized suggestions in near real-time based on their questions. Experimental results demonstrate that IDM-GPT delivers satisfactory performance across multiple traffic-related tasks, providing comprehensive and actionable insights that support effective traffic management and urban mobility improvement.
Authors: Shramay Palta, Nirupama Chandrasekaran, Rachel Rudinger, Scott Counts
Abstract: Using a sample of 25,000 Bing Copilot conversations, we study how the agent responds to users of varying levels of domain expertise and the resulting impact on user experience along multiple dimensions. Our findings show that across a variety of topical domains, the agent largely responds at proficient or expert levels of expertise (77% of conversations) which correlates with positive user experience regardless of the user's level of expertise. Misalignment, such that the agent responds at a level of expertise below that of the user, has a negative impact on overall user experience, with the impact more profound for more complex tasks. We also show that users engage more, as measured by the number of words in the conversation, when the agent responds at a level of expertise commensurate with that of the user. Our findings underscore the importance of alignment between user and AI when designing human-centered AI systems, to ensure satisfactory and productive interactions.
Authors: Daniel Weiner, Raj Korpan
Abstract: In role-playing games (RPGs), the level of immersion is critical-especially when an in-game agent conveys tasks, hints, or ideas to the player. For an agent to accurately interpret the player's emotional state and contextual nuances, a foundational level of understanding is required, which can be achieved using a Large Language Model (LLM). Maintaining the LLM's focus across multiple context changes, however, necessitates a more robust approach, such as integrating the LLM with a dedicated task allocation model to guide its performance throughout gameplay. In response to this need, we introduce Voting-Based Task Assignment (VBTA), a framework inspired by human reasoning in task allocation and completion. VBTA assigns capability profiles to agents and task descriptions to tasks, then generates a suitability matrix that quantifies the alignment between an agent's abilities and a task's requirements. Leveraging six distinct voting methods, a pre-trained LLM, and integrating conflict-based search (CBS) for path planning, VBTA efficiently identifies and assigns the most suitable agent to each task. While existing approaches focus on generating individual aspects of gameplay, such as single quests, or combat encounters, our method shows promise when generating both unique combat encounters and narratives because of its generalizable nature.
Authors: Chenlu Ju, Jiaxin Liu, Shobhit Sinha, Hao Xue, Flora Salim
Abstract: This work leverages Large Language Models (LLMs) to simulate human mobility, addressing challenges like high costs and privacy concerns in traditional models. Our hierarchical framework integrates persona generation, activity selection, and destination prediction, using real-world demographic and psychological data to create realistic movement patterns. Both physical models and language models are employed to explore and demonstrate different methodologies for human mobility simulation. By structuring data with summarization and weighted density metrics, the system ensures scalable memory management while retaining actionable insights. Preliminary results indicate that LLM-driven simulations align with observed real-world patterns, offering scalable, interpretable insights for social problems such as urban planning, traffic management, and public health. The framework's ability to dynamically generate personas and activities enables it to provide adaptable and realistic daily routines. This study demonstrates the transformative potential of LLMs in advancing mobility modeling for societal and urban applications. The source code and interactive demo for our framework are available at https://github.com/cju0/TrajLLM.
Authors: Xin Liu, Ziyue Zhang, Jingxin Nie
Abstract: Traditional psychological experiments utilizing naturalistic stimuli face challenges in manual annotation and ecological validity. To address this, we introduce a novel paradigm leveraging multimodal large language models (LLMs) as proxies to extract rich semantic information from naturalistic images through a Visual Question Answering (VQA) strategy for analyzing human visual semantic representation. LLM-derived representations successfully predict established neural activity patterns measured by fMRI (e.g., faces, buildings), validating its feasibility and revealing hierarchical semantic organization across cortical regions. A brain semantic network constructed from LLM-derived representations identifies meaningful clusters reflecting functional and contextual associations. This innovative methodology offers a powerful solution for investigating brain semantic organization with naturalistic stimuli, overcoming limitations of traditional annotation methods and paving the way for more ecologically valid explorations of human cognition.
Authors: Jeesu Jung, Chanjun Park, Sangkeun Jung
Abstract: Recent advancements in Large Language Model (LLM) alignment have sought to mitigate the cost of human annotations by leveraging pretrained models to generate preference data. However, existing methods often compare responses from models with substantially different capabilities, yielding superficial distinctions that fail to provide meaningful guidance on what constitutes a superior response. To address this limitation, we propose Kinship-Aware pReference MApping (KARMA), a novel framework that systematically pairs responses from models with comparable competencies. By constraining preference comparisons to outputs of similar complexity and quality, KARMA enhances the informativeness of preference data and improves the granularity of alignment signals. Empirical evaluations demonstrate that our kinship-aware approach leads to more consistent and interpretable alignment outcomes, ultimately facilitating a more principled and reliable pathway for aligning LLM behavior with human preferences.
Authors: Weipeng Jiang, Juan Zhai, Shiqing Ma, Ziyan Lei, Xiaofei Xie, Yige Wang, Chao Shen
Abstract: In recent years, Large Language Models (LLMs) have faced increasing demands to selectively remove sensitive information, protect privacy, and comply with copyright regulations through unlearning, by Machine Unlearning. While evaluating unlearning effectiveness is crucial, existing benchmarks are limited in scale and comprehensiveness, typically containing only a few hundred test cases. We identify two critical challenges in generating holistic audit datasets: ensuring audit adequacy and handling knowledge redundancy between forget and retain dataset. To address these challenges, we propose HANKER, an automated framework for holistic audit dataset generation leveraging knowledge graphs to achieve fine-grained coverage and eliminate redundant knowledge. Applying HANKER to the popular MUSE benchmark, we successfully generated over 69,000 and 111,000 audit cases for the News and Books datasets respectively, identifying thousands of knowledge memorization instances that the previous benchmark failed to detect. Our empirical analysis uncovers how knowledge redundancy significantly skews unlearning effectiveness metrics, with redundant instances artificially inflating the observed memorization measurements ROUGE from 19.7% to 26.1% and Entailment Scores from 32.4% to 35.2%, highlighting the necessity of systematic deduplication for accurate assessment.
Authors: Huangyuan Su, Aaron Walsman, Daniel Garces, Sham Kakade, Stephanie Gil
Abstract: In this project, our goal is to determine how to leverage the world-knowledge of pretrained large language models for efficient and robust learning in multiagent decision making. We examine this in a taxi routing and assignment problem where agents must decide how to best pick up passengers in order to minimize overall waiting time. While this problem is situated on a graphical road network, we show that with the proper prompting zero-shot performance is quite strong on this task. Furthermore, with limited fine-tuning along with the one-at-a-time rollout algorithm for look ahead, LLMs can out-compete existing approaches with 50 times fewer environmental interactions. We also explore the benefits of various linguistic prompting approaches and show that including certain easy-to-compute information in the prompt significantly improves performance. Finally, we highlight the LLM's built-in semantic understanding, showing its ability to adapt to environmental factors through simple prompts.
Authors: Longling Geng, Edward Y. Chang
Abstract: This benchmark suite provides a comprehensive evaluation framework for assessing both individual LLMs and multi-agent systems in real-world planning scenarios. The suite encompasses eleven designed problems that progress from basic to highly complex, incorporating key aspects such as multi-agent coordination, inter-agent dependencies, and dynamic environmental disruptions. Each problem can be scaled along three dimensions: the number of parallel planning threads, the complexity of inter-dependencies, and the frequency of unexpected disruptions requiring real-time adaptation. The benchmark includes detailed specifications, evaluation metrics, and baseline implementations using contemporary frameworks like LangGraph, enabling rigorous testing of both single-agent and multi-agent planning capabilities. Through standardized evaluation criteria and scalable complexity, this benchmark aims to drive progress in developing more robust and adaptable AI planning systems for real-world applications.
Authors: Jingtao Zhan, Jiahao Zhao, Jiayu Li, Yiqun Liu, Bo Zhang, Qingyao Ai, Jiaxin Mao, Hongning Wang, Min Zhang, Shaoping Ma
Abstract: How does intelligence emerge? We propose that intelligence is not a sudden gift or random occurrence, but rather a necessary trait for species to survive through Natural Selection. If a species passes the test of Natural Selection, it demonstrates the intelligence to survive in nature. Extending this perspective, we introduce Intelligence Test, a method to quantify the intelligence of any subject on any task. Like how species evolve by trial and error, Intelligence Test quantifies intelligence by the number of failed attempts before success. Fewer failures correspond to higher intelligence. When the expectation and variance of failure counts are both finite, it signals the achievement of an autonomous level of intelligence. Using Intelligence Test, we comprehensively evaluate existing AI systems. Our results show that while AI systems achieve a level of autonomy in simple tasks, they are still far from autonomous in more complex tasks, such as vision, search, recommendation, and language. While scaling model size might help, this would come at an astronomical cost. Projections suggest that achieving general autonomy would require unimaginable $10^{26}$ parameters. Even if Moore's Law continuously holds, such a parameter scale would take $70$ years. This staggering cost highlights the complexity of human tasks and the inadequacies of current AI. To further understand this phenomenon, we conduct a theoretical analysis. Our simulations suggest that human tasks possess a criticality property. As a result, autonomy requires a deep understanding of the task's underlying mechanisms. Current AI, however, does not fully grasp these mechanisms and instead relies on superficial mimicry, making it difficult to reach an autonomous level. We believe Intelligence Test can not only guide the future development of AI but also offer profound insights into the intelligence of humans ourselves.
Authors: Juraj Gottweis, Wei-Hung Weng, Alexander Daryin, Tao Tu, Anil Palepu, Petar Sirkovic, Artiom Myaskovsky, Felix Weissenberger, Keran Rong, Ryutaro Tanno, Khaled Saab, Dan Popovici, Jacob Blum, Fan Zhang, Katherine Chou, Avinatan Hassidim, Burak Gokturk, Amin Vahdat, Pushmeet Kohli, Yossi Matias, Andrew Carroll, Kavita Kulkarni, Nenad Tomasev, Yuan Guan, Vikram Dhillon, Eeshit Dhaval Vaishnav, Byron Lee, Tiago R D Costa, Jos\'e R Penad\'es, Gary Peltz, Yunhan Xu, Annalisa Pawlosky, Alan Karthikesalingam, Vivek Natarajan
Abstract: Scientific discovery relies on scientists generating novel hypotheses that undergo rigorous experimental validation. To augment this process, we introduce an AI co-scientist, a multi-agent system built on Gemini 2.0. The AI co-scientist is intended to help uncover new, original knowledge and to formulate demonstrably novel research hypotheses and proposals, building upon prior evidence and aligned to scientist-provided research objectives and guidance. The system's design incorporates a generate, debate, and evolve approach to hypothesis generation, inspired by the scientific method and accelerated by scaling test-time compute. Key contributions include: (1) a multi-agent architecture with an asynchronous task execution framework for flexible compute scaling; (2) a tournament evolution process for self-improving hypotheses generation. Automated evaluations show continued benefits of test-time compute, improving hypothesis quality. While general purpose, we focus development and validation in three biomedical areas: drug repurposing, novel target discovery, and explaining mechanisms of bacterial evolution and anti-microbial resistance. For drug repurposing, the system proposes candidates with promising validation findings, including candidates for acute myeloid leukemia that show tumor inhibition in vitro at clinically applicable concentrations. For novel target discovery, the AI co-scientist proposed new epigenetic targets for liver fibrosis, validated by anti-fibrotic activity and liver cell regeneration in human hepatic organoids. Finally, the AI co-scientist recapitulated unpublished experimental results via a parallel in silico discovery of a novel gene transfer mechanism in bacterial evolution. These results, detailed in separate, co-timed reports, demonstrate the potential to augment biomedical and scientific discovery and usher an era of AI empowered scientists.
Authors: Sen Yang, Yafu Li, Wai Lam, Yu Cheng
Abstract: Large language models (LLMs) often struggle with complex reasoning tasks due to their limitations in addressing the vast reasoning space and inherent ambiguities of natural language. We propose the Mixture-of-Search-Agents (MoSA) paradigm, a novel approach leveraging the collective expertise of multiple LLMs to enhance search-based reasoning. MoSA integrates diverse reasoning pathways by combining independent exploration with iterative refinement among LLMs, mitigating the limitations of single-model approaches. Using Monte Carlo Tree Search (MCTS) as a backbone, MoSA enables multiple agents to propose and aggregate reasoning steps, resulting in improved accuracy. Our comprehensive evaluation across four reasoning benchmarks demonstrates MoSA's consistent performance improvements over single-agent and other multi-agent baselines, particularly in complex mathematical and commonsense reasoning tasks.
Authors: Achmad Anggawirya Alimin, Dominik P. Goldstein, Lukas Schulze Balhorn, Artur M. Schweidtmann
Abstract: We propose a methodology that allows communication with Piping and Instrumentation Diagrams (P&IDs) using natural language. In particular, we represent P&IDs through the DEXPI data model as labeled property graphs and integrate them with Large Language Models (LLMs). The approach consists of three main parts: 1) P&IDs are cast into a graph representation from the DEXPI format using our pyDEXPI Python package. 2) A tool for generating P&ID knowledge graphs from pyDEXPI. 3) Integration of the P&ID knowledge graph to LLMs using graph-based retrieval augmented generation (graph-RAG). This approach allows users to communicate with P&IDs using natural language. It extends LLM's ability to retrieve contextual data from P&IDs and mitigate hallucinations. Leveraging the LLM's large corpus, the model is also able to interpret process information in PIDs, which could help engineers in their daily tasks. In the future, this work will also open up opportunities in the context of other generative Artificial Intelligence (genAI) solutions on P&IDs, and AI-assisted HAZOP studies.
Authors: Anastasios Nentidis, Charilaos Akasiadis, Angelos Charalambidis, Alexander Artikis
Abstract: In Knowledge Graphs (KGs), where the schema of the data is usually defined by particular ontologies, reasoning is a necessity to perform a range of tasks, such as retrieval of information, question answering, and the derivation of new knowledge. However, information to populate KGs is often extracted (semi-) automatically from natural language resources, or by integrating datasets that follow different semantic schemas, resulting in KG inconsistency. This, however, hinders the process of reasoning. In this survey, we focus on how to perform reasoning on inconsistent KGs, by analyzing the state of the art towards three complementary directions: a) the detection of the parts of the KG that cause the inconsistency, b) the fixing of an inconsistent KG to render it consistent, and c) the inconsistency-tolerant reasoning. We discuss existing work from a range of relevant fields focusing on how, and in which cases they are related to the above directions. We also highlight persisting challenges and future directions.
Authors: Humza Sami, Mubashir ul Islam, Samy Charas, Asav Gandhi, Pierre-Emmanuel Gaillardon, Valerio Tenace
Abstract: Recent advancements in Large Language Models (LLMs) have substantially evolved Multi-Agent Systems (MASs) capabilities, enabling systems that not only automate tasks but also leverage near-human reasoning capabilities. To achieve this, LLM-based MASs need to be built around two critical principles: (i) a robust architecture that fully exploits LLM potential for specific tasks -- or related task sets -- and ($ii$) an effective methodology for equipping LLMs with the necessary capabilities to perform tasks and manage information efficiently. It goes without saying that a priori architectural designs can limit the scalability and domain adaptability of a given MAS. To address these challenges, in this paper we introduce Nexus: a lightweight Python framework designed to easily build and manage LLM-based MASs. Nexus introduces the following innovations: (i) a flexible multi-supervisor hierarchy, (ii) a simplified workflow design, and (iii) easy installation and open-source flexibility: Nexus can be installed via pip and is distributed under a permissive open-source license, allowing users to freely modify and extend its capabilities. Experimental results demonstrate that architectures built with Nexus exhibit state-of-the-art performance across diverse domains. In coding tasks, Nexus-driven MASs achieve a 99% pass rate on HumanEval and a flawless 100% on VerilogEval-Human, outperforming cutting-edge reasoning language models such as o3-mini and DeepSeek-R1. Moreover, these architectures display robust proficiency in complex reasoning and mathematical problem solving, achieving correct solutions for all randomly selected problems from the MATH dataset. In the realm of multi-objective optimization, Nexus-based architectures successfully address challenging timing closure tasks on designs from the VTR benchmark suite, while guaranteeing, on average, a power saving of nearly 30%.
Authors: Enrico Saccon, Ahmet Tikna, Davide De Martini, Edoardo Lamon, Luigi Palopoli, Marco Roveri
Abstract: This paper presents a novel framework, called PLANTOR (PLanning with Natural language for Task-Oriented Robots), that integrates Large Language Models (LLMs) with Prolog-based knowledge management and planning for multi-robot tasks. The system employs a two-phase generation of a robot-oriented knowledge base, ensuring reusability and compositional reasoning, as well as a three-step planning procedure that handles temporal dependencies, resource constraints, and parallel task execution via mixed-integer linear programming. The final plan is converted into a Behaviour Tree for direct use in ROS2. We tested the framework in multi-robot assembly tasks within a block world and an arch-building scenario. Results demonstrate that LLMs can produce accurate knowledge bases with modest human feedback, while Prolog guarantees formal correctness and explainability. This approach underscores the potential of LLM integration for advanced robotics tasks requiring flexible, scalable, and human-understandable planning.
Authors: Pierre Peigne-Lefebvre, Mikolaj Kniejski, Filip Sondej, Matthieu David, Jason Hoelscher-Obermaier, Christian Schroeder de Witt, Esben Kran
Abstract: As AI agents are increasingly adopted to collaborate on complex objectives, ensuring the security of autonomous multi-agent systems becomes crucial. We develop simulations of agents collaborating on shared objectives to study these security risks and security trade-offs. We focus on scenarios where an attacker compromises one agent, using it to steer the entire system toward misaligned outcomes by corrupting other agents. In this context, we observe infectious malicious prompts - the multi-hop spreading of malicious instructions. To mitigate this risk, we evaluated several strategies: two "vaccination" approaches that insert false memories of safely handling malicious input into the agents' memory stream, and two versions of a generic safety instruction strategy. While these defenses reduce the spread and fulfillment of malicious instructions in our experiments, they tend to decrease collaboration capability in the agent network. Our findings illustrate potential trade-off between security and collaborative efficiency in multi-agent systems, providing insights for designing more secure yet effective AI collaborations.
Authors: Hongyi Ling, Shubham Parashar, Sambhav Khurana, Blake Olson, Anwesha Basu, Gaurangi Sinha, Zhengzhong Tu, James Caverlee, Shuiwang Ji
Abstract: We consider enhancing large language models (LLMs) for complex planning tasks. While existing methods allow LLMs to explore intermediate steps to make plans, they either depend on unreliable self-verification or external verifiers to evaluate these steps, which demand significant data and computations. Here, we propose automated heuristics discovery (AutoHD), a novel approach that enables LLMs to explicitly generate heuristic functions to guide inference-time search, allowing accurate evaluation of intermediate states. These heuristic functions are further refined through a heuristic evolution process, improving their robustness and effectiveness. Our proposed method requires no additional model training or fine-tuning, and the explicit definition of heuristic functions generated by the LLMs provides interpretability and insights into the reasoning process. Extensive experiments across diverse benchmarks demonstrate significant gains over multiple baselines, including nearly twice the accuracy on some datasets, establishing our approach as a reliable and interpretable solution for complex planning tasks.
Authors: William Solow, Sandhya Saisubramanian, Alan Fern
Abstract: We introduce WOFOSTGym, a novel crop simulation environment designed to train reinforcement learning (RL) agents to optimize agromanagement decisions for annual and perennial crops in single and multi-farm settings. Effective crop management requires optimizing yield and economic returns while minimizing environmental impact, a complex sequential decision-making problem well suited for RL. However, the lack of simulators for perennial crops in multi-farm contexts has hindered RL applications in this domain. Existing crop simulators also do not support multiple annual crops. WOFOSTGym addresses these gaps by supporting 23 annual crops and two perennial crops, enabling RL agents to learn diverse agromanagement strategies in multi-year, multi-crop, and multi-farm settings. Our simulator offers a suite of challenging tasks for learning under partial observability, non-Markovian dynamics, and delayed feedback. WOFOSTGym's standard RL interface allows researchers without agricultural expertise to explore a wide range of agromanagement problems. Our experiments demonstrate the learned behaviors across various crop varieties and soil types, highlighting WOFOSTGym's potential for advancing RL-driven decision support in agriculture.
Authors: Qi Yu, Zhichen Zeng, Yuchen Yan, Lei Ying, R. Srikant, Hanghang Tong
Abstract: Network alignment, which aims to find node correspondence across different networks, is the cornerstone of various downstream multi-network and Web mining tasks. Most of the embedding-based methods indirectly model cross-network node relationships by contrasting positive and negative node pairs sampled from hand-crafted strategies, which are vulnerable to graph noises and lead to potential misalignment of nodes. Another line of work based on the optimal transport (OT) theory directly models cross-network node relationships and generates noise-reduced alignments. However, OT methods heavily rely on fixed, pre-defined cost functions that prohibit end-to-end training and are hard to generalize. In this paper, we aim to unify the embedding and OT-based methods in a mutually beneficial manner and propose a joint optimal transport and embedding framework for network alignment named JOENA. For one thing (OT for embedding), through a simple yet effective transformation, the noise-reduced OT mapping serves as an adaptive sampling strategy directly modeling all cross-network node pairs for robust embedding learning.For another (embedding for OT), on top of the learned embeddings, the OT cost can be gradually trained in an end-to-end fashion, which further enhances the alignment quality. With a unified objective, the mutual benefits of both methods can be achieved by an alternating optimization schema with guaranteed convergence. Extensive experiments on real-world networks validate the effectiveness and scalability of JOENA, achieving up to 16% improvement in MRR and 20x speedup compared with the state-of-the-art alignment methods.
Authors: Max Ku, Thomas Chong, Jonathan Leung, Krish Shah, Alvin Yu, Wenhu Chen
Abstract: Understanding domain-specific theorems often requires more than just text-based reasoning; effective communication through structured visual explanations is crucial for deeper comprehension. While large language models (LLMs) demonstrate strong performance in text-based theorem reasoning, their ability to generate coherent and pedagogically meaningful visual explanations remains an open challenge. In this work, we introduce TheoremExplainAgent, an agentic approach for generating long-form theorem explanation videos (over 5 minutes) using Manim animations. To systematically evaluate multimodal theorem explanations, we propose TheoremExplainBench, a benchmark covering 240 theorems across multiple STEM disciplines, along with 5 automated evaluation metrics. Our results reveal that agentic planning is essential for generating detailed long-form videos, and the o3-mini agent achieves a success rate of 93.8% and an overall score of 0.77. However, our quantitative and qualitative studies show that most of the videos produced exhibit minor issues with visual element layout. Furthermore, multimodal explanations expose deeper reasoning flaws that text-based explanations fail to reveal, highlighting the importance of multimodal explanations.
Authors: Karthik Shivashankar, Antonio Martini
Abstract: Machine learning (ML) codebases face unprecedented challenges in maintaining code quality and sustainability as their complexity grows exponentially. While traditional code smell detection tools exist, they fail to address ML-specific issues that can significantly impact model performance, reproducibility, and maintainability. This paper introduces MLScent, a novel static analysis tool that leverages sophisticated Abstract Syntax Tree (AST) analysis to detect anti-patterns and code smells specific to ML projects. MLScent implements 76 distinct detectors across major ML frameworks including TensorFlow (13 detectors), PyTorch (12 detectors), Scikit-learn (9 detectors), and Hugging Face (10 detectors), along with data science libraries like Pandas and NumPy (8 detectors each). The tool's architecture also integrates general ML smell detection (16 detectors), and specialized analysis for data preprocessing and model training workflows. Our evaluation demonstrates MLScent's effectiveness through both quantitative classification metrics and qualitative assessment via user studies feedback with ML practitioners. Results show high accuracy in identifying framework-specific anti-patterns, data handling issues, and general ML code smells across real-world projects.
Authors: Md Motaleb Hossen Manik
Abstract: Background: AI-powered code generation, fueled by Large Language Models (LLMs), is revolutionizing software development. Models like OpenAI's Codex and GPT-4, alongside DeepSeek, leverage vast code and natural language datasets. However, ensuring code quality, correctness, and managing complex tasks remains challenging, necessitating thorough evaluation. Methodology: This research compares ChatGPT (version o1) and DeepSeek (version R1) for Python code generation using online judge coding challenges. It evaluates correctness (online judge verdicts, up to three attempts), code quality (Pylint/Flake8), and efficiency (execution time/memory usage). Results: DeepSeek demonstrated higher correctness, particularly on algorithmic tasks, often achieving 'Accepted' on the first attempt. ChatGPT sometimes requires multiple attempts or failures. ChatGPT encountered fewer issues, used comparable or slightly less memory, consumed less execution times and wrote fewer lines of code. Conclusion: DeepSeek exhibited superior correctness in Python code generation, often requiring fewer attempts, suggesting an advantage in algorithmic problem-solving. Both models showed almost similar efficiency in execution time and memory use. Finally, this research provides insights for developers choosing AI coding assistants and informs future AI-driven software development research.
Authors: Ariful Haque, Sunzida Siddique, Md. Mahfuzur Rahman, Ahmed Rafi Hasan, Laxmi Rani Das, Marufa Kamal, Tasnim Masura, Kishor Datta Gupta
Abstract: The integration of Large Language Models (LLMs) such as GitHub Copilot, ChatGPT, Cursor AI, and Codeium AI into software development has revolutionized the coding landscape, offering significant productivity gains, automation, and enhanced debugging capabilities. These tools have proven invaluable for generating code snippets, refactoring existing code, and providing real-time support to developers. However, their widespread adoption also presents notable challenges, particularly in terms of security vulnerabilities, code quality, and ethical concerns. This paper provides a comprehensive analysis of the benefits and risks associated with AI-powered coding tools, drawing on user feedback, security analyses, and practical use cases. We explore the potential for these tools to replicate insecure coding practices, introduce biases, and generate incorrect or non-sensical code (hallucinations). In addition, we discuss the risks of data leaks, intellectual property violations and the need for robust security measures to mitigate these threats. By comparing the features and performance of these tools, we aim to guide developers in making informed decisions about their use, ensuring that the benefits of AI-assisted coding are maximized while minimizing associated risks.
Authors: Ankur Sinha, Chaitanya Agarwal, Pekka Malo
Abstract: Large language models (LLMs) excel at generating human-like responses but often struggle with interactive tasks that require access to real-time information. This limitation poses challenges in finance, where models must access up-to-date information, such as recent news or price movements, to support decision-making. To address this, we introduce Financial Agent, a knowledge-grounding approach for LLMs to handle financial queries using real-time text and tabular data. Our contributions are threefold: First, we develop a Financial Context Dataset of over 50,000 financial queries paired with the required context. Second, we train FinBloom 7B, a custom 7 billion parameter LLM, on 14 million financial news articles from Reuters and Deutsche Presse-Agentur, alongside 12 million Securities and Exchange Commission (SEC) filings. Third, we fine-tune FinBloom 7B using the Financial Context Dataset to serve as a Financial Agent. This agent generates relevant financial context, enabling efficient real-time data retrieval to answer user queries. By reducing latency and eliminating the need for users to manually provide accurate data, our approach significantly enhances the capability of LLMs to handle dynamic financial tasks. Our proposed approach makes real-time financial decisions, algorithmic trading and other related tasks streamlined, and is valuable in contexts with high-velocity data flows.
Authors: Jiayimei Wang, Tao Ni, Wei-Bin Lee, Qingchuan Zhao
Abstract: The increasing complexity of software systems has driven significant advancements in program analysis, as traditional methods unable to meet the demands of modern software development. To address these limitations, deep learning techniques, particularly Large Language Models (LLMs), have gained attention due to their context-aware capabilities in code comprehension. Recognizing the potential of LLMs, researchers have extensively explored their application in program analysis since their introduction. Despite existing surveys on LLM applications in cybersecurity, comprehensive reviews specifically addressing their role in program analysis remain scarce. In this survey, we systematically review the application of LLMs in program analysis, categorizing the existing work into static analysis, dynamic analysis, and hybrid approaches. Moreover, by examining and synthesizing recent studies, we identify future directions and challenges in the field. This survey aims to demonstrate the potential of LLMs in advancing program analysis practices and offer actionable insights for security researchers seeking to enhance detection frameworks or develop domain-specific models.
Authors: Gabriel Patron, Zhiwei Xu, Ishan Kapnadak, Felipe Maia Polo
Abstract: Modern recommender systems follow the guiding principle of serving the right user, the right item at the right time. One of their main limitations is that they are typically limited to items already in the catalog. We propose REcommendations BEyond CAtalogs, REBECA, a new class of probabilistic diffusion-based recommender systems that synthesize new items tailored to individual tastes rather than retrieve items from the catalog. REBECA combines efficient training in embedding space with a novel diffusion prior that only requires users' past ratings of items. We evaluate REBECA on real-world data and propose novel personalization metrics for generative recommender systems. Extensive experiments demonstrate that REBECA produces high-quality, personalized recommendations, generating images that align with users' unique preferences.
Authors: Ilqar Ramazanli, Hamid Eghbalzadeh, Xiaoyi Liu, Yang Wang, Jiaxiang Fu, Kaushik Rangadurai, Sem Park, Bo Long, Xue Feng
Abstract: Perturbation-based regularization techniques address many challenges in industrial-scale large models, particularly with sparse labels, and emphasize consistency and invariance for perturbation in model predictions. One of the popular regularization techniques has been various forms of self-consistency, which involve making small modifications to input data while preserving contextual information and enforcing similar predictions through auxiliary loss functions. In this work, we explore the first successful application of perturbation-based regularization algorithms in large-scale ads ranking models, and further propose a novel regularization algorithm, namely, Loss-Balanced Small Perturbation Regularization (LSPR) that can be used in potentially any deep learning model. We have successfully demonstrate that both Self-Consistency Regularization approaches (SCR) and LSPR are scalable and can improve ads delivery systems. By conducting industrial-scale experiments, and numerical analysis, we additionally show that our proposed LSPR, performs consistently better compared to SCR, across various groups and signal availability setups. Finally, we report a successful application of the proposed LSPR in a billion-scale industrial ranking system, which to the best of our knowledge, is the first of its kind, and it is specially designed to address the various scalability challenges (e.g, various surfaces, geological locations, clients and so on) as we will mention in this paper.
Authors: Shaola Ren, Li Ke, Longtao Huang, Dehong Gao, Hui Xue
Abstract: Automatically extracting effective queries is challenging in information retrieval, especially in toxic content exploration, as such content is likely to be disguised. With the recent achievements in generative Large Language Model (LLM), we are able to leverage the capabilities of LLMs to extract effective queries for similar content exploration directly. This study proposes QExplorer, an approach of large language model based Query Extraction for toxic content Exploration. The QExplorer approach involves a 2-stage training process: instruction Supervised FineTuning (SFT) and preference alignment using Direct Preference Optimization (DPO), as well as the datasets construction with feedback of search system. To verify the effectiveness of QExplorer, a series of offline and online experiments are conducted on our real-world system. The offline empirical results demonstrate that the performance of our automatic query extraction outperforms that of several LLMs and humans. The online deployment shows a significant increase in the detection of toxic items.
Authors: Hang Zhou, Yucheng Wang, Huijing Zhan
Abstract: Multi-modal recommendation systems aim to enhance performance by integrating an item's content features across various modalities with user behavior data. Effective utilization of features from different modalities requires addressing two challenges: preserving semantic commonality across modalities (modality-shared) and capturing unique characteristics for each modality (modality-specific). Most existing approaches focus on aligning feature spaces across modalities, which helps represent modality-shared features. However, modality-specific distinctions are often neglected, especially when there are significant semantic variations between modalities. To address this, we propose a Modality Distinctiveness Enhancement (MDE) framework that prioritizes extracting modality-specific information to improve recommendation accuracy while maintaining shared features. MDE enhances differences across modalities through a novel multi-modal fusion module and introduces a node-level trade-off mechanism to balance cross-modal alignment and differentiation. Extensive experiments on three public datasets show that our approach significantly outperforms other state-of-the-art methods, demonstrating the effectiveness of jointly considering modality-shared and modality-specific features.
Authors: Xinyuan Wang, Yanchi Liu, Wei Cheng, Xujiang Zhao, Zhengzhang Chen, Wenchao Yu, Yanjie Fu, Haifeng Chen
Abstract: Large Language Models (LLMs) exhibit potential artificial generic intelligence recently, however, their usage is costly with high response latency. Given mixed LLMs with their own strengths and weaknesses, LLM routing aims to identify the most suitable model for each query in the stream to maximize response quality and minimize cost and latency. However, the challenges involve: (1) dynamic trade-offs among quality, cost, and latency; (2) enabling continual learning in deployed systems; and (3) navigating a varying (e.g., new LLM addition or old LLM removal) set of LLM candidates over time. To bridge these gaps, we develop MixLLM, a dynamic contextual-bandit-based routing system for query-LLM assignment. Specifically, we first leverage query tags to enhance query embeddings for the routing task. Next, we design lightweight prediction models to estimate the response qualities and costs of queries over LLMs. We then devise a meta-decision maker to choose the query-LLM assignments to best tradeoff response quality, cost, and latency. Finally, the system benefits from continual training, allowing it to adapt to evolving queries and user feedback over time. Our extensive experiments show that MixLLM achieves the best trade-offs in response quality, cost, and latency (97.25% of GPT-4's quality at 24.18% of the cost under the time constraint).
Authors: Gur Keinan, Omer Ben-Porat
Abstract: While recommender systems (RSs) traditionally rely on extensive individual user data, regulatory and technological shifts necessitate reliance on aggregated user information. This shift significantly impacts the recommendation process, requiring RSs to engage in intensive exploration to identify user preferences. However, this approach risks user churn due to potentially unsatisfactory recommendations. In this paper, we propose a model that addresses the dual challenges of leveraging aggregated user information and mitigating churn risk. Our model assumes that the RS operates with a probabilistic prior over user types and aggregated satisfaction levels for various content types. We demonstrate that optimal policies naturally transition from exploration to exploitation in finite time, develop a branch-and-bound algorithm for computing these policies, and empirically validate its effectiveness.
Authors: Krishna Chaitanya Sunkara (Independent Researcher), Krishnaiah Narukulla (Independent Researcher)
Abstract: The conventional resource search in cloud infrastructure relies on keyword-based searches or GUIDs, which demand exact matches and significant user effort to locate resources. These conventional search approaches often fail to interpret the intent behind natural language queries, making resource discovery inefficient and inaccessible to users. Though there exists some form of NLP based search engines, they are limited and focused more on analyzing the NLP query itself and extracting identifiers to find the resources. But they fail to search resources based on their behavior or operations or their capabilities or relationships or features or business relevance or the dynamic changing state or the knowledge these resources have. The search criteria has been changing with the inundation of AI based services which involved discovering not just the requested resources and identifiers but seeking insights. The real intent of a search has never been to just to list the resources but with some actual context such as to understand causes of some behavior in the system, compliance checks, capacity estimations, network constraints, or troubleshooting or business insights. This paper proposes an advanced Natural Language Processing (NLP) enhanced by ontology-based semantics to enable intuitive, human-readable queries which allows users to actually discover the intent-of-search itself. By constructing an ontology of cloud resources, their interactions, and behaviors, the proposed framework enables dynamic intent extraction and relevance ranking using Latent Semantic Indexing (LSI) and AI models. It introduces an automated pipeline which integrates ontology extraction by AI powered data crawlers, building a semantic knowledge base for context aware resource discovery.
Authors: Aditi Mavalankar, Hassan Mansoor, Zita Marinho, Masha Samsikova, Tom Schaul
Abstract: Scaling up inference-time compute has proven to be a valuable strategy in improving the performance of Large Language Models (LLMs) without fine-tuning. An important task that can benefit from additional inference-time compute is self-repair; given an initial flawed response, or guess, the LLM corrects its own mistake and produces an improved response, or fix. We leverage the in-context learning ability of LLMs to perform self-repair in the coding domain. The key contribution of our paper is an approach that synthesises and selects an ordered set of golden example pairs, or AuPairs, of these initial guesses and subsequent fixes for the corresponding problems. Each such AuPair is provided as a single in-context example at inference time to generate a repaired solution. For an inference-time compute budget of $N$ LLM calls per problem, $N$ AuPairs are used to generate $N$ repaired solutions, out of which the highest-scoring solution is selected as the final answer. The underlying intuition is that if the LLM is given a different example of fixing an incorrect guess each time, it can subsequently generate a diverse set of repaired solutions. Our algorithm selects these AuPairs in a manner that maximises complementarity and usefulness. We demonstrate the results of our algorithm on 5 LLMs across 7 competitive programming datasets for the code repair task. Our algorithm yields a significant boost in performance compared to best-of-$N$ and self-repair, and also exhibits strong generalisation across datasets and models. Moreover, our approach shows significantly stronger scaling with inference-time compute budget compared to baselines.
Authors: Tong Ye, Weigang Huang, Xuhong Zhang, Tengfei Ma, Peiyu Liu, Jianwei Yin, Wenhai Wang
Abstract: Large Language Models (LLMs), particularly Code LLMs, have demonstrated impressive performance in code generation. Current research primarily focuses on the correctness of generated code, while efficiency remains less explored. Recent works have focused on modifying the initial version of the code to improve its efficiency. However, such refinements are limited by the algorithmic design and overall logic of the initial code, resulting in only incremental improvements. In contrast, when human developers write high-quality code, they typically begin by designing several potential solutions at the logical level, evaluating various algorithms and their complexities, and then proceeding to implement and optimize the solution. In this study, we introduce \tool: \uline{L}arge \uline{L}anguage \uline{M}odel for Code \uline{Effi}ciency, a novel framework that enables LLMs to generate code that balances both efficiency and correctness. Specifically, \tool divides the efficiency optimization process into two domains: algorithmic exploration in the logic domain and implementation optimization in the code domain. The correctness of the code is then guaranteed through a synthetic test case refinement process. This approach, which prioritizes efficiency before ensuring correctness, offers a new paradigm for efficient code generation. Experiments demonstrate that \tool consistently improves both efficiency and correctness, achieving new state-of-the-art performance in code efficiency benchmarks across various LLM backbones.
Authors: Lukas Schulze Balhorn, Niels Seijsener, Kevin Dao, Minji Kim, Dominik P. Goldstein, Ge H. M. Driessen, Artur M. Schweidtmann
Abstract: A piping and instrumentation diagram (P&ID) is a central reference document in chemical process engineering. Currently, chemical engineers manually review P&IDs through visual inspection to find and rectify errors. However, engineering projects can involve hundreds to thousands of P&ID pages, creating a significant revision workload. This study proposes a rule-based method to support engineers with error detection and correction in P&IDs. The method is based on a graph representation of P&IDs, enabling automated error detection and correction, i.e., autocorrection, through rule graphs. We use our pyDEXPI Python package to generate P&ID graphs from DEXPI-standard P&IDs. In this study, we developed 33 rules based on chemical engineering knowledge and heuristics, with five selected rules demonstrated as examples. A case study on an illustrative P&ID validates the reliability and effectiveness of the rule-based autocorrection method in revising P&IDs.
Authors: Xuemeng Song, Haoqiang Lin, Haokun Wen, Bohan Hou, Mingzhu Xu, Liqiang Nie
Abstract: Composed Image Retrieval (CIR) is an emerging yet challenging task that allows users to search for target images using a multimodal query, comprising a reference image and a modification text specifying the user's desired changes to the reference image. Given its significant academic and practical value, CIR has become a rapidly growing area of interest in the computer vision and machine learning communities, particularly with the advances in deep learning. To the best of our knowledge, there is currently no comprehensive review of CIR to provide a timely overview of this field. Therefore, we synthesize insights from over 120 publications in top conferences and journals, including ACM TOIS, SIGIR, and CVPR In particular, we systematically categorize existing supervised CIR and zero-shot CIR models using a fine-grained taxonomy. For a comprehensive review, we also briefly discuss approaches for tasks closely related to CIR, such as attribute-based CIR and dialog-based CIR. Additionally, we summarize benchmark datasets for evaluation and analyze existing supervised and zero-shot CIR methods by comparing experimental results across multiple datasets. Furthermore, we present promising future directions in this field, offering practical insights for researchers interested in further exploration.
Authors: Samuel Miller, Daking Rai, Ziyu Yao
Abstract: Recently, language models (LMs) have shown impressive proficiency in code generation tasks, especially when fine-tuned on code-specific datasets, commonly known as Code LMs. However, our understanding of the internal decision-making processes of Code LMs, such as how they use their (syntactic or semantic) knowledge, remains limited, which could lead to unintended harm as they are increasingly used in real life. This motivates us to conduct one of the first Mechanistic Interpretability works to understand how Code LMs perform a syntactic completion task, specifically the closing parenthesis task, on the CodeLlama-7b model (Roziere et al. 2023). Our findings reveal that the model requires middle-later layers until it can confidently predict the correct label for the closing parenthesis task. Additionally, we identify that while both multi-head attention (MHA) and feed-forward (FF) sub-layers play essential roles, MHA is particularly crucial. Furthermore, we also discover attention heads that keep track of the number of already closed parentheses precisely but may or may not promote a correct number of closing parentheses that are still missing, leading to a positive or negative impact on the model's performance.
Authors: Sudev Kumar Padhi, Archana Tiwari, Sk. Subidh Ali
Abstract: Advancements in digital technologies make it easy to modify the content of digital images. Hence, ensuring digital images integrity and authenticity is necessary to protect them against various attacks that manipulate them. We present a Deep Learning (DL) based dual invisible watermarking technique for performing source authentication, content authentication, and protecting digital content copyright of images sent over the internet. Beyond securing images, the proposed technique demonstrates robustness to content-preserving image manipulations. It is also impossible to imitate or overwrite watermarks because the cryptographic hash of the image and the dominant features of the image in the form of perceptual hash are used as watermarks. We highlighted the need for source authentication to safeguard image integrity and authenticity, along with identifying similar content for copyright protection. After exhaustive testing, we obtained a high peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM), which implies there is a minute change in the original image after embedding our watermarks. Our trained model achieves high watermark extraction accuracy and to the best of our knowledge, this is the first deep learning-based dual watermarking technique proposed in the literature.
Authors: Aman Goel, Xian Carrie Wu, Zhe Wang, Dmitriy Bespalov, Yanjun Qi
Abstract: Jailbreaking large-language models (LLMs) involves testing their robustness against adversarial prompts and evaluating their ability to withstand prompt attacks that could elicit unauthorized or malicious responses. In this paper, we present TurboFuzzLLM, a mutation-based fuzzing technique for efficiently finding a collection of effective jailbreaking templates that, when combined with harmful questions, can lead a target LLM to produce harmful responses through black-box access via user prompts. We describe the limitations of directly applying existing template-based attacking techniques in practice, and present functional and efficiency-focused upgrades we added to mutation-based fuzzing to generate effective jailbreaking templates automatically. TurboFuzzLLM achieves $\geq$ 95\% attack success rates (ASR) on public datasets for leading LLMs (including GPT-4o \& GPT-4 Turbo), shows impressive generalizability to unseen harmful questions, and helps in improving model defenses to prompt attacks.
Authors: Ranjan Sapkota, Shaina Raza, Manoj Karkee
Abstract: Despite increasing discussions on open-source Artificial Intelligence (AI), existing research lacks a discussion on the transparency and accessibility of state-of-the-art (SoTA) Large Language Models (LLMs). The Open Source Initiative (OSI) has recently released its first formal definition of open-source software. This definition, when combined with standard dictionary definitions and the sparse published literature, provide an initial framework to support broader accessibility to AI models such as LLMs, but more work is essential to capture the unique dynamics of openness in AI. In addition, concerns about open-washing, where models claim openness but lack full transparency, has been raised, which limits the reproducibility, bias mitigation, and domain adaptation of these models. In this context, our study critically analyzes SoTA LLMs from the last five years, including ChatGPT, DeepSeek, LLaMA, and others, to assess their adherence to transparency standards and the implications of partial openness. Specifically, we examine transparency and accessibility from two perspectives: open-source vs. open-weight models. Our findings reveal that while some models are labeled as open-source, this does not necessarily mean they are fully open-sourced. Even in the best cases, open-source models often do not report model training data, and code as well as key metrics, such as weight accessibility, and carbon emissions. To the best of our knowledge, this is the first study that systematically examines the transparency and accessibility of over 100 different SoTA LLMs through the dual lens of open-source and open-weight models. The findings open avenues for further research and call for responsible and sustainable AI practices to ensure greater transparency, accountability, and ethical deployment of these models.(DeepSeek transparency, ChatGPT accessibility, open source, DeepSeek open source)
Authors: Solha Kang, Wesley De Neve, Francois Rameau, Utku Ozbulak
Abstract: The past decade has witnessed a substantial increase in the number of startups and companies offering AI-based solutions for clinical decision support in medical institutions. However, the critical nature of medical decision-making raises several concerns about relying on external software. Key issues include potential variations in image modalities and the medical devices used to obtain these images, potential legal issues, and adversarial attacks. Fortunately, the open-source nature of machine learning research has made foundation models publicly available and straightforward to use for medical applications. This accessibility allows medical institutions to train their own AI-based models, thereby mitigating the aforementioned concerns. Given this context, an important question arises: how much data do medical institutions need to train effective AI models? In this study, we explore this question in relation to breast cancer detection, a particularly contested area due to the prevalence of this disease, which affects approximately 1 in every 8 women. Through large-scale experiments on various patient sizes in the training set, we show that medical institutions do not need a decade's worth of MRI images to train an AI model that performs competitively with the state-of-the-art, provided the model leverages foundation models. Furthermore, we observe that for patient counts greater than 50, the number of patients in the training set has a negligible impact on the performance of models and that simple ensembles further improve the results without additional complexity.
Authors: Yukun Chen, Shuo Shao, Enhao Huang, Yiming Li, Pin-Yu Chen, Zhan Qin, Kui Ren
Abstract: Backdoor attacks on deep neural networks (DNNs) have emerged as a significant security threat, allowing adversaries to implant hidden malicious behaviors during the model training phase. Pre-processing-based defense, which is one of the most important defense paradigms, typically focuses on input transformations or backdoor trigger inversion (BTI) to deactivate or eliminate embedded backdoor triggers during the inference process. However, these methods suffer from inherent limitations: transformation-based defenses often fail to balance model utility and defense performance, while BTI-based defenses struggle to accurately reconstruct trigger patterns without prior knowledge. In this paper, we propose REFINE, an inversion-free backdoor defense method based on model reprogramming. REFINE consists of two key components: \textbf{(1)} an input transformation module that disrupts both benign and backdoor patterns, generating new benign features; and \textbf{(2)} an output remapping module that redefines the model's output domain to guide the input transformations effectively. By further integrating supervised contrastive loss, REFINE enhances the defense capabilities while maintaining model utility. Extensive experiments on various benchmark datasets demonstrate the effectiveness of our REFINE and its resistance to potential adaptive attacks.
Authors: Ivoline Ngong, Swanand Kadhe, Hao Wang, Keerthiram Murugesan, Justin D. Weisz, Amit Dhurandhar, Karthikeyan Natesan Ramamurthy
Abstract: Conversational agents are increasingly woven into individuals' personal lives, yet users often underestimate the privacy risks involved. The moment users share information with these agents (e.g., LLMs), their private information becomes vulnerable to exposure. In this paper, we characterize the notion of contextual privacy for user interactions with LLMs. It aims to minimize privacy risks by ensuring that users (sender) disclose only information that is both relevant and necessary for achieving their intended goals when interacting with LLMs (untrusted receivers). Through a formative design user study, we observe how even "privacy-conscious" users inadvertently reveal sensitive information through indirect disclosures. Based on insights from this study, we propose a locally-deployable framework that operates between users and LLMs, and identifies and reformulates out-of-context information in user prompts. Our evaluation using examples from ShareGPT shows that lightweight models can effectively implement this framework, achieving strong gains in contextual privacy while preserving the user's intended interaction goals through different approaches to classify information relevant to the intended goals.
Authors: Xuxu Liu, Siyuan Liang, Mengya Han, Yong Luo, Aishan Liu, Xiantao Cai, Zheng He, Dacheng Tao
Abstract: Generative large language models are crucial in natural language processing, but they are vulnerable to backdoor attacks, where subtle triggers compromise their behavior. Although backdoor attacks against LLMs are constantly emerging, existing benchmarks remain limited in terms of sufficient coverage of attack, metric system integrity, backdoor attack alignment. And existing pre-trained backdoor attacks are idealized in practice due to resource access constraints. Therefore we establish $\textit{ELBA-Bench}$, a comprehensive and unified framework that allows attackers to inject backdoor through parameter efficient fine-tuning ($\textit{e.g.,}$ LoRA) or without fine-tuning techniques ($\textit{e.g.,}$ In-context-learning). $\textit{ELBA-Bench}$ provides over 1300 experiments encompassing the implementations of 12 attack methods, 18 datasets, and 12 LLMs. Extensive experiments provide new invaluable findings into the strengths and limitations of various attack strategies. For instance, PEFT attack consistently outperform without fine-tuning approaches in classification tasks while showing strong cross-dataset generalization with optimized triggers boosting robustness; Task-relevant backdoor optimization techniques or attack prompts along with clean and adversarial demonstrations can enhance backdoor attack success while preserving model performance on clean samples. Additionally, we introduce a universal toolbox designed for standardized backdoor attack research, with the goal of propelling further progress in this vital area.
Authors: Jianjian Li, Junquan Fan, Feng Tang, Gang Huang, Shitao Zhu, Songlin Liu, Nian Xie, Wulong Liu, Yong Liao
Abstract: The rapid success of Vision Large Language Models (VLLMs) often depends on the high-resolution images with abundant visual tokens, which hinders training and deployment efficiency. Current training-free visual token compression methods exhibit serious performance degradation in tasks involving high-resolution, text-oriented image understanding and reasoning. In this paper, we propose an efficient visual token compression framework for text-oriented VLLMs in high-resolution scenarios. In particular, we employ a light-weight self-distillation pre-training stage to compress the visual tokens, requiring a limited numbers of image-text pairs and minimal learnable parameters. Afterwards, to mitigate potential performance degradation of token-compressed models, we construct a high-quality post-train stage. To validate the effectiveness of our method, we apply it to an advanced VLLMs, InternVL2. Experimental results show that our approach significantly reduces computational overhead while outperforming the baselines across a range of text-oriented benchmarks. We will release the models and code soon.
Authors: Rabimba Karanjai, Sam Blackshear, Lei Xu, Weidong Shi
Abstract: The rapid growth of the blockchain ecosystem and the increasing value locked in smart contracts necessitate robust security measures. While languages like Solidity and Move aim to improve smart contract security, vulnerabilities persist. This paper presents Smartify, a novel multi-agent framework leveraging Large Language Models (LLMs) to automatically detect and repair vulnerabilities in Solidity and Move smart contracts. Unlike traditional methods that rely solely on vast pre-training datasets, Smartify employs a team of specialized agents working on different specially fine-tuned LLMs to analyze code based on underlying programming concepts and language-specific security principles. We evaluated Smartify on a dataset for Solidity and a curated dataset for Move, demonstrating its effectiveness in fixing a wide range of vulnerabilities. Our results show that Smartify (Gemma2+codegemma) achieves state-of-the-art performance, surpassing existing LLMs and enhancing general-purpose models' capabilities, such as Llama 3.1. Notably, Smartify can incorporate language-specific knowledge, such as the nuances of Move, without requiring massive language-specific pre-training datasets. This work offers a detailed analysis of various LLMs' performance on smart contract repair, highlighting the strengths of our multi-agent approach and providing a blueprint for developing more secure and reliable decentralized applications in the growing blockchain landscape. We also provide a detailed recipe for extending this to other similar use cases.
Authors: Jianwei Wang, Junyao Yang, Haoran Li, Huiping Zhuang, Cen Chen, Ziqian Zeng
Abstract: The success of large language models (LLMs) has attracted many individuals to fine-tune them for domain-specific tasks by uploading their data. However, in sensitive areas like healthcare and finance, privacy concerns often arise. One promising solution is to sample synthetic data with Differential Privacy (DP) guarantees to replace private data. However, these synthetic data contain significant flawed data, which are considered as noise. Existing solutions typically rely on naive filtering by comparing ROUGE-L scores or embedding similarities, which are ineffective in addressing the noise. To address this issue, we propose RewardDS, a novel privacy-preserving framework that fine-tunes a reward proxy model and uses reward signals to guide the synthetic data generation. Our RewardDS introduces two key modules, Reward Guided Filtering and Self-Optimizing Refinement, to both filter and refine the synthetic data, effectively mitigating the noise. Extensive experiments across medical, financial, and code generation domains demonstrate the effectiveness of our method.
Authors: Peng Yifeng, Wu Zhizheng, Chen Chen
Abstract: Modern large language models (LLMs) exhibit critical vulnerabilities to poison pill attacks: localized data poisoning that alters specific factual knowledge while preserving overall model utility. We systematically demonstrate these attacks exploit inherent architectural properties of LLMs, achieving 54.6% increased retrieval inaccuracy on long-tail knowledge versus dominant topics and up to 25.5% increase retrieval inaccuracy on compressed models versus original architectures. Through controlled mutations (e.g., temporal/spatial/entity alterations) and, our method induces localized memorization deterioration with negligible impact on models' performance on regular standard benchmarks (e.g., <2% performance drop on MMLU/GPQA), leading to potential detection evasion. Our findings suggest: (1) Disproportionate vulnerability in long-tail knowledge may result from reduced parameter redundancy; (2) Model compression may increase attack surfaces, with pruned/distilled models requiring 30% fewer poison samples for equivalent damage; (3) Associative memory enables both spread of collateral damage to related concepts and amplification of damage from simultaneous attack, particularly for dominant topics. These findings raise concerns over current scaling paradigms since attack costs are lowering while defense complexity is rising. Our work establishes poison pills as both a security threat and diagnostic tool, revealing critical security-efficiency trade-offs in language model compression that challenges prevailing safety assumptions.
Authors: Linshan Wu, Jiaxin Zhuang, Yanning Zhou, Sunan He, Jiabo Ma, Luyang Luo, Xi Wang, Xuefeng Ni, Xiaoling Zhong, Mingxiang Wu, Yinghua Zhao, Xiaohui Duan, Varut Vardhanabhuti, Pranav Rajpurkar, Hao Chen
Abstract: Tumor is a leading cause of death worldwide, with an estimated 10 million deaths attributed to tumor-related diseases every year. AI-driven tumor recognition unlocks new possibilities for more precise and intelligent tumor screening and diagnosis. However, the progress is heavily hampered by the scarcity of annotated datasets, which demands extensive annotation efforts by radiologists. To tackle this challenge, we introduce FreeTumor, an innovative Generative AI (GAI) framework to enable large-scale tumor synthesis for mitigating data scarcity. Specifically, FreeTumor effectively leverages a combination of limited labeled data and large-scale unlabeled data for tumor synthesis training. Unleashing the power of large-scale data, FreeTumor is capable of synthesizing a large number of realistic tumors on images for augmenting training datasets. To this end, we create the largest training dataset for tumor synthesis and recognition by curating 161,310 publicly available Computed Tomography (CT) volumes from 33 sources, with only 2.3% containing annotated tumors. To validate the fidelity of synthetic tumors, we engaged 13 board-certified radiologists in a Visual Turing Test to discern between synthetic and real tumors. Rigorous clinician evaluation validates the high quality of our synthetic tumors, as they achieved only 51.1% sensitivity and 60.8% accuracy in distinguishing our synthetic tumors from real ones. Through high-quality tumor synthesis, FreeTumor scales up the recognition training datasets by over 40 times, showcasing a notable superiority over state-of-the-art AI methods including various synthesis methods and foundation models. These findings indicate promising prospects of FreeTumor in clinical applications, potentially advancing tumor treatments and improving the survival rates of patients.
Authors: Mingli Zhu, Shaokui Wei, Hongyuan Zha, Baoyuan Wu
Abstract: Recent studies have highlighted the vulnerability of deep neural networks to backdoor attacks, where models are manipulated to rely on embedded triggers within poisoned samples, despite the presence of both benign and trigger information. While several defense methods have been proposed, they often struggle to balance backdoor mitigation with maintaining benign performance.In this work, inspired by the concept of optical polarizer-which allows light waves of specific polarizations to pass while filtering others-we propose a lightweight backdoor defense approach, NPD. This method integrates a neural polarizer (NP) as an intermediate layer within the compromised model, implemented as a lightweight linear transformation optimized via bi-level optimization. The learnable NP filters trigger information from poisoned samples while preserving benign content. Despite its effectiveness, we identify through empirical studies that NPD's performance degrades when the target labels (required for purification) are inaccurately estimated. To address this limitation while harnessing the potential of targeted adversarial mitigation, we propose class-conditional neural polarizer-based defense (CNPD). The key innovation is a fusion module that integrates the backdoored model's predicted label with the features to be purified. This architecture inherently mimics targeted adversarial defense mechanisms without requiring label estimation used in NPD. We propose three implementations of CNPD: the first is r-CNPD, which trains a replicated NP layer for each class and, during inference, selects the appropriate NP layer for defense based on the predicted class from the backdoored model. To efficiently handle a large number of classes, two variants are designed: e-CNPD, which embeds class information as additional features, and a-CNPD, which directs network attention using class information.
Authors: Yao Su, Keqi Han, Mingjie Zeng, Lichao Sun, Liang Zhan, Carl Yang, Lifang He, Xiangnan Kong
Abstract: Brain imaging analysis is fundamental in neuroscience, providing valuable insights into brain structure and function. Traditional workflows follow a sequential pipeline-brain extraction, registration, segmentation, parcellation, network generation, and classification-treating each step as an independent task. These methods rely heavily on task-specific training data and expert intervention to correct intermediate errors, making them particularly burdensome for high-dimensional neuroimaging data, where annotations and quality control are costly and time-consuming. We introduce UniBrain, a unified end-to-end framework that integrates all processing steps into a single optimization process, allowing tasks to interact and refine each other. Unlike traditional approaches that require extensive task-specific annotations, UniBrain operates with minimal supervision, leveraging only low-cost labels (i.e., classification and extraction) and a single labeled atlas. By jointly optimizing extraction, registration, segmentation, parcellation, network generation, and classification, UniBrain enhances both accuracy and computational efficiency while significantly reducing annotation effort. Experimental results demonstrate its superiority over existing methods across multiple tasks, offering a more scalable and reliable solution for neuroimaging analysis. Our code and data can be found at https://github.com/Anonymous7852/UniBrain
Authors: Fangqi Liu, Rishav Sen, Jose Paolo Talusan, Ava Pettet, Aaron Kandel, Yoshinori Suzue, Ayan Mukhopadhyay, Abhishek Dubey
Abstract: Strategic aggregation of electric vehicle batteries as energy reservoirs can optimize power grid demand, benefiting smart and connected communities, especially large office buildings that offer workplace charging. This involves optimizing charging and discharging to reduce peak energy costs and net peak demand, monitored over extended periods (e.g., a month), which involves making sequential decisions under uncertainty and delayed and sparse rewards, a continuous action space, and the complexity of ensuring generalization across diverse conditions. Existing algorithmic approaches, e.g., heuristic-based strategies, fall short in addressing real-time decision-making under dynamic conditions, and traditional reinforcement learning (RL) models struggle with large state-action spaces, multi-agent settings, and the need for long-term reward optimization. To address these challenges, we introduce a novel RL framework that combines the Deep Deterministic Policy Gradient approach (DDPG) with action masking and efficient MILP-driven policy guidance. Our approach balances the exploration of continuous action spaces to meet user charging demands. Using real-world data from a major electric vehicle manufacturer, we show that our approach comprehensively outperforms many well-established baselines and several scalable heuristic approaches, achieving significant cost savings while meeting all charging requirements. Our results show that the proposed approach is one of the first scalable and general approaches to solving the V2B energy management challenge.
Authors: PIN AI Team, Bill Qingyun Sun, Laura Florescu, Boliang Zhang, Regan Peng, Smile Hu, Shouqiao Wang, Ben Wu, Xi Wang, Davide Crapis, Gavin Zhen Guo
Abstract: Personal AI assistants (e.g., Apple Intelligence, Meta AI) offer proactive recommendations that simplify everyday tasks, but their reliance on sensitive user data raises concerns about privacy and trust. To address these challenges, we introduce the Guardian of Data (GOD), a secure, privacy-preserving framework for training and evaluating AI assistants directly on-device. Unlike traditional benchmarks, the GOD model measures how well assistants can anticipate user needs-such as suggesting gifts-while protecting user data and autonomy. Functioning like an AI school, it addresses the cold start problem by simulating user queries and employing a curriculum-based approach to refine the performance of each assistant. Running within a Trusted Execution Environment (TEE), it safeguards user data while applying reinforcement and imitation learning to refine AI recommendations. A token-based incentive system encourages users to share data securely, creating a data flywheel that drives continuous improvement. By integrating privacy, personalization, and trust, the GOD model provides a scalable, responsible path for advancing personal AI assistants. For community collaboration, part of the framework is open-sourced at https://github.com/PIN-AI/God-Model.
Authors: Tomas Nieponice, Veronica Valeros, Sebastian Garcia
Abstract: We introduce ARACNE, a fully autonomous LLM-based pentesting agent tailored for SSH services that can execute commands on real Linux shell systems. Introduces a new agent architecture with multi-LLM model support. Experiments show that ARACNE can reach a 60\% success rate against the autonomous defender ShelLM and a 57.58\% success rate against the Over The Wire Bandit CTF challenges, improving over the state-of-the-art. When winning, the average number of actions taken by the agent to accomplish the goals was less than 5. The results show that the use of multi-LLM is a promising approach to increase accuracy in the actions.
Authors: Hang Wang, Qiaoyi Fang, Junshan Zhang
Abstract: The past few years have witnessed a rapid growth of the deployment of automated vehicles (AVs). Clearly, AVs and human-driven vehicles (HVs) will co-exist for many years, and AVs will have to operate around HVs, pedestrians, cyclists, and more, calling for fundamental breakthroughs in AI designed for mixed traffic to achieve mixed autonomy. Thus motivated, we study heterogeneous decision making by AVs and HVs in a mixed traffic environment, aiming to capture the interactions between human and machine decision-making and develop an AI foundation that enables vehicles to operate safely and efficiently. There are a number of challenges to achieve mixed autonomy, including 1) humans drivers make driving decisions with bounded rationality, and it remains open to develop accurate models for HVs' decision making; and 2) uncertainty-aware planning plays a critical role for AVs to take safety maneuvers in response to the human behavior. In this paper, we introduce a formulation of AV-HV interaction, where the HV makes decisions with bounded rationality and the AV employs uncertainty-aware planning based on the prediction on HV's future actions. We conduct a comprehensive analysis on AV and HV's learning regret to answer the questions: 1) {How does the learning performance depend on HV's bounded rationality and AV's planning}; 2) {How do different decision making strategies impact the overall learning performance}? Our findings reveal some intriguing phenomena, such as Goodhart's Law in AV's learning performance and compounding effects in HV's decision making process. By examining the dynamics of the regrets, we gain insights into the interplay between human and machine decision making.
Authors: Zachary McBride Lazri, Anirudh Nakra, Ivan Brugere, Danial Dervovic, Antigoni Polychroniadou, Furong Huang, Dana Dachman-Soled, Min Wu
Abstract: Fairness constraints applied to machine learning (ML) models in static contexts have been shown to potentially produce adverse outcomes among demographic groups over time. To address this issue, emerging research focuses on creating fair solutions that persist over time. While many approaches treat this as a single-agent decision-making problem, real-world systems often consist of multiple interacting entities that influence outcomes. Explicitly modeling these entities as agents enables more flexible analysis of their interventions and the effects they have on a system's underlying dynamics. A significant challenge in conducting research on multi-agent systems is the lack of realistic environments that leverage the limited real-world data available for analysis. To address this gap, we introduce the concept of a Multi-Agent Fair Environment (MAFE) and present and analyze three MAFEs that model distinct social systems. Experimental results demonstrate the utility of our MAFEs as testbeds for developing multi-agent fair algorithms.
Authors: Zhizhi Peng, Taotao Wang, Chonghe Zhao, Guofu Liao, Zibin Lin, Yifeng Liu, Bin Cao, Long Shi, Qing Yang, Shengli Zhang
Abstract: As machine learning technologies advance rapidly across various domains, concerns over data privacy and model security have grown significantly. These challenges are particularly pronounced when models are trained and deployed on cloud platforms or third-party servers due to the computational resource limitations of users' end devices. In response, zero-knowledge proof (ZKP) technology has emerged as a promising solution, enabling effective validation of model performance and authenticity in both training and inference processes without disclosing sensitive data. Thus, ZKP ensures the verifiability and security of machine learning models, making it a valuable tool for privacy-preserving AI. Although some research has explored the verifiable machine learning solutions that exploit ZKP, a comprehensive survey and summary of these efforts remain absent. This survey paper aims to bridge this gap by reviewing and analyzing all the existing Zero-Knowledge Machine Learning (ZKML) research from June 2017 to December 2024. We begin by introducing the concept of ZKML and outlining its ZKP algorithmic setups under three key categories: verifiable training, verifiable inference, and verifiable testing. Next, we provide a comprehensive categorization of existing ZKML research within these categories and analyze the works in detail. Furthermore, we explore the implementation challenges faced in this field and discuss the improvement works to address these obstacles. Additionally, we highlight several commercial applications of ZKML technology. Finally, we propose promising directions for future advancements in this domain.
Authors: Yu Bo, Weian Mao, Yanjun Shao, Weiqiang Bai, Peng Ye, Xinzhu Ma, Junbo Zhao, Hao Chen, Chunhua Shen
Abstract: In recent years, a variety of methods based on Transformer and state space model (SSM) architectures have been proposed, advancing foundational DNA language models. However, there is a lack of comparison between these recent approaches and the classical architecture convolutional networks (CNNs) on foundation model benchmarks. This raises the question: are CNNs truly being surpassed by these recent approaches based on transformer and SSM architectures? In this paper, we develop a simple but well-designed CNN-based method termed ConvNova. ConvNova identifies and proposes three effective designs: 1) dilated convolutions, 2) gated convolutions, and 3) a dual-branch framework for gating mechanisms. Through extensive empirical experiments, we demonstrate that ConvNova significantly outperforms recent methods on more than half of the tasks across several foundation model benchmarks. For example, in histone-related tasks, ConvNova exceeds the second-best method by an average of 5.8%, while generally utilizing fewer parameters and enabling faster computation. In addition, the experiments observed findings that may be related to biological characteristics. This indicates that CNNs are still a strong competitor compared to Transformers and SSMs. We anticipate that this work will spark renewed interest in CNN-based methods for DNA foundation models.
Authors: Zike Yuan, Ming Liu, Hui Wang, Bing Qin
Abstract: Graph-theoretic problems arise in real-world applications like logistics, communication networks, and traffic optimization. These problems are often complex, noisy, and irregular, posing challenges for traditional algorithms. Large language models (LLMs) offer potential solutions but face challenges, including limited accuracy and input length constraints. To address these challenges, we propose MA-GTS (Multi-Agent Graph Theory Solver), a multi-agent framework that decomposes these complex problems through agent collaboration. MA-GTS maps the implicitly expressed text-based graph data into clear, structured graph representations and dynamically selects the most suitable algorithm based on problem constraints and graph structure scale. This approach ensures that the solution process remains efficient and the resulting reasoning path is interpretable. We validate MA-GTS using the G-REAL dataset, a real-world-inspired graph theory dataset we created. Experimental results show that MA-GTS outperforms state-of-the-art approaches in terms of efficiency, accuracy, and scalability, with strong results across multiple benchmarks (G-REAL 94.2%, GraCoRe 96.9%, NLGraph 98.4%).MA-GTS is open-sourced at https://github.com/ZIKEYUAN/MA-GTS.git.
Authors: Hao Shen, Zhouhong Gu, Haokai Hong, Weili Han
Abstract: The widespread adoption of Large Language Models (LLMs) has raised significant privacy concerns regarding the exposure of personally identifiable information (PII) in user prompts. To address this challenge, we propose a query-unrelated PII masking strategy and introduce PII-Bench, the first comprehensive evaluation framework for assessing privacy protection systems. PII-Bench comprises 2,842 test samples across 55 fine-grained PII categories, featuring diverse scenarios from single-subject descriptions to complex multi-party interactions. Each sample is carefully crafted with a user query, context description, and standard answer indicating query-relevant PII. Our empirical evaluation reveals that while current models perform adequately in basic PII detection, they show significant limitations in determining PII query relevance. Even state-of-the-art LLMs struggle with this task, particularly in handling complex multi-subject scenarios, indicating substantial room for improvement in achieving intelligent PII masking.
Authors: Ching-Chun Chang, Isao Echizen
Abstract: Steganography is the art and science of covert writing, with a broad range of applications interwoven within the realm of cybersecurity. As artificial intelligence continues to evolve, its ability to synthesise realistic content emerges as a threat in the hands of cybercriminals who seek to manipulate and misrepresent the truth. Such synthetic content introduces a non-trivial risk of overwriting the subtle changes made for the purpose of steganography. When the signals in both the spatial and temporal domains are vulnerable to unforeseen overwriting, it calls for reflection on what can remain invariant after all. This study proposes a paradigm in steganography for audiovisual media, where messages are concealed beyond both spatial and temporal domains. A chain of multimodal agents is developed to deconstruct audiovisual content into a cover text, embed a message within the linguistic domain, and then reconstruct the audiovisual content through synchronising both aural and visual modalities with the resultant stego text. The message is encoded by biasing the word sampling process of a language generation model and decoded by analysing the probability distribution of word choices. The accuracy of message transmission is evaluated under both zero-bit and multi-bit capacity settings. Fidelity is assessed through both biometric and semantic similarities, capturing the identities of the recorded face and voice, as well as the core ideas conveyed through the media. Secrecy is examined through statistical comparisons between cover and stego texts. Robustness is tested across various scenarios, including audiovisual compression, face-swapping, voice-cloning and their combinations.
Authors: Milan Vojnovic, Se-Young Yun
Abstract: In this note, we examine the aggregation of preferences achieved by the Group Policy Optimisation (GRPO) algorithm, a reinforcement learning method used to train advanced artificial intelligence models such as DeepSeek-R1-Zero and DeepSeekMath. The GRPO algorithm trains a policy using a reward preference model, which is computed by sampling a set of outputs for a given context, observing the corresponding rewards, and applying shift-and-scale normalisation to these reward values. Additionally, it incorporates a penalty function to discourage deviations from a reference policy. We present a framework that enables us to characterise the stationary policies of the GRPO algorithm. This analysis reveals that the aggregation of preferences differs fundamentally from standard logarithmic pooling, which is implemented by other approaches such as RLHF. The precise form of preference aggregation arises from the way the reward preference model is defined and from the penalty function, which we show to essentially correspond to the reverse Kullback-Leibler (KL) divergence between the aggregation policy and the reference policy. Interestingly, we demonstrate that for groups of size two, the reward preference model corresponds to pairwise comparison preferences, similar to those in other alignment methods based on pairwise comparison feedback. We provide explicit characterisations of the aggregate preference for binary questions, for groups of size two, and in the limit of large group size. This provides insights into the dependence of the aggregate preference on parameters such as the regularisation constant and the confidence margin of question answers. Finally, we discuss the aggregation of preferences obtained by modifying the GRPO algorithm to use direct KL divergence as the penalty or to use rewards without scale normalisation.
Authors: Zohar Ringel, Noa Rubin, Edo Mor, Moritz Helias, Inbar Seroussi
Abstract: Deep learning algorithms have made incredible strides in the past decade yet due to the complexity of these algorithms, the science of deep learning remains in its early stages. Being an experimentally driven field, it is natural to seek a theory of deep learning within the physics paradigm. As deep learning is largely about learning functions and distributions over functions, statistical field theory, a rich and versatile toolbox for tackling complex distributions over functions (fields) is an obvious choice of formalism. Research efforts carried out in the past few years have demonstrated the ability of field theory to provide useful insights on generalization, implicit bias, and feature learning effects. Here we provide a pedagogical review of this emerging line of research.
Authors: Erick da Silva Farias, Eduardo Palhares Junior
Abstract: The automatic detection of human conflicts through videos is a crucial area in computer vision, with significant applications in monitoring and public safety policies. However, the scarcity of public datasets and the complexity of human interactions make this task challenging. This study investigates the integration of advanced deep learning techniques, including Attention Mechanism, Convolutional Neural Networks (CNNs), and Bidirectional Long ShortTerm Memory (BiLSTM), to improve the detection of violent behaviors in videos. The research explores how the use of the attention mechanism can help focus on the most relevant parts of the video, enhancing the accuracy and robustness of the model. The experiments indicate that the combination of CNNs with BiLSTM and the attention mechanism provides a promising solution for conflict monitoring, offering insights into the effectiveness of different strategies. This work opens new possibilities for the development of automated surveillance systems that can operate more efficiently in real-time detection of violent events.
Authors: Radu Marinescu, Debarun Bhattacharjya, Junkyu Lee, Tigran Tchrakian, Javier Carnerero Cano, Yufang Hou, Elizabeth Daly, Alessandra Pascale
Abstract: Large language models (LLMs) have demonstrated vast capabilities on generative tasks in recent years, yet they struggle with guaranteeing the factual correctness of the generated content. This makes these models unreliable in realistic situations where factually accurate responses are expected. In this paper, we propose FactReasoner, a new factuality assessor that relies on probabilistic reasoning to assess the factuality of a long-form generated response. Specifically, FactReasoner decomposes the response into atomic units, retrieves relevant contexts for them from an external knowledge source, and constructs a joint probability distribution over the atoms and contexts using probabilistic encodings of the logical relationships (entailment, contradiction) between the textual utterances corresponding to the atoms and contexts. FactReasoner then computes the posterior probability of whether atomic units in the response are supported by the retrieved contexts. Our experiments on labeled and unlabeled benchmark datasets demonstrate clearly that FactReasoner improves considerably over state-of-the-art prompt-based approaches in terms of both factual precision and recall.
Authors: Amol Khanna, Fred Lu, Edward Raff
Abstract: Linear $L_1$-regularized models have remained one of the simplest and most effective tools in data science. Over the past decade, screening rules have risen in popularity as a way to eliminate features when producing the sparse regression weights of $L_1$ models. However, despite the increasing need of privacy-preserving models for data analysis, to the best of our knowledge, no differentially private screening rule exists. In this paper, we develop the first private screening rule for linear regression. We initially find that this screening rule is too strong: it screens too many coefficients as a result of the private screening step. However, a weakened implementation of private screening reduces overscreening and improves performance.
Authors: Zhewei Kang, Xuandong Zhao, Dawn Song
Abstract: Best-of-N selection is a key technique for improving the reasoning performance of Large Language Models (LLMs) through increased test-time computation. Current state-of-the-art methods often employ computationally intensive reward models for response evaluation and selection. Reward-free alternatives, like self-consistency and universal self-consistency, are limited in their ability to handle open-ended generation tasks or scale effectively. To address these limitations, we propose self-certainty, a novel and efficient metric that leverages the inherent probability distribution of LLM outputs to estimate response quality without requiring external reward models. We hypothesize that higher distributional self-certainty, aggregated across multiple samples, correlates with improved response accuracy, as it reflects greater confidence in the generated output. Through extensive experiments on various reasoning tasks, we demonstrate that self-certainty (1) scales effectively with increasing sample size $N$, akin to reward models but without the computational overhead; (2) complements chain-of-thought, improving reasoning performance beyond greedy decoding; and (3) generalizes to open-ended tasks where traditional self-consistency methods fall short. Our findings establish self-certainty as a practical and efficient way for improving LLM reasoning capabilities. The code is available at https://github.com/backprop07/Self-Certainty
Authors: M. E. Smith, N. Yilmaz, T. Watts, P. M. Scheikl, J. Ge, A. Deguet, A. Kuntz, A. Krieger
Abstract: Existing tracheal tumor resection methods often lack the precision required for effective airway clearance, and robotic advancements offer new potential for autonomous resection. We present a vision-guided, autonomous approach for palliative resection of tracheal tumors. This system models the tracheal surface with a fifth-degree polynomial to plan tool trajectories, while a custom Faster R-CNN segmentation pipeline identifies the trachea and tumor boundaries. The electrocautery tool angle is optimized using handheld surgical demonstrations, and trajectories are planned to maintain a 1 mm safety clearance from the tracheal surface. We validated the workflow successfully in five consecutive experiments on ex-vivo animal tissue models, successfully clearing the airway obstruction without trachea perforation in all cases (with more than 90% volumetric tumor removal). These results support the feasibility of an autonomous resection platform, paving the way for future developments in minimally-invasive autonomous resection.
Authors: Grace Guinan, Addison Salvador, Michelle A. Smeaton, Andrew Glaws, Hilary Egan, Brian C. Wyatt, Babak Anasori, Kevin R. Fiedler, Matthew J. Olszta, Steven R. Spurgeon
Abstract: What does materials science look like in the "Age of Artificial Intelligence?" Each materials domain-synthesis, characterization, and modeling-has a different answer to this question, motivated by unique challenges and constraints. This work focuses on the tremendous potential of autonomous characterization within electron microscopy. We present our recent advancements in developing domain-aware, multimodal models for microscopy analysis capable of describing complex atomic systems. We then address the critical gap between the theoretical promise of autonomous microscopy and its current practical limitations, showcasing recent successes while highlighting the necessary developments to achieve robust, real-world autonomy.
Authors: Miguel Herencia Garc\'ia del Castillo, Ricardo Moya Garcia, Manuel Jes\'us Cerezo Maz\'on, Ekaitz Arriola Garcia, Pablo Men\'endez Fern\'andez-Miranda
Abstract: In this article, we present a Latent Diffusion Model (LDM) for the generation of brain Magnetic Resonance Imaging (MRI), conditioning its generation based on pathology (Healthy, Glioblastoma, Sclerosis, Dementia) and acquisition modality (T1w, T1ce, T2w, Flair, PD). To evaluate the quality of the generated images, the Fr\'echet Inception Distance (FID) and Multi-Scale Structural Similarity Index (MS-SSIM) metrics were employed. The results indicate that the model generates images with a distribution similar to real ones, maintaining a balance between visual fidelity and diversity. Additionally, the model demonstrates extrapolation capability, enabling the generation of configurations that were not present in the training data. The results validate the potential of the model to increase in the number of samples in clinical datasets, balancing underrepresented classes, and evaluating AI models in medicine, contributing to the development of diagnostic tools in radiology without compromising patient privacy.
Authors: Matthew Barker, Andrew Bell, Evan Thomas, James Carr, Thomas Andrews, Umang Bhatt
Abstract: While Retrieval Augmented Generation (RAG) has emerged as a popular technique for improving Large Language Model (LLM) systems, it introduces a large number of choices, parameters and hyperparameters that must be made or tuned. This includes the LLM, embedding, and ranker models themselves, as well as hyperparameters governing individual RAG components. Yet, collectively optimizing the entire configuration in a RAG or LLM system remains under-explored - especially in multi-objective settings - due to intractably large solution spaces, noisy objective evaluations, and the high cost of evaluations. In this work, we introduce the first approach for multi-objective parameter optimization of cost, latency, safety and alignment over entire LLM and RAG systems. We find that Bayesian optimization methods significantly outperform baseline approaches, obtaining a superior Pareto front on two new RAG benchmark tasks. We conclude our work with important considerations for practitioners who are designing multi-objective RAG systems, highlighting nuances such as how optimal configurations may not generalize across tasks and objectives.
Authors: Markus Bertl, Alan Mott, Salvatore Sinno, Bhavika Bhalgamiya
Abstract: The digitization of healthcare presents numerous challenges, including the complexity of biological systems, vast data generation, and the need for personalized treatment plans. Traditional computational methods often fall short, leading to delayed and sometimes ineffective diagnoses and treatments. Quantum Computing (QC) and Quantum Machine Learning (QML) offer transformative advancements with the potential to revolutionize medicine. This paper summarizes areas where QC promises unprecedented computational power, enabling faster, more accurate diagnostics, personalized treatments, and enhanced drug discovery processes. However, integrating quantum technologies into precision medicine also presents challenges, including errors in algorithms and high costs. We show that mathematically-based techniques for specifying, developing, and verifying software (formal methods) can enhance the reliability and correctness of QC. By providing a rigorous mathematical framework, formal methods help to specify, develop, and verify systems with high precision. In genomic data analysis, formal specification languages can precisely (1) define the behavior and properties of quantum algorithms designed to identify genetic markers associated with diseases. Model checking tools can systematically explore all possible states of the algorithm to (2) ensure it behaves correctly under all conditions, while theorem proving techniques provide mathematical (3) proof that the algorithm meets its specified properties, ensuring accuracy and reliability. Additionally, formal optimization techniques can (4) enhance the efficiency and performance of quantum algorithms by reducing resource usage, such as the number of qubits and gate operations. Therefore, we posit that formal methods can significantly contribute to enabling QC to realize its full potential as a game changer in precision medicine.
Authors: Zhuoran Lu, Qian Zhou, Yi Wang
Abstract: Generative AI significantly enhances player agency in interactive narratives (IN) by enabling just-in-time content generation that adapts to player actions. While delegating generation to AI makes IN more interactive, it becomes challenging for authors to control the space of possible narratives - within which the final story experienced by the player emerges from their interaction with AI. In this paper, we present WhatELSE, an AI-bridged IN authoring system that creates narrative possibility spaces from example stories. WhatELSE provides three views (narrative pivot, outline, and variants) to help authors understand the narrative space and corresponding tools leveraging linguistic abstraction to control the boundaries of the narrative space. Taking innovative LLM-based narrative planning approaches, WhatELSE further unfolds the narrative space into executable game events. Through a user study (N=12) and technical evaluations, we found that WhatELSE enables authors to perceive and edit the narrative space and generates engaging interactive narratives at play-time.
Authors: Hediyeh Baban, Sai A Pidapar, Aashutosh Nema, Sichen Lu
Abstract: We introduce a novel multi-agent collaboration framework designed to enhance the accuracy and robustness of text classification models. Leveraging BERT as the primary classifier, our framework dynamically escalates low-confidence predictions to a specialized multi-agent system comprising Lexical, Contextual, Logic, Consensus, and Explainability agents. This collaborative approach allows for comprehensive analysis and consensus-driven decision-making, significantly improving classification performance across diverse text classification tasks. Empirical evaluations on benchmark datasets demonstrate that our framework achieves a 5.5% increase in accuracy compared to standard BERT-based classifiers, underscoring its effectiveness and academic novelty in advancing multi-agent systems within natural language processing.
Authors: Kevin Pu, Daniel Lazaro, Ian Arawjo, Haijun Xia, Ziang Xiao, Tovi Grossman, Yan Chen
Abstract: AI programming tools enable powerful code generation, and recent prototypes attempt to reduce user effort with proactive AI agents, but their impact on programming workflows remains unexplored. We introduce and evaluate Codellaborator, a design probe LLM agent that initiates programming assistance based on editor activities and task context. We explored three interface variants to assess trade-offs between increasingly salient AI support: prompt-only, proactive agent, and proactive agent with presence and context (Codellaborator). In a within-subject study (N=18), we find that proactive agents increase efficiency compared to prompt-only paradigm, but also incur workflow disruptions. However, presence indicators and \revise{interaction context support} alleviated disruptions and improved users' awareness of AI processes. We underscore trade-offs of Codellaborator on user control, ownership, and code understanding, emphasizing the need to adapt proactivity to programming processes. Our research contributes to the design exploration and evaluation of proactive AI systems, presenting design implications on AI-integrated programming workflow.
Authors: Yuexi Chen, Yimin Xiao, Kazi Tasnim Zinat, Naomi Yamashita, Ge Gao, Zhicheng Liu
Abstract: Understanding collaborative writing dynamics between native speakers (NS) and non-native speakers (NNS) is critical for enhancing collaboration quality and team inclusivity. In this paper, we partnered with communication researchers to develop visual analytics solutions for comparing NS and NNS behaviors in 162 writing sessions across 27 teams. The primary challenges in analyzing writing behaviors are data complexity and the uncertainties introduced by automated methods. In response, we present \textsc{COALA}, a novel visual analytics tool that improves model interpretability by displaying uncertainties in author clusters, generating behavior summaries using large language models, and visualizing writing-related actions at multiple granularities. We validated the effectiveness of \textsc{COALA} through user studies with domain experts (N=2+2) and researchers with relevant experience (N=8). We present the insights discovered by participants using \textsc{COALA}, suggest features for future AI-assisted collaborative writing tools, and discuss the broader implications for analyzing collaborative processes beyond writing.
Authors: Devansh Saxena, Ji-Youn Jung, Jodi Forlizzi, Kenneth Holstein, John Zimmerman
Abstract: AI systems are often introduced with high expectations, yet many fail to deliver, resulting in unintended harm and missed opportunities for benefit. We frequently observe significant "AI Mismatches", where the system's actual performance falls short of what is needed to ensure safety and co-create value. These mismatches are particularly difficult to address once development is underway, highlighting the need for early-stage intervention. Navigating complex, multi-dimensional risk factors that contribute to AI Mismatches is a persistent challenge. To address it, we propose an AI Mismatch approach to anticipate and mitigate risks early on, focusing on the gap between realistic model performance and required task performance. Through an analysis of 774 AI cases, we extracted a set of critical factors, which informed the development of seven matrices that map the relationships between these factors and highlight high-risk areas. Through case studies, we demonstrate how our approach can help reduce risks in AI development.
Authors: Konstantina Palla, Jos\'e Luis Redondo Garc\'ia, Claudia Hauff, Francesco Fabbri, Henrik Lindstr\"om, Daniel R. Taber, Andreas Damianou, Mounia Lalmas
Abstract: Content moderation plays a critical role in shaping safe and inclusive online environments, balancing platform standards, user expectations, and regulatory frameworks. Traditionally, this process involves operationalising policies into guidelines, which are then used by downstream human moderators for enforcement, or to further annotate datasets for training machine learning moderation models. However, recent advancements in large language models (LLMs) are transforming this landscape. These models can now interpret policies directly as textual inputs, eliminating the need for extensive data curation. This approach offers unprecedented flexibility, as moderation can be dynamically adjusted through natural language interactions. This paradigm shift raises important questions about how policies are operationalised and the implications for content moderation practices. In this paper, we formalise the emerging policy-as-prompt framework and identify five key challenges across four domains: Technical Implementation (1. translating policy to prompts, 2. sensitivity to prompt structure and formatting), Sociotechnical (3. the risk of technological determinism in policy formation), Organisational (4. evolving roles between policy and machine learning teams), and Governance (5. model governance and accountability). Through analysing these challenges across technical, sociotechnical, organisational, and governance dimensions, we discuss potential mitigation approaches. This research provides actionable insights for practitioners and lays the groundwork for future exploration of scalable and adaptive content moderation systems in digital ecosystems.
Authors: Robert Marlin, Raja Jurdak, Alsharif Abuadbba
Abstract: The widespread adoption of Electric Vehicles (EVs) poses critical challenges for energy providers, particularly in predicting charging time (temporal prediction), ensuring user privacy, and managing resources efficiently in mobility-driven networks. This paper introduces the Hierarchical Federated Learning Transformer Network (H-FLTN) framework to address these challenges. H-FLTN employs a three-tier hierarchical architecture comprising EVs, community Distributed Energy Resource Management Systems (DERMS), and the Energy Provider Data Centre (EPDC) to enable accurate spatio-temporal predictions of EV charging needs while preserving privacy. Temporal prediction is enhanced using Transformer-based learning, capturing complex dependencies in charging behavior. Privacy is ensured through Secure Aggregation, Additive Secret Sharing, and Peer-to-Peer (P2P) Sharing with Augmentation, which allow only secret shares of model weights to be exchanged while securing all transmissions. To improve training efficiency and resource management, H-FLTN integrates Dynamic Client Capping Mechanism (DCCM) and Client Rotation Management (CRM), ensuring that training remains both computationally and temporally efficient as the number of participating EVs increases. DCCM optimises client participation by limiting excessive computational loads, while CRM balances training contributions across epochs, preventing imbalanced participation. Our simulation results based on large-scale empirical vehicle mobility data reveal that DCCM and CRM reduce the training time complexity with increasing EVs from linear to constant. Its integration into real-world smart city infrastructure enhances energy demand forecasting, resource allocation, and grid stability, ensuring reliability and sustainability in future mobility ecosystems.
Authors: Chaitanya Kapoor, Sudhanshu Srivastava, Meenakshi Khosla
Abstract: Understanding convergent learning -- the extent to which artificial and biological neural networks develop similar representations -- is crucial for neuroscience and AI, as it reveals shared learning principles and guides brain-like model design. While several studies have noted convergence in early and late layers of vision networks, key gaps remain. First, much existing work relies on a limited set of metrics, overlooking transformation invariances required for proper alignment. We compare three metrics that ignore specific irrelevant transformations: linear regression (ignoring affine transformations), Procrustes (ignoring rotations and reflections), and permutation/soft-matching (ignoring unit order). Notably, orthogonal transformations align representations nearly as effectively as more flexible linear ones, and although permutation scores are lower, they significantly exceed chance, indicating a robust representational basis. A second critical gap lies in understanding when alignment emerges during training. Contrary to expectations that convergence builds gradually with task-specific learning, our findings reveal that nearly all convergence occurs within the first epoch -- long before networks achieve optimal performance. This suggests that shared input statistics, architectural biases, or early training dynamics drive convergence rather than the final task solution. Finally, prior studies have not systematically examined how changes in input statistics affect alignment. Our work shows that out-of-distribution (OOD) inputs consistently amplify differences in later layers, while early layers remain aligned for both in-distribution and OOD inputs, suggesting that this alignment is driven by generalizable features stable across distribution shifts. These findings fill critical gaps in our understanding of representational convergence, with implications for neuroscience and AI.
Authors: Alireza Daghighfarsoodeh, Chung-Yu Wang, Hamed Taherkhani, Melika Sepidband, Mohammad Abdollahi, Hadi Hemmati, Hung Viet Pham
Abstract: Deep learning (DL) has revolutionized areas such as computer vision, natural language processing, and more. However, developing DL systems is challenging due to the complexity of DL workflows. Large Language Models (LLMs), such as GPT, Claude, Llama, Mistral, etc., have emerged as promising tools to assist in DL code generation, offering potential solutions to these challenges. Despite this, existing benchmarks such as DS-1000 are limited, as they primarily focus on small DL code snippets related to pre/post-processing tasks and lack a comprehensive coverage of the full DL pipeline, including different DL phases and input data types. To address this, we introduce DeepBench, a novel benchmark dataset designed for function-level DL code generation. DeepBench categorizes DL problems based on three key aspects: phases such as pre-processing, model construction, and training; tasks, including classification, regression, and recommendation; and input data types such as tabular, image, and text. GPT-4o -- the state-of-the-art LLM -- achieved 31% accuracy on DeepBench, significantly lower than its 60% on DS-1000. We observed similar difficulty for other LLMs (e.g., 28% vs. 54% for Claude, 21% vs. 41% for LLaMA, and 15% vs. 20% for Mistral). This result underscores DeepBench's greater complexity. We also construct a taxonomy of issues and bugs found in LLM-generated DL code, which highlights the distinct challenges that LLMs face when generating DL code compared to general code. Furthermore, our analysis also reveals substantial performance variations across categories, with differences of up to 7% among phases and 37% among tasks. These disparities suggest that DeepBench offers valuable insights into the LLMs' performance and areas for potential improvement in the DL domain.
Authors: Eric Oliver, Sagnik Dakshit
Abstract: Deep learning's growing prevalence has driven its widespread use in healthcare, where AI and sensor advancements enhance diagnosis, treatment, and monitoring. In mobile health, AI-powered tools enable early diagnosis and continuous monitoring of conditions like stress. Wearable technologies and multimodal physiological data have made stress detection increasingly viable, but model efficacy depends on data quality, quantity, and modality. This study develops transformer models for stress detection using the WESAD dataset, training on electrocardiograms (ECG), electrodermal activity (EDA), electromyography (EMG), respiration rate (RESP), temperature (TEMP), and 3-axis accelerometer (ACC) signals. The results demonstrate the effectiveness of single-modality transformers in analyzing physiological signals, achieving state-of-the-art performance with accuracy, precision and recall values in the range of $99.73\%$ to $99.95\%$ for stress detection. Furthermore, this study explores cross-modal performance and also explains the same using 2D visualization of the learned embedding space and quantitative analysis based on data variance. Despite the large body of work on stress detection and monitoring, the robustness and generalization of these models across different modalities has not been explored. This research represents one of the initial efforts to interpret embedding spaces for stress detection, providing valuable information on cross-modal performance.
Authors: Nathalie Riche, Anna Offenwanger, Frederic Gmeiner, David Brown, Hugo Romat, Michel Pahud, Nicolai Marquardt, Kori Inkpen, Ken Hinckley
Abstract: Chat-based prompts respond with verbose linear-sequential texts, making it difficult to explore and refine ambiguous intents, back up and reinterpret, or shift directions in creative AI-assisted design work. AI-Instruments instead embody "prompts" as interface objects via three key principles: (1) Reification of user-intent as reusable direct-manipulation instruments; (2) Reflection of multiple interpretations of ambiguous user-intents (Reflection-in-intent) as well as the range of AI-model responses (Reflection-in-response) to inform design "moves" towards a desired result; and (3) Grounding to instantiate an instrument from an example, result, or extrapolation directly from another instrument. Further, AI-Instruments leverage LLM's to suggest, vary, and refine new instruments, enabling a system that goes beyond hard-coded functionality by generating its own instrumental controls from content. We demonstrate four technology probes, applied to image generation, and qualitative insights from twelve participants, showing how AI-Instruments address challenges of intent formulation, steering via direct manipulation, and non-linear iterative workflows to reflect and resolve ambiguous intents.
Authors: Frederic Gmeiner, Nicolai Marquardt, Michael Bentley, Hugo Romat, Michel Pahud, David Brown, Asta Roseway, Nikolas Martelaro, Kenneth Holstein, Ken Hinckley, Nathalie Riche
Abstract: Despite Generative AI (GenAI) systems' potential for enhancing content creation, users often struggle to effectively integrate GenAI into their creative workflows. Core challenges include misalignment of AI-generated content with user intentions (intent elicitation and alignment), user uncertainty around how to best communicate their intents to the AI system (prompt formulation), and insufficient flexibility of AI systems to support diverse creative workflows (workflow flexibility). Motivated by these challenges, we created IntentTagger: a system for slide creation based on the notion of Intent Tags - small, atomic conceptual units that encapsulate user intent - for exploring granular and non-linear micro-prompting interactions for Human-GenAI co-creation workflows. Our user study with 12 participants provides insights into the value of flexibly expressing intent across varying levels of ambiguity, meta-intent elicitation, and the benefits and challenges of intent tag-driven workflows. We conclude by discussing the broader implications of our findings and design considerations for GenAI-supported content creation workflows.
Authors: Yuwei Yan, Yu Shang, Qingbin Zeng, Yu Li, Keyu Zhao, Zhiheng Zheng, Xuefei Ning, Tianji Wu, Shengen Yan, Yu Wang, Fengli Xu, Yong Li
Abstract: The AgentSociety Challenge is the first competition in the Web Conference that aims to explore the potential of Large Language Model (LLM) agents in modeling user behavior and enhancing recommender systems on web platforms. The Challenge consists of two tracks: the User Modeling Track and the Recommendation Track. Participants are tasked to utilize a combined dataset from Yelp, Amazon, and Goodreads, along with an interactive environment simulator, to develop innovative LLM agents. The Challenge has attracted 295 teams across the globe and received over 1,400 submissions in total over the course of 37 official competition days. The participants have achieved 21.9% and 20.3% performance improvement for Track 1 and Track 2 in the Development Phase, and 9.1% and 15.9% in the Final Phase, representing a significant accomplishment. This paper discusses the detailed designs of the Challenge, analyzes the outcomes, and highlights the most successful LLM agent designs. To support further research and development, we have open-sourced the benchmark environment at https://tsinghua-fib-lab.github.io/AgentSocietyChallenge.
URLs: https://tsinghua-fib-lab.github.io/AgentSocietyChallenge.
Authors: Akhil Nagariya, Dimitar Filev, Srikanth Saripalli, Gaurav Pandey
Abstract: In the area of autonomous driving, navigating off-road terrains presents a unique set of challenges, from unpredictable surfaces like grass and dirt to unexpected obstacles such as bushes and puddles. In this work, we present a novel learning-based local planner that addresses these challenges by directly capturing human driving nuances from real-world demonstrations using only a monocular camera. The key features of our planner are its ability to navigate in challenging off-road environments with various terrain types and its fast learning capabilities. By utilizing minimal human demonstration data (5-10 mins), it quickly learns to navigate in a wide array of off-road conditions. The local planner significantly reduces the real world data required to learn human driving preferences. This allows the planner to apply learned behaviors to real-world scenarios without the need for manual fine-tuning, demonstrating quick adjustment and adaptability in off-road autonomous driving technology.
Authors: Nicolas Michel, Maorong Wang, Jiangpeng He, Toshihiko Yamasaki
Abstract: Continual Learning (CL) addresses the problem of learning from a data sequence where the distribution changes over time. Recently, efficient solutions leveraging Pre-Trained Models (PTM) have been widely explored in the offline CL (offCL) scenario, where the data corresponding to each incremental task is known beforehand and can be seen multiple times. However, such solutions often rely on 1) prior knowledge regarding task changes and 2) hyper-parameter search, particularly regarding the learning rate. Both assumptions remain unavailable in online CL (onCL) scenarios, where incoming data distribution is unknown and the model can observe each datum only once. Therefore, existing offCL strategies fall largely behind performance-wise in onCL, with some proving difficult or impossible to adapt to the online scenario. In this paper, we tackle both problems by leveraging Online Prototypes (OP) and Class-Wise Hypergradients (CWH). OP leverages stable output representations of PTM by updating its value on the fly to act as replay samples without requiring task boundaries or storing past data. CWH learns class-dependent gradient coefficients during training to improve over sub-optimal learning rates. We show through experiments that both introduced strategies allow for a consistent gain in accuracy when integrated with existing approaches. We will make the code fully available upon acceptance.
Authors: Jiayi Fu, Xuandong Zhao, Chengyuan Yao, Heng Wang, Qi Han, Yanghua Xiao
Abstract: Reinforcement Learning from Human Feedback (RLHF) is essential for aligning large language models (LLMs) with human values. However, RLHF is susceptible to reward hacking, where the agent exploits flaws in the reward function rather than learning the intended behavior, thus degrading alignment. While reward shaping helps stabilize RLHF and partially mitigate reward hacking, a systematic investigation into shaping techniques and their underlying principles remains lacking. To bridge this gap, we present a comprehensive study of the prevalent reward shaping methods. Our analysis suggests three key design principles: (1) RL reward is ideally bounded, (2) RL benefits from rapid initial growth followed by gradual convergence, and (3) RL reward is best formulated as a function of centered reward. Guided by these insights, we propose Preference As Reward (PAR), a novel approach that leverages the latent preferences embedded within the reward model itself as the signal for reinforcement learning. We evaluated PAR on two base models, Gemma2-2B and Llama3-8B, using two datasets, Ultrafeedback-Binarized and HH-RLHF. Experimental results demonstrate PAR's superior performance over other reward shaping methods. On the AlpacaEval 2.0 benchmark, PAR achieves a win rate at least 5 percentage points higher than competing approaches. Furthermore, PAR exhibits remarkable data efficiency, requiring only a single reference reward for optimal performance, and maintains robustness against reward hacking even after two full epochs of training. Code is available at https://github.com/PorUna-byte/PAR.
Authors: Yuqing Wang, Xiao Yang
Abstract: This study addresses the challenge of resource scheduling optimization in edge-cloud collaborative computing using deep reinforcement learning (DRL). The proposed DRL-based approach improves task processing efficiency, reduces overall processing time, enhances resource utilization, and effectively controls task migrations. Experimental results demonstrate the superiority of DRL over traditional scheduling algorithms, particularly in managing complex task allocation, dynamic workloads, and multiple resource constraints. Despite its advantages, further improvements are needed to enhance learning efficiency, reduce training time, and address convergence issues. Future research should focus on increasing the algorithm's fault tolerance to handle more complex and uncertain scheduling scenarios, thereby advancing the intelligence and efficiency of edge-cloud computing systems.
Authors: Jun-En Ding, Dongsheng Luo, Anna Zilverstand, Feng Liu
Abstract: Analyzing functional brain networks using functional magnetic resonance imaging (fMRI) is crucial for understanding psychiatric disorders and addictive behaviors. While existing fMRI-based graph convolutional networks (GCNs) show considerable promise for feature extraction, they often fall short in characterizing complex relationships between brain regions and demographic factors and accounting for interpretable variables linked to psychiatric conditions. We propose NeuroTree to overcome these limitations, integrating a k-hop AGE-GCN with neural ordinary differential equations (ODEs). This framework leverages an attention mechanism to optimize functional connectivity (FC), thereby enhancing dynamic FC feature learning for brain disease classification. Furthermore, NeuroTree effectively decodes fMRI network features into tree structures, which improves the capture of high-order brain regional pathway features and enables the identification of hierarchical neural behavioral patterns essential for understanding disease-related brain subnetworks. Our empirical evaluations demonstrate that NeuroTree achieves state-of-the-art performance across two distinct mental disorder datasets and provides valuable insights into age-related deterioration patterns. These findings underscore the model's efficacy in predicting psychiatric disorders and elucidating their underlying neural mechanisms.
Authors: Jungsoo Park, Junmo Kang, Gabriel Stanovsky, Alan Ritter
Abstract: The surge of LLM studies makes synthesizing their findings challenging. Meta-analysis can uncover important trends across studies, but its use is limited by the time-consuming nature of manual data extraction. Our study presents a semi-automated approach for meta-analysis that accelerates data extraction using LLMs. It automatically identifies relevant arXiv papers, extracts experimental results and related attributes, and organizes them into a structured dataset. We conduct a comprehensive meta-analysis of frontier LLMs using an automatically extracted dataset, reducing the effort of paper surveying and data extraction by more than 93\% compared to manual approaches. We validate our dataset by showing that it reproduces key findings from a recent manual meta-analysis about Chain-of-Thought (CoT), and also uncovers new insights that go beyond it, showing for example that in-context examples benefit multimodal tasks but offer limited gains in mathematical tasks compared to CoT. Our automatically updatable dataset enables continuous tracking of target models by extracting evaluation studies as new data becomes available. Through our scientific artifacts and empirical analysis, we provide novel insights into LLMs while facilitating ongoing meta-analyses of their behavior.
Authors: Gyeongje Cho, Yeonkyoung So, Jaejin Lee
Abstract: Multiple-choice benchmarks, consisting of various prompts and choices, are among the most widely used methods to assess a language model's natural language understanding capability. Given a specific prompt, we typically compute $P(Choice|Prompt)$ to evaluate how likely a language model is to generate the correct choice compared to incorrect ones. However, we observe that performance measured using this approach reflects not only the model's comprehension of the prompt but also its inherent biases for certain choices regardless of the prompt. This issue makes it challenging to accurately measure a model's natural language understanding, as models may select the answer without fully understanding the prompt. To address this limitation, we propose a novel metric called ANPMI, which normalizes Pointwise Mutual Information (PMI) by $-\log P(Choice)$. ANPMI provides a more accurate assessment of the model's natural language understanding by ensuring that it is challenging to answer a question without properly understanding the prompt.
Authors: Ruifeng Tan, Weixiang Hong, Jiayue Tang, Xibin Lu, Ruijun Ma, Xiang Zheng, Jia Li, Jiaqiang Huang, Tong-Yi Zhang
Abstract: Battery Life Prediction (BLP), which relies on time series data produced by battery degradation tests, is crucial for battery utilization, optimization, and production. Despite impressive advancements, this research area faces three key challenges. Firstly, the limited size of existing datasets impedes insights into modern battery life data. Secondly, most datasets are restricted to small-capacity lithium-ion batteries tested under a narrow range of diversity in labs, raising concerns about the generalizability of findings. Thirdly, inconsistent and limited benchmarks across studies obscure the effectiveness of baselines and leave it unclear if models popular in other time series fields are effective for BLP. To address these challenges, we propose BatteryLife, a comprehensive dataset and benchmark for BLP. BatteryLife integrates 16 datasets, offering a 2.4 times sample size compared to the previous largest dataset, and provides the most diverse battery life resource with batteries from 8 formats, 80 chemical systems, 12 operating temperatures, and 646 charge/discharge protocols, including both laboratory and industrial tests. Notably, BatteryLife is the first to release battery life datasets of zinc-ion batteries, sodium-ion batteries, and industry-tested large-capacity lithium-ion batteries. With the comprehensive dataset, we revisit the effectiveness of baselines popular in this and other time series fields. Furthermore, we propose CyclePatch, a plug-in technique that can be employed in a series of neural networks. Extensive benchmarking of 18 methods reveals that models popular in other time series fields can be unsuitable for BLP, and CyclePatch consistently improves model performance establishing state-of-the-art benchmarks. Moreover, BatteryLife evaluates model performance across aging conditions and domains. BatteryLife is available at https://github.com/Ruifeng-Tan/BatteryLife.
Authors: Muhammad A. Muttaqien, Tomohiro Motoda, Ryo Hanai, Domae Yukiyasu
Abstract: This paper introduces a novel pipeline to enhance the precision of object masking for robotic manipulation within the specific domain of masking products in convenience stores. The approach integrates two advanced AI models, CLIP and SAM, focusing on their synergistic combination and the effective use of multimodal data (image and text). Emphasis is placed on utilizing gradient-based attention mechanisms and customized datasets to fine-tune performance. While CLIP, SAM, and Grad- CAM are established components, their integration within this structured pipeline represents a significant contribution to the field. The resulting segmented masks, generated through this combined approach, can be effectively utilized as inputs for robotic systems, enabling more precise and adaptive object manipulation in the context of convenience store products.
Authors: Yunmei Huang, Songlin Hou, Zachary Nelson Horve, Songlin Fei
Abstract: The precise identification of tree species is fundamental to forestry, conservation, and environmental monitoring. Though many studies have demonstrated that high accuracy can be achieved using bark-based species classification, these models often function as "black boxes", limiting interpretability, trust, and adoption in critical forestry applications. Attribution-based Explainable AI (XAI) methods have been used to address this issue in related works. However, XAI applications are often dependent on local features (such as a head shape or paw in animal applications) and cannot describe global visual features (such as ruggedness or smoothness) that are present in texture-dominant images such as tree bark. Concept-based XAI methods, on the other hand, offer explanations based on global visual features with concepts, but they tend to require large overhead in building external concept image datasets and the concepts can be vague and subjective without good means of precise quantification. To address these challenges, we propose a lightweight post-hoc method to interpret visual models for tree species classification using operators and quantifiable concepts. Our approach eliminates computational overhead, enables the quantification of complex concepts, and evaluates both concept importance and the model's reasoning process. To the best of our knowledge, our work is the first study to explain bark vision models in terms of global visual features with concepts. Using a human-annotated dataset as ground truth, our experiments demonstrate that our method significantly outperforms TCAV and Llama3.2 in concept importance ranking based on Kendall's Tau, highlighting its superior alignment with human perceptions.
Authors: Zichuan Fu, Wentao Song, Yejing Wang, Xian Wu, Yefeng Zheng, Yingying Zhang, Derong Xu, Xuetao Wei, Tong Xu, Xiangyu Zhao
Abstract: Recent advances in transformer-based Large Language Models (LLMs) have demonstrated remarkable capabilities across various tasks. However, their quadratic computational complexity concerning sequence length remains a significant bottleneck for processing long documents. As a result, many efforts like sparse attention and state space models have been proposed to improve the efficiency of LLMs over long sequences. Though effective, these approaches compromise the performance or introduce structural complexity. This calls for a simple yet efficient model that preserves the fundamental Transformer architecture. To this end, we introduce SWAT, which enables efficient long-context handling via Sliding Window Attention Training. This paper first attributes the inefficiency of Transformers to the attention sink phenomenon resulting from the high variance of softmax operation. Then, we replace softmax with the sigmoid function and utilize a balanced ALiBi and Rotary Position Embedding for efficient information compression and retention. Experiments demonstrate that SWAT achieves SOTA performance compared with state-of-the-art linear recurrent architectures on eight benchmarks. Code is available at https://anonymous.4open.science/r/SWAT-attention.
Authors: Kerem Zaman, Shashank Srivastava
Abstract: Large Language Models (LLMs) offer natural language explanations as an alternative to feature attribution methods for model interpretability. However, despite their plausibility, they may not reflect the model's internal reasoning faithfully, which is crucial for understanding the model's true decision-making processes. Although several faithfulness metrics have been proposed, a unified evaluation framework remains absent. To address this gap, we present Causal Diagnosticity, a framework to evaluate faithfulness metrics for natural language explanations. Our framework employs the concept of causal diagnosticity, and uses model-editing methods to generate faithful-unfaithful explanation pairs. Our benchmark includes four tasks: fact-checking, analogy, object counting, and multi-hop reasoning. We evaluate a variety of faithfulness metrics, including post-hoc explanation and chain-of-thought-based methods. We find that all tested faithfulness metrics often fail to surpass a random baseline. Our work underscores the need for improved metrics and more reliable interpretability methods in LLMs.
Authors: Jungin Kim, Shinwoo Park, Yo-Sub Han
Abstract: Code watermarking identifies AI-generated code by embedding patterns into the code during generation. Effective watermarking requires meeting two key conditions: the watermark should be reliably detectable, and the code should retain its original functionality. However, existing methods often modify tokens that are critical for program logic, such as keywords in conditional expressions or operators in arithmetic computations. These modifications can cause syntax errors or functional failures, limiting the practical use of watermarking. We present STONE, a method that preserves functional integrity by selectively inserting watermarks only into non-syntax tokens. By excluding tokens essential for code execution, STONE minimizes the risk of functional degradation. In addition, we introduce CWEM, a comprehensive evaluation metric that evaluates watermarking techniques based on correctness, detectability, and naturalness. While correctness and detectability have been widely used, naturalness remains underexplored despite its importance. Unnatural patterns can reveal the presence of a watermark, making it easier for adversaries to remove. We evaluate STONE using CWEM and compare its performance with the state-of-the-art approach. The results show that STONE achieves an average improvement of 7.69% in CWEM across Python, C++, and Java. Our code is available in https://github.com/inistory/STONE-watermarking/.
Authors: Soobin Park, Hankyung Kim, Youn-kyung Lim
Abstract: Image-generative AI provides new opportunities to transform personal data into alternative visual forms. In this paper, we illustrate the potential of AI-generated images in facilitating meaningful engagement with personal data. In a formative autobiographical design study, we explored the design and use of AI-generated images derived from personal data. Informed by this study, we designed a web-based application as a probe that represents personal data through generative images utilizing Open AI's GPT-4 model and DALL-E 3. We then conducted a 21-day diary study and interviews using the probe with 16 participants to investigate users' in-depth experiences with images generated by AI in everyday lives. Our findings reveal new qualities of experiences in users' engagement with data, highlighting how participants constructed personal meaning from their data through imagination and speculation on AI-generated images. We conclude by discussing the potential and concerns of leveraging image-generative AI for personal data meaning-making.
Authors: Jacob Dunefsky, Arman Cohan
Abstract: Steering vectors have emerged as a promising approach for interpreting and controlling LLMs, but current methods typically require large contrastive datasets that are often impractical to construct and may capture spurious correlations. We propose directly optimizing steering vectors through gradient descent on a single training example, and systematically investigate how these vectors generalize. We consider several steering optimization techniques, including multiple novel ones, and find that the resulting vectors effectively mediate safety-relevant behaviors in multiple models. Indeed, in experiments on an alignment-faking model, we are able to optimize one-shot steering vectors that induce harmful behavior on benign examples and whose negations suppress harmful behavior on malign examples. And in experiments on refusal suppression, we demonstrate that one-shot optimized steering vectors can transfer across inputs, yielding a Harmbench attack success rate of 96.9%. Furthermore, to quantitatively assess steering effectiveness in instruction-tuned models, we develop a novel evaluation framework using sequence probabilities from the corresponding base model. With this framework, we analyze how steering vectors modulate an instruction-tuned LLM's ability to recover from outputting false information, and find that this ability derives from the base model. Overall, our findings suggest that optimizing steering vectors on a single example can mediate misaligned behavior in LLMs, and provide a path toward better understanding the relationship between LLM behavior and activation space structure.
Authors: Junxiao Ma, Jingjing Wang, Jiamin Luo, Peiying Yu, Guodong Zhou
Abstract: Prior studies on Video Anomaly Detection (VAD) mainly focus on detecting whether each video frame is abnormal or not in the video, which largely ignore the structured video semantic information (i.e., what, when, and where does the abnormal event happen). With this in mind, we propose a new chat-paradigm \textbf{M}ulti-scene Video Abnormal Event Extraction and Localization (M-VAE) task, aiming to extract the abnormal event quadruples (i.e., subject, event type, object, scene) and localize such event. Further, this paper believes that this new task faces two key challenges, i.e., global-local spatial modeling and global-local spatial balancing. To this end, this paper proposes a Global-local Spatial-sensitive Large Language Model (LLM) named Sherlock, i.e., acting like Sherlock Holmes to track down the criminal events, for this M-VAE task. Specifically, this model designs a Global-local Spatial-enhanced MoE (GSM) module and a Spatial Imbalance Regulator (SIR) to address the two challenges respectively. Extensive experiments on our M-VAE instruction dataset show the significant advantages of Sherlock over several advanced Video-LLMs. This justifies the importance of global-local spatial information for the M-VAE task and the effectiveness of Sherlock in capturing such information.
Authors: Shi Fu, Yingjie Wang, Yuzhu Chen, Xinmei Tian, Dacheng Tao
Abstract: High-quality data is essential for training large generative models, yet the vast reservoir of real data available online has become nearly depleted. Consequently, models increasingly generate their own data for further training, forming Self-consuming Training Loops (STLs). However, the empirical results have been strikingly inconsistent: some models degrade or even collapse, while others successfully avoid these failures, leaving a significant gap in theoretical understanding to explain this discrepancy. This paper introduces the intriguing notion of recursive stability and presents the first theoretical generalization analysis, revealing how both model architecture and the proportion between real and synthetic data influence the success of STLs. We further extend this analysis to transformers in in-context learning, showing that even a constant-sized proportion of real data ensures convergence, while also providing insights into optimal synthetic data sizing.
Authors: Youngtae Kim, Soonju Jeong, Sardar Arslan, Dhananjay Agnihotri, Yahya Ahmed, Ali Nawaz, Jinhee Song, Hyewon Kim
Abstract: This study proposes a two-phase methodology for detecting and classifying auxiliary insulation in structural components. In the detection phase, a YOLOv8x model is trained on a dataset of complete structural blueprints, each annotated with bounding boxes indicating areas that should contain insulation. In the classification phase, these detected insulation patches are cropped and categorized into two classes: present or missing. These are then used to train a YOLOv8x-CLS model that determines the presence or absence of auxiliary insulation. Preprocessing steps for both datasets included annotation, augmentation, and appropriate cropping of the insulation regions. The detection model achieved a mean average precision (mAP) score of 82%, while the classification model attained an accuracy of 98%. These findings demonstrate the effectiveness of the proposed approach in automating insulation detection and classification, providing a foundation for further advancements in this domain.
Authors: Kaishuai Xu, Tiezheng Yu, Wenjun Hou, Yi Cheng, Liangyou Li, Xin Jiang, Lifeng Shang, Qun Liu, Wenjie Li
Abstract: Large Language Models (LLMs) are being used more and more extensively for automated evaluation in various scenarios. Previous studies have attempted to fine-tune open-source LLMs to replicate the evaluation explanations and judgments of powerful proprietary models, such as GPT-4. However, these methods are largely limited to text-based analyses under predefined general criteria, resulting in reduced adaptability for unseen instructions and demonstrating instability in evaluating adherence to quantitative and structural constraints. To address these limitations, we propose a novel evaluation framework, ARJudge, that adaptively formulates evaluation criteria and synthesizes both text-based and code-driven analyses to evaluate LLM responses. ARJudge consists of two components: a fine-tuned Analyzer that generates multi-faceted evaluation analyses and a tuning-free Refiner that combines and refines all analyses to make the final judgment. We construct a Composite Analysis Corpus that integrates tasks for evaluation criteria generation alongside text-based and code-driven analysis generation to train the Analyzer. Our results demonstrate that ARJudge outperforms existing fine-tuned evaluators in effectiveness and robustness. Furthermore, it demonstrates the importance of multi-faceted evaluation and code-driven analyses in enhancing evaluation capabilities.
Authors: Fanglei Xue, Meihan Zhang, Shuqi Li, Xinyu Gao, James A. Wohlschlegel, Wenbing Huang, Yi Yang, Weixian Deng
Abstract: Targeted protein degradation (TPD) induced by small molecules has emerged as a rapidly evolving modality in drug discovery, targeting proteins traditionally considered "undruggable". Proteolysis-targeting chimeras (PROTACs) and molecular glue degraders (MGDs) are the primary small molecules that induce TPD. Both types of molecules form a ternary complex linking an E3 ligase with a target protein, a crucial step for drug discovery. While significant advances have been made in binary structure prediction for proteins and small molecules, ternary structure prediction remains challenging due to obscure interaction mechanisms and insufficient training data. Traditional methods relying on manually assigned rules perform poorly and are computationally demanding due to extensive random sampling. In this work, we introduce DeepTernary, a novel deep learning-based approach that directly predicts ternary structures in an end-to-end manner using an encoder-decoder architecture. DeepTernary leverages an SE(3)-equivariant graph neural network (GNN) with both intra-graph and ternary inter-graph attention mechanisms to capture intricate ternary interactions from our collected high-quality training dataset, TernaryDB. The proposed query-based Pocket Points Decoder extracts the 3D structure of the final binding ternary complex from learned ternary embeddings, demonstrating state-of-the-art accuracy and speed in existing PROTAC benchmarks without prior knowledge from known PROTACs. It also achieves notable accuracy on the more challenging MGD benchmark under the blind docking protocol. Remarkably, our experiments reveal that the buried surface area calculated from predicted structures correlates with experimentally obtained degradation potency-related metrics. Consequently, DeepTernary shows potential in effectively assisting and accelerating the development of TPDs for previously undruggable targets.
Authors: Tianyun Liu
Abstract: Traditional text-to-speech (TTS) methods primarily focus on establishing a mapping between phonemes and mel-spectrograms. However, during the phoneme encoding stage, there is often a lack of real mel-spectrogram auxiliary information, which results in the encoding process lacking true semantic understanding. At the same time, traditional TTS systems often struggle to balance the inference speed of the model with the quality of the synthesized speech. Methods that generate high-quality synthesized speech tend to have slower inference speeds, while faster inference methods often sacrifice speech quality. In this paper, I propose Clip-TTS, a TTS method based on the Clip architecture. This method uses the Clip framework to establish a connection between text content and real mel-spectrograms during the text encoding stage, enabling the text encoder to directly learn the true semantics of the global context, thereby ensuring the quality of the synthesized speech. In terms of model architecture, I adopt the basic structure of Transformer, which allows Clip-TTS to achieve fast inference speeds. Experimental results show that on the LJSpeech and Baker datasets, the speech generated by Clip-TTS achieves state-of-the-art MOS scores, and it also performs excellently on multi-emotion datasets.Audio samples are available at: https://ltydd1314.github.io/.
Authors: Ziyuan Zhong, Junyang Zhou
Abstract: In this paper, we propose an innovative dynamic classification algorithm designed to achieve the objective of zero missed detections and minimal false positives. The algorithm partitions the data into N equivalent training subsets and N prediction subsets using a supervised model, followed by independent predictions from N separate predictive models. This enables each predictive model to operate within a smaller data range, thereby improving overall accuracy. Additionally, the algorithm leverages data generated through supervised learning to further refine prediction results, filtering out predictions that do not meet accuracy requirements without the need to introduce additional models. Experimental results demonstrate that, when data partitioning errors are minimal, the dynamic classification algorithm achieves exceptional performance with zero missed detections and minimal false positives, significantly outperforming existing model ensembles. Even in cases where classification errors are larger, the algorithm remains comparable to state of the art models. The key innovations of this study include self-supervised classification learning, the use of small-range subset predictions, and the direct rejection of substandard predictions. While the current algorithm still has room for improvement in terms of automatic parameter tuning and classification model efficiency, it has demonstrated outstanding performance across multiple datasets. Future research will focus on optimizing the classification component to further enhance the algorithm's robustness and adaptability.
Authors: Hongye Jin, Pei Chen, Jingfeng Yang, Zhengyang Wang, Meng Jiang, Yifan Gao, Binxuan Huang, Xinyang Zhang, Zheng Li, Tianyi Liu, Huasheng Li, Bing Yin
Abstract: Large Language Models (LLMs) have demonstrated remarkable performance across a wide range of natural language processing tasks. However, they are often distracted by irrelevant or noisy context in input sequences that degrades output quality. This problem affects both long- and short-context scenarios, such as retrieval-augmented generation, table question-answering, and in-context learning. We reveal that LLMs can implicitly identify whether input sequences contain useful information at early layers, prior to token generation. Leveraging this insight, we introduce Early Noise Dropping (\textsc{END}), a novel approach to mitigate this issue without requiring fine-tuning the LLMs. \textsc{END} segments input sequences into chunks and employs a linear prober on the early layers of LLMs to differentiate between informative and noisy chunks. By discarding noisy chunks early in the process, \textsc{END} preserves critical information, reduces distraction, and lowers computational overhead. Extensive experiments demonstrate that \textsc{END} significantly improves both performance and efficiency across different LLMs on multiple evaluation datasets. Furthermore, by investigating LLMs' implicit understanding to the input with the prober, this work also deepens understanding of how LLMs do reasoning with contexts internally.
Authors: Weiyan Wang, Xingjian Shi, Ruiqi Shu, Yuan Gao, Rui Ray Chen, Kun Wang, Fan Xu, Jinbao Xue, Shuaipeng Li, Yangyu Tao, Di Wang, Hao Wu, Xiaomeng Huang
Abstract: In practice, physical spatiotemporal forecasting can suffer from data scarcity, because collecting large-scale data is non-trivial, especially for extreme events. Hence, we propose \method{}, a novel probabilistic framework to realize iterative self-training with new self-ensemble strategies, achieving better physical consistency and generalization on extreme events. Following any base forecasting model, we can encode its deterministic outputs into a latent space and retrieve multiple codebook entries to generate probabilistic outputs. Then BeamVQ extends the beam search from discrete spaces to the continuous state spaces in this field. We can further employ domain-specific metrics (e.g., Critical Success Index for extreme events) to filter out the top-k candidates and develop the new self-ensemble strategy by combining the high-quality candidates. The self-ensemble can not only improve the inference quality and robustness but also iteratively augment the training datasets during continuous self-training. Consequently, BeamVQ realizes the exploration of rare but critical phenomena beyond the original dataset. Comprehensive experiments on different benchmarks and backbones show that BeamVQ consistently reduces forecasting MSE (up to 39%), enhancing extreme events detection and proving its effectiveness in handling data scarcity.
Authors: Yangfan Xu, Qu Hao, Lilian Zhang, Jun Mao, Xiaofeng He, Wenqi Wu, Changhao Chen
Abstract: Visual SLAM is essential for mobile robots, drone navigation, and VR/AR, but traditional RGB camera systems struggle in low-light conditions, driving interest in thermal SLAM, which excels in such environments. However, thermal imaging faces challenges like low contrast, high noise, and limited large-scale annotated datasets, restricting the use of deep learning in outdoor scenarios. We present DarkSLAM, a noval deep learning-based monocular thermal SLAM system designed for large-scale localization and reconstruction in complex lighting conditions.Our approach incorporates the Efficient Channel Attention (ECA) mechanism in visual odometry and the Selective Kernel Attention (SKA) mechanism in depth estimation to enhance pose accuracy and mitigate thermal depth degradation. Additionally, the system includes thermal depth-based loop closure detection and pose optimization, ensuring robust performance in low-texture thermal scenes. Extensive outdoor experiments demonstrate that DarkSLAM significantly outperforms existing methods like SC-Sfm-Learner and Shin et al., delivering precise localization and 3D dense mapping even in challenging nighttime environments.
Authors: Shuyi Liu, Simiao Cui, Haoran Bu, Yuming Shang, Xi Zhang
Abstract: Large language models (LLMs) have demonstrated remarkable capabilities across various applications, highlighting the urgent need for comprehensive safety evaluations. In particular, the enhanced Chinese language proficiency of LLMs, combined with the unique characteristics and complexity of Chinese expressions, has driven the emergence of Chinese-specific benchmarks for safety assessment. However, these benchmarks generally fall short in effectively exposing LLM safety vulnerabilities. To address the gap, we introduce JailBench, the first comprehensive Chinese benchmark for evaluating deep-seated vulnerabilities in LLMs, featuring a refined hierarchical safety taxonomy tailored to the Chinese context. To improve generation efficiency, we employ a novel Automatic Jailbreak Prompt Engineer (AJPE) framework for JailBench construction, which incorporates jailbreak techniques to enhance assessing effectiveness and leverages LLMs to automatically scale up the dataset through context-learning. The proposed JailBench is extensively evaluated over 13 mainstream LLMs and achieves the highest attack success rate against ChatGPT compared to existing Chinese benchmarks, underscoring its efficacy in identifying latent vulnerabilities in LLMs, as well as illustrating the substantial room for improvement in the security and trustworthiness of LLMs within the Chinese context. Our benchmark is publicly available at https://github.com/STAIR-BUPT/JailBench.
Authors: Jakub Macina, Nico Daheim, Ido Hakimi, Manu Kapur, Iryna Gurevych, Mrinmaya Sachan
Abstract: Evaluating the pedagogical capabilities of AI-based tutoring models is critical for making guided progress in the field. Yet, we lack a reliable, easy-to-use, and simple-to-run evaluation that reflects the pedagogical abilities of models. To fill this gap, we present MathTutorBench, an open-source benchmark for holistic tutoring model evaluation. MathTutorBench contains a collection of datasets and metrics that broadly cover tutor abilities as defined by learning sciences research in dialog-based teaching. To score the pedagogical quality of open-ended teacher responses, we train a reward model and show it can discriminate expert from novice teacher responses with high accuracy. We evaluate a wide set of closed- and open-weight models on MathTutorBench and find that subject expertise, indicated by solving ability, does not immediately translate to good teaching. Rather, pedagogy and subject expertise appear to form a trade-off that is navigated by the degree of tutoring specialization of the model. Furthermore, tutoring appears to become more challenging in longer dialogs, where simpler questioning strategies begin to fail. We release the benchmark, code, and leaderboard openly to enable rapid benchmarking of future models.
Authors: Lei Zhao, Sizhou Chen, Linfeng Feng, Xiao-Lei Zhang, Xuelong Li
Abstract: Text-to-audio (TTA), which generates audio signals from textual descriptions, has received huge attention in recent years. However, recent works focused on text to monaural audio only. As we know, spatial audio provides more immersive auditory experience than monaural audio, e.g. in virtual reality. To address this issue, we propose a text-to-spatial-audio (TTSA) generation framework named DualSpec.Specifically, it first trains variational autoencoders (VAEs) for extracting the latent acoustic representations from sound event audio. Then, given text that describes sound events and event directions, the proposed method uses the encoder of a pretrained large language model to transform the text into text features. Finally, it trains a diffusion model from the latent acoustic representations and text features for the spatial audio generation. In the inference stage, only the text description is needed to generate spatial audio. Particularly, to improve the synthesis quality and azimuth accuracy of the spatial sound events simultaneously, we propose to use two kinds of acoustic features. One is the Mel spectrograms which is good for improving the synthesis quality, and the other is the short-time Fourier transform spectrograms which is good at improving the azimuth accuracy. We provide a pipeline of constructing spatial audio dataset with text prompts, for the training of the VAEs and diffusion model. We also introduce new spatial-aware evaluation metrics to quantify the azimuth errors of the generated spatial audio recordings. Experimental results demonstrate that the proposed method can generate spatial audio with high directional and event consistency.
Authors: Margaret Li, Sneha Kudugunta, Luke Zettlemoyer
Abstract: Modern foundation models rely heavily on using scaling laws to guide crucial training decisions. Researchers often extrapolate the optimal architecture and hyper parameters settings from smaller training runs by describing the relationship between, loss, or task performance, and scale. All components of this process vary, from the specific equation being fit, to the training setup, to the optimization method. Each of these factors may affect the fitted law, and therefore, the conclusions of a given study. We discuss discrepancies in the conclusions that several prior works reach, on questions such as the optimal token to parameter ratio. We augment this discussion with our own analysis of the critical impact that changes in specific details may effect in a scaling study, and the resulting altered conclusions. Additionally, we survey over 50 papers that study scaling trends: while 45 of these papers quantify these trends using a power law, most under-report crucial details needed to reproduce their findings. To mitigate this, we we propose a checklist for authors to consider while contributing to scaling law research.
Authors: Hongyi Cal, ie Li, Wenzhen Dong
Abstract: The effectiveness of instruction fine-tuning for Large Language Models is fundamentally constrained by the quality and efficiency of training datasets. This work introduces Low-Confidence Gold (LCG), a novel filtering framework that employs centroid-based clustering and confidence-guided selection for identifying valuable instruction pairs. Through a semi-supervised approach using a lightweight classifier trained on representative samples, LCG curates high-quality subsets while preserving data diversity. Experimental evaluation demonstrates that models fine-tuned on LCG-filtered subsets of 6K samples achieve superior performance compared to existing methods, with substantial improvements on MT-bench and consistent gains across comprehensive evaluation metrics. The framework's efficacy while maintaining model performance establishes a promising direction for efficient instruction tuning.
Authors: Qiancheng Xu, Yongqi Li, Heming Xia, Fan Liu, Min Yang, Wenjie Li
Abstract: Tool learning has emerged as a promising direction by extending Large Language Models' (LLMs) capabilities with external tools. Existing tool learning studies primarily focus on the general-purpose tool-use capability, which addresses explicit user requirements in instructions. However, they overlook the importance of personalized tool-use capability, leading to an inability to handle implicit user preferences. To address the limitation, we first formulate the task of personalized tool learning, which integrates user's interaction history towards personalized tool usage. To fill the gap of missing benchmarks, we construct PEToolBench, featuring diverse user preferences reflected in interaction history under three distinct personalized settings, and encompassing a wide range of tool-use scenarios. Moreover, we propose a framework PEToolLLaMA to adapt LLMs to the personalized tool learning task, which is trained through supervised fine-tuning and direct preference optimization. Extensive experiments on PEToolBench demonstrate the superiority of PEToolLLaMA over existing LLMs.
Authors: Jinbo Wang, Mingze Wang, Zhanpeng Zhou, Junchi Yan, Weinan E, Lei Wu
Abstract: Transformers consist of diverse building blocks, such as embedding layers, normalization layers, self-attention mechanisms, and point-wise feedforward networks. Thus, understanding the differences and interactions among these blocks is important. In this paper, we uncover a clear Sharpness Disparity across these blocks, which emerges early in training and intriguingly persists throughout the training process. Motivated by this finding, we propose Blockwise Learning Rate (LR), a strategy that tailors the LR to each block's sharpness, accelerating large language model (LLM) pre-training. By integrating Blockwise LR into AdamW, we consistently achieve lower terminal loss and nearly $2\times$ speedup compared to vanilla AdamW. We demonstrate this acceleration across GPT-2 and LLaMA, with model sizes ranging from 0.12B to 1.1B and datasets of OpenWebText and MiniPile. Finally, we incorporate Blockwise LR into Adam-mini (Zhang et al., 2024), a recently proposed memory-efficient variant of Adam, achieving a combined $2\times$ speedup and $2\times$ memory saving. These results underscore the potential of exploiting the sharpness disparity to improve LLM training.
Authors: Nahom Abishu Hayla, A. Mohammed Seid, Aiman Erbad, Tilahun M. Getu, Ala Al-Fuqaha, Mohsen Guizani
Abstract: Although multi-tier vehicular Metaverse promises to transform vehicles into essential nodes -- within an interconnected digital ecosystem -- using efficient resource allocation and seamless vehicular twin (VT) migration, this can hardly be achieved by the existing techniques operating in a highly dynamic vehicular environment, since they can hardly balance multi-objective optimization problems such as latency reduction, resource utilization, and user experience (UX). To address these challenges, we introduce a novel multi-tier resource allocation and VT migration framework that integrates Graph Convolutional Networks (GCNs), a hierarchical Stackelberg game-based incentive mechanism, and Multi-Agent Deep Reinforcement Learning (MADRL). The GCN-based model captures both spatial and temporal dependencies within the vehicular network; the Stackelberg game-based incentive mechanism fosters cooperation between vehicles and infrastructure; and the MADRL algorithm jointly optimizes resource allocation and VT migration in real time. By modeling this dynamic and multi-tier vehicular Metaverse as a Markov Decision Process (MDP), we develop a MADRL-based algorithm dubbed the Multi-Objective Multi-Agent Deep Deterministic Policy Gradient (MO-MADDPG), which can effectively balances the various conflicting objectives. Extensive simulations validate the effectiveness of this algorithm that is demonstrated to enhance scalability, reliability, and efficiency while considerably improving latency, resource utilization, migration cost, and overall UX by 12.8%, 9.7%, 14.2%, and 16.1%, respectively.
Authors: Liangdong Liu, Zhitong Zheng, Cong Wang, Tianhuang Su, Zhenyu Yang
Abstract: Large language models (LLMs) have wide applications in the field of natural language processing(NLP), such as GPT-4 and Llama. However, with the exponential growth of model parameter sizes, LLMs bring significant resource overheads. Low-bit quantization, as a key technique, reduces memory usage and computational demands by decreasing the bit-width of model parameters, activations, and gradients. Previous quantization methods for LLMs have largely employed Post-Training Quantization (PTQ) and Quantization-Aware Training (QAT). PTQ does not require any retraining of the original model, while QAT involves optimizing precision during training to achieve the best quantization parameters. The BitNet team proposed a radically different approach, where quantization is performed from the start of model training, utilizing low-precision binary weights during the training process. This approach has led to the emergence of many binary quantization techniques for large language models. This paper provides a comprehensive review of these binary quantization techniques. Specifically, we will introduce binary quantization techniques in deep neural networks and further explore their application to LLMs, reviewing their various contributions, implementations, and applications.
Authors: Jaehyeon Son, Soochan Lee, Gunhee Kim
Abstract: Recent studies have shown that Transformers can perform in-context reinforcement learning (RL) by imitating existing RL algorithms, enabling sample-efficient adaptation to unseen tasks without parameter updates. However, these models also inherit the suboptimal behaviors of the RL algorithms they imitate. This issue primarily arises due to the gradual update rule employed by those algorithms. Model-based planning offers a promising solution to this limitation by allowing the models to simulate potential outcomes before taking action, providing an additional mechanism to deviate from the suboptimal behavior. Rather than learning a separate dynamics model, we propose Distillation for In-Context Planning (DICP), an in-context model-based RL framework where Transformers simultaneously learn environment dynamics and improve policy in-context. We evaluate DICP across a range of discrete and continuous environments, including Darkroom variants and Meta-World. Our results show that DICP achieves state-of-the-art performance while requiring significantly fewer environment interactions than baselines, which include both model-free counterparts and existing meta-RL methods.
Authors: Marc Martinez-Gost, Ana P\'erez-Neira, Miguel \'Angel Lagunas
Abstract: This paper utilizes the properties of type-based multiple access (TBMA) to investigate its effectiveness as a robust approach for over-the-air computation (AirComp) in the presence of Byzantine attacks, this is, adversarial strategies where malicious nodes intentionally distort their transmissions to corrupt the aggregated result. Unlike classical direct aggregation (DA) AirComp, which aggregates data in the amplitude of the signals and are highly vulnerable to attacks, TBMA distributes data over multiple radio resources, enabling the receiver to construct a histogram representation of the transmitted data. This structure allows the integration of classical robust estimators and supports the computation of diverse functions beyond the arithmetic mean, which is not feasible with DA. Through extensive simulations, we demonstrate that robust TBMA significantly outperforms DA, maintaining high accuracy even under adversarial conditions, and showcases its applicability in federated learning (FEEL) scenarios. Additionally, TBMA reduces channel state information (CSI) requirements, lowers energy consumption, and enhances resiliency by leveraging the diversity of the transmitted data. These results establish TBMA as a scalable and robust solution for AirComp, paving the way for secure and efficient aggregation in next-generation networks.
Authors: Zerui Li, Gengze Zhou, Haodong Hong, Yanyan Shao, Wenqi Lyu, Yanyuan Qiao, Qi Wu
Abstract: Vision-and-Language Navigation (VLN) empowers agents to associate time-sequenced visual observations with corresponding instructions to make sequential decisions. However, generalization remains a persistent challenge, particularly when dealing with visually diverse scenes or transitioning from simulated environments to real-world deployment. In this paper, we address the mismatch between human-centric instructions and quadruped robots with a low-height field of view, proposing a Ground-level Viewpoint Navigation (GVNav) approach to mitigate this issue. This work represents the first attempt to highlight the generalization gap in VLN across varying heights of visual observation in realistic robot deployments. Our approach leverages weighted historical observations as enriched spatiotemporal contexts for instruction following, effectively managing feature collisions within cells by assigning appropriate weights to identical features across different viewpoints. This enables low-height robots to overcome challenges such as visual obstructions and perceptual mismatches. Additionally, we transfer the connectivity graph from the HM3D and Gibson datasets as an extra resource to enhance spatial priors and a more comprehensive representation of real-world scenarios, leading to improved performance and generalizability of the waypoint predictor in real-world environments. Extensive experiments demonstrate that our Ground-level Viewpoint Navigation (GVnav) approach significantly improves performance in both simulated environments and real-world deployments with quadruped robots.
Authors: Fengbin Guan, Zihao Yu, Yiting Lu, Xin Li, Zhibo Chen
Abstract: Video quality assessment tasks rely heavily on the rich features required for video understanding, such as semantic information, texture, and temporal motion. The existing video foundational model, InternVideo2, has demonstrated strong potential in video understanding tasks due to its large parameter size and large-scale multimodal data pertaining. Building on this, we explored the transferability of InternVideo2 to video quality assessment under compression scenarios. To design a lightweight model suitable for this task, we proposed a distillation method to equip the smaller model with rich compression quality priors. Additionally, we examined the performance of different backbones during the distillation process. The results showed that, compared to other methods, our lightweight model distilled from InternVideo2 achieved excellent performance in compression video quality assessment.
Authors: Samuele Pasini, Gianluca Maragliano, Jinhan Kim, Paolo Tonella
Abstract: Cross-site scripting (XSS) poses a significant threat to web application security. While Deep Learning (DL) has shown remarkable success in detecting XSS attacks, it remains vulnerable to adversarial attacks due to the discontinuous nature of its input-output mapping. These adversarial attacks employ mutation-based strategies for different components of XSS attack vectors, allowing adversarial agents to iteratively select mutations to evade detection. Our work replicates a state-of-the-art XSS adversarial attack, highlighting threats to validity in the reference work and extending it toward a more effective evaluation strategy. Moreover, we introduce an XSS Oracle to mitigate these threats. The experimental results show that our approach achieves an escape rate above 96% when the threats to validity of the replicated technique are addressed.
Authors: Dejan Grba
Abstract: Generative artificial intelligence (generative AI) has entered the mainstream culture and become a subject of extensive academic investigation. However, the character and background of its impact on art require subtler scrutiny and more nuanced contextualization. This paper summarizes a broader study of the roles that AI's conceptual and ideological substrata play in influencing art notions. The focus is on divergent but coalescing and often questionable ideas, values, and political views that generative AI and other art-related AI technologies propagate from the computer science and AI/tech industry to the contemporary art and culture. The paper maps the main areas of this complex relationship and concisely critiques their key aspects.
Authors: Bojana Ba\v{s}aragin, Darija Medvecki, Gorana Goji\'c, Milena Oparnica, Dragi\v{s}a Mi\v{s}kovi\'c
Abstract: This study introduces a novel Natural Language Processing pipeline that enhances customer service efficiency at Telekom Srbija, a leading Serbian telecommunications company, through automated email topic detection and labelling. Central to the pipeline is BERTopic, a modular architecture that allows unsupervised topic modelling. After a series of preprocessing and post-processing steps, we assign one of 12 topics and several additional labels to incoming emails, allowing customer service to filter and access them through a custom-made application. The model's performance was evaluated by assessing the speed and correctness of the automatically assigned topics across a test dataset of 100 customer emails. The pipeline shows broad applicability across languages, particularly for those that are low-resourced and morphologically rich. The system now operates in the company's production environment, streamlining customer service operations through automated email classification.
Authors: Guikun Chen, Xu Zhang, Yi Yang, Wenguan Wang
Abstract: Chemical reaction data is a pivotal asset, driving advances in competitive fields such as pharmaceuticals, materials science, and industrial chemistry. Its proprietary nature renders it sensitive, as it often includes confidential insights and competitive advantages organizations strive to protect. However, in contrast to this need for confidentiality, the current standard training paradigm for machine learning-based retrosynthesis gathers reaction data from multiple sources into one single edge to train prediction models. This paradigm poses considerable privacy risks as it necessitates broad data availability across organizational boundaries and frequent data transmission between entities, potentially exposing proprietary information to unauthorized access or interception during storage and transfer. In the present study, we introduce the chemical knowledge-informed framework (CKIF), a privacy-preserving approach for learning retrosynthesis models. CKIF enables distributed training across multiple chemical organizations without compromising the confidentiality of proprietary reaction data. Instead of gathering raw reaction data, CKIF learns retrosynthesis models through iterative, chemical knowledge-informed aggregation of model parameters. In particular, the chemical properties of predicted reactants are leveraged to quantitatively assess the observable behaviors of individual models, which in turn determines the adaptive weights used for model aggregation. On a variety of reaction datasets, CKIF outperforms several strong baselines by a clear margin (e.g., ~20% performance improvement over FedAvg on USPTO-50K), showing its feasibility and superiority to stimulate further research on privacy-preserving retrosynthesis.
Authors: Behnaz Elhaminia, Abdullah Alsalemi, Esha Nasir, Mostafa Jahanifar, Ruqayya Awan, Lawrence S. Young, Nasir M. Rajpoot, Fayyaz Minhas, Shan E Ahmed Raza
Abstract: Whole slide image (WSI) registration is an essential task for analysing the tumour microenvironment (TME) in histopathology. It involves the alignment of spatial information between WSIs of the same section or serial sections of a tissue sample. The tissue sections are usually stained with single or multiple biomarkers before imaging, and the goal is to identify neighbouring nuclei along the Z-axis for creating a 3D image or identifying subclasses of cells in the TME. This task is considerably more challenging compared to radiology image registration, such as magnetic resonance imaging or computed tomography, due to various factors. These include gigapixel size of images, variations in appearance between differently stained tissues, changes in structure and morphology between non-consecutive sections, and the presence of artefacts, tears, and deformations. Currently, there is a noticeable gap in the literature regarding a review of the current approaches and their limitations, as well as the challenges and opportunities they present. We aim to provide a comprehensive understanding of the available approaches and their application for various purposes. Furthermore, we investigate current deep learning methods used for WSI registration, emphasising their diverse methodologies. We examine the available datasets and explore tools and software employed in the field. Finally, we identify open challenges and potential future trends in this area of research.
Authors: Lars Benedikt Kaesberg, Jonas Becker, Jan Philip Wahle, Terry Ruas, Bela Gipp
Abstract: Much of the success of multi-agent debates depends on carefully choosing the right parameters. Among them, the decision-making protocol stands out. Systematic comparison of decision protocols is difficult because studies alter multiple discussion parameters beyond the protocol. So far, it has been largely unknown how decision-making addresses the challenges of different tasks. This work systematically evaluates the impact of seven decision protocols (e.g., majority voting, unanimity consensus). We change only one variable at a time (i.e., decision protocol) to analyze how different methods affect the collaboration between agents and test different protocols on knowledge (MMLU, MMLU-Pro, GPQA) and reasoning datasets (StrategyQA, MuSR, SQuAD 2.0). Our results show that voting protocols improve performance by 13.2% in reasoning tasks and consensus protocols by 2.8% in knowledge tasks over the other decision protocol. Increasing the number of agents improves performance, while more discussion rounds before voting reduces it. To improve decision-making by increasing answer diversity, we propose two new methods, All-Agents Drafting (AAD) and Collective Improvement (CI). Our methods improve task performance by up to 3.3% with AAD and up to 7.4% with CI. This work demonstrates the importance of decision-making in multi-agent debates beyond scaling.
Authors: Yijiang River Dong, Tiancheng Hu, Yinhong Liu, Ahmet \"Ust\"un, Nigel Collier
Abstract: While Reinforcement Learning from Human Feedback (RLHF) is widely used to align Large Language Models (LLMs) with human preferences, it typically assumes homogeneous preferences across users, overlooking diverse human values and minority viewpoints. Although personalized preference learning addresses this by tailoring separate preferences for individual users, the field lacks standardized methods to assess its effectiveness. We present a multi-faceted evaluation framework that measures not only performance but also fairness, unintended effects, and adaptability across varying levels of preference divergence. Through extensive experiments comparing eight personalization methods across three preference datasets, we demonstrate that performance differences between methods could reach 36% when users strongly disagree, and personalization can introduce up to 20% safety misalignment. These findings highlight the critical need for holistic evaluation approaches to advance the development of more effective and inclusive preference learning systems.
Authors: Rebekka G\"orge, Michael Mock, H\'ector Allende-Cid
Abstract: Social categories and stereotypes are embedded in language and can introduce data bias into Large Language Models (LLMs). Despite safeguards, these biases often persist in model behavior, potentially leading to representational harm in outputs. While sociolinguistic research provides valuable insights into the formation of stereotypes, NLP approaches for stereotype detection rarely draw on this foundation and often lack objectivity, precision, and interpretability. To fill this gap, in this work we propose a new approach that detects and quantifies the linguistic indicators of stereotypes in a sentence. We derive linguistic indicators from the Social Category and Stereotype Communication (SCSC) framework which indicate strong social category formulation and stereotyping in language, and use them to build a categorization scheme. To automate this approach, we instruct different LLMs using in-context learning to apply the approach to a sentence, where the LLM examines the linguistic properties and provides a basis for a fine-grained assessment. Based on an empirical evaluation of the importance of different linguistic indicators, we learn a scoring function that measures the linguistic indicators of a stereotype. Our annotations of stereotyped sentences show that these indicators are present in these sentences and explain the strength of a stereotype. In terms of model performance, our results show that the models generally perform well in detecting and classifying linguistic indicators of category labels used to denote a category, but sometimes struggle to correctly evaluate the associated behaviors and characteristics. Using more few-shot examples within the prompts, significantly improves performance. Model performance increases with size, as Llama-3.3-70B-Instruct and GPT-4 achieve comparable results that surpass those of Mixtral-8x7B-Instruct, GPT-4-mini and Llama-3.1-8B-Instruct.
Authors: Henry Peng Zou, Zhengyao Gu, Yue Zhou, Yankai Chen, Weizhi Zhang, Liancheng Fang, Yibo Wang, Yangning Li, Kay Liu, Philip S. Yu
Abstract: Test-time computing approaches, which leverage additional computational resources during inference, have been proven effective in enhancing large language model performance. This work introduces a novel, linearly scaling approach, TestNUC, that improves test-time predictions by leveraging the local consistency of neighboring unlabeled data-it classifies an input instance by considering not only the model's prediction on that instance but also on neighboring unlabeled instances. We evaluate TestNUC across eight diverse datasets, spanning intent classification, topic mining, domain discovery, and emotion detection, demonstrating its consistent superiority over baseline methods such as standard prompting and self-consistency. Furthermore, TestNUC can be seamlessly integrated with existing test-time computing approaches, substantially boosting their performance. Our analysis reveals that TestNUC scales effectively with increasing amounts of unlabeled data and performs robustly across different embedding models, making it practical for real-world applications. Our code is available at https://github.com/HenryPengZou/TestNUC.
Authors: Daniel Rose, Chia-Chien Hung, Marco Lepri, Israa Alqassem, Kiril Gashteovski, Carolin Lawrence
Abstract: Differential Diagnosis (DDx) is a fundamental yet complex aspect of clinical decision-making, in which physicians iteratively refine a ranked list of possible diseases based on symptoms, antecedents, and medical knowledge. While recent advances in large language models have shown promise in supporting DDx, existing approaches face key limitations, including single-dataset evaluations, isolated optimization of components, unrealistic assumptions about complete patient profiles, and single-attempt diagnosis. We introduce a Modular Explainable DDx Agent (MEDDxAgent) framework designed for interactive DDx, where diagnostic reasoning evolves through iterative learning, rather than assuming a complete patient profile is accessible. MEDDxAgent integrates three modular components: (1) an orchestrator (DDxDriver), (2) a history taking simulator, and (3) two specialized agents for knowledge retrieval and diagnosis strategy. To ensure robust evaluation, we introduce a comprehensive DDx benchmark covering respiratory, skin, and rare diseases. We analyze single-turn diagnostic approaches and demonstrate the importance of iterative refinement when patient profiles are not available at the outset. Our broad evaluation demonstrates that MEDDxAgent achieves over 10% accuracy improvements in interactive DDx across both large and small LLMs, while offering critical explainability into its diagnostic reasoning process.
Authors: Melanie Schaller, Mathis Kruse, Antonio Ortega, Marius Lindauer, Bodo Rosenhahn
Abstract: Addressing sensor drift is essential in industrial measurement systems, where precise data output is necessary for maintaining accuracy and reliability in monitoring processes, as it progressively degrades the performance of machine learning models over time. Our findings indicate that the standard cross-validation method used in existing model training overestimates performance by inadequately accounting for drift. This is primarily because typical cross-validation techniques allow data instances to appear in both training and testing sets, thereby distorting the accuracy of the predictive evaluation. As a result, these models are unable to precisely predict future drift effects, compromising their ability to generalize and adapt to evolving data conditions. This paper presents two solutions: (1) a novel sensor drift compensation learning paradigm for validating models, and (2) automated machine learning (AutoML) techniques to enhance classification performance and compensate sensor drift. By employing strategies such as data balancing, meta-learning, automated ensemble learning, hyperparameter optimization, feature selection, and boosting, our AutoML-DC (Drift Compensation) model significantly improves classification performance against sensor drift. AutoML-DC further adapts effectively to varying drift severities.
Authors: Lauren Klein, Meredith Martin, Andr\'e Brock, Maria Antoniak, Melanie Walsh, Jessica Marie Johnson, Lauren Tilton, David Mimno
Abstract: This paper presents a set of provocations for considering the uses, impact, and harms of generative AI from the perspective of humanities researchers. We provide a working definition of humanities research, summarize some of its most salient theories and methods, and apply these theories and methods to the current landscape of AI. Drawing from foundational work in critical data studies, along with relevant humanities scholarship, we elaborate eight claims with broad applicability to current conversations about generative AI: 1) Models make words, but people make meaning; 2) Generative AI requires an expanded definition of culture; 3) Generative AI can never be representative; 4) Bigger models are not always better models; 5) Not all training data is equivalent; 6) Openness is not an easy fix; 7) Limited access to compute enables corporate capture; and 8) AI universalism creates narrow human subjects. We conclude with a discussion of the importance of resisting the extraction of humanities research by computer science and related fields.
Authors: Jinyu Cai, Yusei Ishimizu, Mingyue Zhang, Munan Li, Jialong Li, Kenji Tei
Abstract: Social media platforms frequently impose restrictive policies to moderate user content, prompting the emergence of creative evasion language strategies. This paper presents a multi-agent framework based on Large Language Models (LLMs) to simulate the iterative evolution of language strategies under regulatory constraints. In this framework, participant agents, as social media users, continuously evolve their language expression, while supervisory agents emulate platform-level regulation by assessing policy violations. To achieve a more faithful simulation, we employ a dual design of language strategies (constraint and expression) to differentiate conflicting goals and utilize an LLM-driven GA (Genetic Algorithm) for the selection, mutation, and crossover of language strategies. The framework is evaluated using two distinct scenarios: an abstract password game and a realistic simulated illegal pet trade scenario. Experimental results demonstrate that as the number of dialogue rounds increases, both the number of uninterrupted dialogue turns and the accuracy of information transmission improve significantly. Furthermore, a user study with 40 participants validates the real-world relevance of the generated dialogues and strategies. Moreover, ablation studies validate the importance of the GA, emphasizing its contribution to long-term adaptability and improved overall results.
Authors: Linshan Jia
Abstract: Feature extraction is crucial in intelligent fault diagnosis of rotating machinery. It is easier for convolutional neural networks(CNNs) to visually recognize and learn fault features by converting the complicated one-dimensional (1D) vibrational signals into two-dimensional (2D) images with simple textures. However, the existing representation methods for encoding 1D signals as images have two main problems, including complicated computation and low separability. Meanwhile, the existing 2D-CNN fault diagnosis methods taking 2D images as the only inputs still suffer from the inevitable information loss because of the conversion process. Considering the above issues, this paper proposes a new 1D-to-2D conversion method called Embedding Gramian Representation (EGR), which is easy to calculate and shows good separability. In EGR, 1D signals are projected in the embedding space and the intrinsic periodicity of vibrational signals is captured enabling the faulty characteristics contained in raw signals to be uncovered. Second, aiming at the information loss problem of existing CNN models with the single input of converted images, a double-branch EGR-based CNN, called EGR-Net, is proposed to learn faulty features from both raw signal feature maps and their corresponding EGRs. The bridge connection is designed to improve the feature learning interaction between the two branches. Widely used open domain gearbox dataset and bearing dataset are used to verify the effectiveness and efficiency of the proposed methods. EGR-Net is compared with traditional and state-of-the-art approaches, and the results show that the proposed method can deliver enhanced performance.
Authors: Nakyeong Yang, Minsung Kim, Seunghyun Yoon, Joongbo Shin, Kyomin Jung
Abstract: Various studies have attempted to remove sensitive or private knowledge from a language model to prevent its unauthorized exposure. However, prior studies have overlooked the complex and interconnected nature of knowledge, where related knowledge must be carefully examined. Specifically, they have failed to evaluate whether an unlearning method faithfully erases interconnected knowledge that should be removed, retaining knowledge that appears relevant but exists in a completely different context. To resolve this problem, we first define a new concept called superficial unlearning, which refers to the phenomenon where an unlearning method either fails to erase the interconnected knowledge it should remove or unintentionally erases irrelevant knowledge. Based on the definition, we introduce a new benchmark, FaithUn, to analyze and evaluate the faithfulness of unlearning in real-world knowledge QA settings. Furthermore, we propose a novel unlearning method, KLUE, which updates only knowledge-related neurons to achieve faithful unlearning. KLUE identifies knowledge neurons using an explainability method and updates only those neurons using selected unforgotten samples. Experimental results demonstrate that widely-used unlearning methods fail to ensure faithful unlearning, while our method shows significant effectiveness in real-world QA unlearning.
Authors: Nikita Shvetsov, Thomas K. Kilvaer, Masoud Tafavvoghi, Anders Sildnes, Kajsa M{\o}llersen, Lill-Tove Rasmussen Busund, Lars Ailo Bongo
Abstract: Developing clinically useful cell-level analysis tools in digital pathology remains challenging due to limitations in dataset granularity, inconsistent annotations, high computational demands, and difficulties integrating new technologies into workflows. To address these issues, we propose a solution that enhances data quality, model performance, and usability by creating a lightweight, extensible cell segmentation and classification model. First, we update data labels through cross-relabeling to refine annotations of PanNuke and MoNuSAC, producing a unified dataset with seven distinct cell types. Second, we leverage the H-Optimus foundation model as a fixed encoder to improve feature representation for simultaneous segmentation and classification tasks. Third, to address foundation models' computational demands, we distill knowledge to reduce model size and complexity while maintaining comparable performance. Finally, we integrate the distilled model into QuPath, a widely used open-source digital pathology platform. Results demonstrate improved segmentation and classification performance using the H-Optimus-based model compared to a CNN-based model. Specifically, average $R^2$ improved from 0.575 to 0.871, and average $PQ$ score improved from 0.450 to 0.492, indicating better alignment with actual cell counts and enhanced segmentation quality. The distilled model maintains comparable performance while reducing parameter count by a factor of 48. By reducing computational complexity and integrating into workflows, this approach may significantly impact diagnostics, reduce pathologist workload, and improve outcomes. Although the method shows promise, extensive validation is necessary prior to clinical deployment.
Authors: Yunyang Li, Zaishuo Xia, Lin Huang, Xinran Wei, Han Yang, Sam Harshe, Zun Wang, Chang Liu, Jia Zhang, Bin Shao, Mark B. Gerstein
Abstract: Density Functional Theory (DFT) is a pivotal method within quantum chemistry and materials science, with its core involving the construction and solution of the Kohn-Sham Hamiltonian. Despite its importance, the application of DFT is frequently limited by the substantial computational resources required to construct the Kohn-Sham Hamiltonian. In response to these limitations, current research has employed deep-learning models to efficiently predict molecular and solid Hamiltonians, with roto-translational symmetries encoded in their neural networks. However, the scalability of prior models may be problematic when applied to large molecules, resulting in non-physical predictions of ground-state properties. In this study, we generate a substantially larger training set (PubChemQH) than used previously and use it to create a scalable model for DFT calculations with physical accuracy. For our model, we introduce a loss function derived from physical principles, which we call Wavefunction Alignment Loss (WALoss). WALoss involves performing a basis change on the predicted Hamiltonian to align it with the observed one; thus, the resulting differences can serve as a surrogate for orbital energy differences, allowing models to make better predictions for molecular orbitals and total energies than previously possible. WALoss also substantially accelerates self-consistent-field (SCF) DFT calculations. Here, we show it achieves a reduction in total energy prediction error by a factor of 1347 and an SCF calculation speed-up by a factor of 18%. These substantial improvements set new benchmarks for achieving accurate and applicable predictions in larger molecular systems.
Authors: Veronika Ro\v{c}kov\'a, Sean O'Hagan
Abstract: The advent of Generative Artificial Intelligence (GAI) has heralded an inflection point that changed how society thinks about knowledge acquisition. While GAI cannot be fully trusted for decision-making, it may still provide valuable information that can be integrated into a decision pipeline. Rather than seeing the lack of certitude and inherent randomness of GAI as a problem, we view it as an opportunity. Indeed, variable answers to given prompts can be leveraged to construct a prior distribution which reflects assuredness of AI predictions. This prior distribution may be combined with tailored datasets for a fully Bayesian analysis with an AI-driven prior. In this paper, we explore such a possibility within a non-parametric Bayesian framework. The basic idea consists of assigning a Dirichlet process prior distribution on the data-generating distribution with AI generative model as its baseline. Hyper-parameters of the prior can be tuned out-of-sample to assess the informativeness of the AI prior. Posterior simulation is achieved by computing a suitably randomized functional on an augmented data that consists of observed (labeled) data as well as fake data whose labels have been imputed using AI. This strategy can be parallelized and rapidly produces iid samples from the posterior by optimization as opposed to sampling from conditionals. Our method enables (predictive) inference and uncertainty quantification leveraging AI predictions in a coherent probabilistic manner.
Authors: Michael Y. Hu, Jackson Petty, Chuan Shi, William Merrill, Tal Linzen
Abstract: Pretraining language models on formal languages can improve their acquisition of natural language, but it is unclear which features of the formal language impart an inductive bias that leads to effective transfer. Drawing on insights from linguistics and complexity theory, we hypothesize that effective transfer occurs when the formal language both captures dependency structures in natural language and remains within the computational limitations of the model architecture. Focusing on transformers, we find that formal languages with both these properties enable language models to achieve lower loss on natural language and better linguistic generalization compared to other languages. In fact, pre-pretraining, or training on formal-then-natural language, reduces loss more efficiently than the same amount of natural language. For a 1B-parameter language model trained on roughly 1.6B tokens of natural language, pre-pretraining achieves the same loss and better linguistic generalization with a 33% smaller token budget. We also give mechanistic evidence of cross-task transfer from formal to natural language: attention heads acquired during formal language pretraining remain crucial for the model's performance on syntactic evaluations.
Authors: Li Ju, Xingyi Yang, Qi Li, Xinchao Wang
Abstract: Graph neural networks (GNNs) are conventionally trained on a per-domain, per-task basis. It creates a significant barrier in transferring the acquired knowledge to different, heterogeneous data setups. This paper introduces GraphBridge, a novel framework to enable knowledge transfer across disparate tasks and domains in GNNs, circumventing the need for modifications to task configurations or graph structures. Specifically, GraphBridge allows for the augmentation of any pre-trained GNN with prediction heads and a bridging network that connects the input to the output layer. This architecture not only preserves the intrinsic knowledge of the original model but also supports outputs of arbitrary dimensions. To mitigate the negative transfer problem, GraphBridg merges the source model with a concurrently trained model, thereby reducing the source bias when applied to the target domain. Our method is thoroughly evaluated across diverse transfer learning scenarios, including Graph2Graph, Node2Node, Graph2Node, and graph2point-cloud. Empirical validation, conducted over 16 datasets representative of these scenarios, confirms the framework's capacity for task- and domain-agnostic transfer learning within graph-like data, marking a significant advancement in the field of GNNs.
Authors: Jiawei Huang, Bingcong Li, Christoph Dann, Niao He
Abstract: Sample efficiency is critical for online Reinforcement Learning from Human Feedback (RLHF). While existing works investigate sample-efficient online exploration strategies, the potential of utilizing misspecified yet relevant reward models to accelerate learning remains underexplored. This paper studies how to transfer knowledge from those imperfect reward models in online RLHF. We start by identifying a novel property of the KL-regularized RLHF objective: \emph{a policy's ability to cover the optimal policy is captured by its sub-optimality}. Building on this insight, we propose a theoretical transfer learning algorithm with provable benefits compared to standard online learning. Our approach achieves low regret in the early stage by quickly adapting to the best available source reward models without prior knowledge of their quality, and over time, it attains an $\tilde{O}(\sqrt{T})$ regret bound \emph{independent} of structural complexity measures. Inspired by our theoretical findings, we develop an empirical algorithm with improved computational efficiency, and demonstrate its effectiveness empirically in summarization tasks.
Authors: Zhiqiang Wang, Haoyu Wang, Lu Hao
Abstract: Webshell is a type of backdoor, and web applications are widely exposed to webshell injection attacks. Therefore, it is important to study webshell detection techniques. In this study, we propose a webshell detection method. We first convert PHP source code to opcodes and then extract Opcode Double-Tuples (ODTs). Next, we combine CodeBert and FastText models for feature representation and classification. To address the challenge that deep learning methods have difficulty detecting long webshell files, we introduce a sliding window attention mechanism. This approach effectively captures malicious behavior within long files. Experimental results show that our method reaches high accuracy in webshell detection, solving the problem of traditional methods that struggle to address new webshell variants and anti-detection techniques.
Authors: Nadya Abdel Madjid, Murad Mebrahtu, Abdelmoamen Nasser, Bilal Hassan, Naoufel Werghi, Jorge Dias, Majid Khonji
Abstract: This paper introduces the Emirates Multi-Task (EMT) dataset - the first publicly available dataset for autonomous driving collected in the Arab Gulf region. The EMT dataset captures the unique road topology, high traffic congestion, and distinctive characteristics of the Gulf region, including variations in pedestrian clothing and weather conditions. It contains over 30,000 frames from a dash-camera perspective, along with 570,000 annotated bounding boxes, covering approximately 150 kilometers of driving routes. The EMT dataset supports three primary tasks: tracking, trajectory forecasting and intention prediction. Each benchmark dataset is complemented with corresponding evaluations: (1) multi-agent tracking experiments, focusing on multi-class scenarios and occlusion handling; (2) trajectory forecasting evaluation using deep sequential and interaction-aware models; and (3) intention benchmark experiments conducted for predicting agents intentions from observed trajectories. The dataset is publicly available at https://avlab.io/emt-dataset, and pre-processing scripts along with evaluation models can be accessed at https://github.com/AV-Lab/emt-dataset.
URLs: https://avlab.io/emt-dataset,, https://github.com/AV-Lab/emt-dataset.
Authors: Taishi Nakamura, Takuya Akiba, Kazuki Fujii, Yusuke Oda, Rio Yokota, Jun Suzuki
Abstract: The Mixture of Experts (MoE) architecture reduces the training and inference cost significantly compared to a dense model of equivalent capacity. Upcycling is an approach that initializes and trains an MoE model using a pre-trained dense model. While upcycling leads to initial performance gains, the training progresses slower than when trained from scratch, leading to suboptimal performance in the long term. We propose Drop-Upcycling - a method that effectively addresses this problem. Drop-Upcycling combines two seemingly contradictory approaches: utilizing the knowledge of pre-trained dense models while statistically re-initializing some parts of the weights. This approach strategically promotes expert specialization, significantly enhancing the MoE model's efficiency in knowledge acquisition. Extensive large-scale experiments demonstrate that Drop-Upcycling significantly outperforms previous MoE construction methods in the long term, specifically when training on hundreds of billions of tokens or more. As a result, our MoE model with 5.9B active parameters achieves comparable performance to a 13B dense model in the same model family, while requiring approximately 1/4 of the training FLOPs. All experimental resources, including source code, training data, model checkpoints and logs, are publicly available to promote reproducibility and future research on MoE.
Authors: Saman Forouzandeh, Pavel N. Krivitsky, Rohitash Chandra
Abstract: Recommender systems leveraging deep learning models have been crucial for assisting users in selecting items aligned with their preferences and interests. However, a significant challenge persists in single-criteria recommender systems, which often overlook the diverse attributes of items that have been addressed by Multi-Criteria Recommender Systems (MCRS). Shared embedding vector for multi-criteria item ratings but have struggled to capture the nuanced relationships between users and items based on specific criteria. In this study, we present a novel representation for Multi-Criteria Recommender Systems (MCRS) based on a multi-edge bipartite graph, where each edge represents one criterion rating of items by users, and Multiview Dual Graph Attention Networks (MDGAT). Employing MDGAT is beneficial and important for adequately considering all relations between users and items, given the presence of both local (criterion-based) and global (multi-criteria) relations. Additionally, we define anchor points in each view based on similarity and employ local and global contrastive learning to distinguish between positive and negative samples across each view and the entire graph. We evaluate our method on two real-world datasets and assess its performance based on item rating predictions. The results demonstrate that our method achieves higher accuracy compared to the baseline method for predicting item ratings on the same datasets. MDGAT effectively capture the local and global impact of neighbours and the similarity between nodes.
Authors: Jiyuan Wang, Weishan Ye, Jialin He, Li Zhang, Gan Huang, Zhuliang Yu, Zhen Liang
Abstract: With the rapid advancement of deep learning, attention mechanisms have become indispensable in electroencephalography (EEG) signal analysis, significantly enhancing Brain-Computer Interface (BCI) applications. This paper presents a comprehensive review of traditional and Transformer-based attention mechanisms, their embedding strategies, and their applications in EEG-based BCI, with a particular emphasis on multimodal data fusion. By capturing EEG variations across time, frequency, and spatial channels, attention mechanisms improve feature extraction, representation learning, and model robustness. These methods can be broadly categorized into traditional attention mechanisms, which typically integrate with convolutional and recurrent networks, and Transformer-based multi-head self-attention, which excels in capturing long-range dependencies. Beyond single-modality analysis, attention mechanisms also enhance multimodal EEG applications, facilitating effective fusion between EEG and other physiological or sensory data. Finally, we discuss existing challenges and emerging trends in attention-based EEG modeling, highlighting future directions for advancing BCI technology. This review aims to provide valuable insights for researchers seeking to leverage attention mechanisms for improved EEG interpretation and application.
Authors: Nikhilesh Prabhakar, Ranveer Singh, Harsha Kokel, Sriraam Natarajan, Prasad Tadepalli
Abstract: Multiagent Reinforcement Learning (MARL) poses significant challenges due to the exponential growth of state and action spaces and the non-stationary nature of multiagent environments. This results in notable sample inefficiency and hinders generalization across diverse tasks. The complexity is further pronounced in relational settings, where domain knowledge is crucial but often underutilized by existing MARL algorithms. To overcome these hurdles, we propose integrating relational planners as centralized controllers with efficient state abstractions and reinforcement learning. This approach proves to be sample-efficient and facilitates effective task transfer and generalization.
Authors: Shiqi Wang, Zhibo Zhang, Libing Fang, Cam-Tu Nguyen, Wenzhon Li
Abstract: Corporate fraud detection aims to automatically recognize companies that conduct wrongful activities such as fraudulent financial statements or illegal insider trading. Previous learning-based methods fail to effectively integrate rich interactions in the company network. To close this gap, we collect 18-year financial records in China to form three graph datasets with fraud labels. We analyze the characteristics of the financial graphs, highlighting two pronounced issues: (1) information overload: the dominance of (noisy) non-company nodes over company nodes hinders the message-passing process in Graph Convolution Networks (GCN); and (2) hidden fraud: there exists a large percentage of possible undetected violations in the collected data. The hidden fraud problem will introduce noisy labels in the training dataset and compromise fraud detection results. To handle such challenges, we propose a novel graph-based method, namely, Knowledge-enhanced GCN with Robust Two-stage Learning (${\rm KeGCN}_{R}$), which leverages Knowledge Graph Embeddings to mitigate the information overload and effectively learns rich representations. The proposed model adopts a two-stage learning method to enhance robustness against hidden frauds. Extensive experimental results not only confirm the importance of interactions but also show the superiority of ${\rm KeGCN}_{R}$ over a number of strong baselines in terms of fraud detection effectiveness and robustness.
Authors: Michael Somma, Thomas Gallien, Branka Stojanovic
Abstract: Anomaly detection in complex dynamical systems is essential for ensuring reliability, safety, and efficiency in industrial and cyber-physical infrastructures. Predictive maintenance helps prevent costly failures, while cybersecurity monitoring has become critical as digitized systems face growing threats. Many of these systems exhibit oscillatory behaviors and bounded motion, requiring anomaly detection methods that capture structured temporal dependencies while adhering to physical consistency principles. In this work, we propose a system-theoretic approach to anomaly detection, grounded in classical embedding theory and physics-inspired consistency principles. We build upon the Fractal Whitney Embedding Prevalence Theorem, extending traditional embedding techniques to complex system dynamics. Additionally, we introduce state-derivative pairs as an embedding strategy to capture system evolution. To enforce temporal coherence, we develop a Temporal Differential Consistency Autoencoder (TDC-AE), incorporating a TDC-Loss that aligns the approximated derivatives of latent variables with their dynamic representations. We evaluate our method on the C-MAPSS dataset, a benchmark for turbofan aeroengine degradation. TDC-AE outperforms LSTMs and Transformers while achieving a 200x reduction in MAC operations, making it particularly suited for lightweight edge computing. Our findings support the hypothesis that anomalies disrupt stable system dynamics, providing a robust, interpretable signal for anomaly detection.
Authors: Christoph Benzm\"uller
Abstract: Deep and shallow embeddings of non-classical logics in classical higher-order logic have been explored, implemented, and used in various automated reasoning tools in recent years. This paper presents a recipe for the simultaneous deployment of different forms of deep and shallow embeddings in classical higher-order logic, enabling not only flexible interactive and automated theorem proving and counterexample finding at meta and object level, but also automated faithfulness proofs between the logic embeddings. The approach, which is fruitful for logic education, research and application, is deliberately illustrated here using simple propositional modal logic. However, the work presented is conceptual in nature and not limited to such a simple logic context.
Authors: Anikait Singh, Sheryl Hsu, Kyle Hsu, Eric Mitchell, Stefano Ermon, Tatsunori Hashimoto, Archit Sharma, Chelsea Finn
Abstract: Effective personalization of LLMs is critical for a broad range of user-interfacing applications such as virtual assistants and content curation. Inspired by the strong in-context learning capabilities of LLMs, we propose Few-Shot Preference Optimization (FSPO), which reframes reward modeling as a meta-learning problem. Under this framework, an LLM learns to quickly adapt to a user via a few labeled preferences from that user, constructing a personalized reward function for them. Additionally, since real-world preference data is scarce and challenging to collect at scale, we propose careful design choices to construct synthetic preference datasets for personalization, generating over 1M synthetic personalized preferences using publicly available LLMs. In particular, to successfully transfer from synthetic data to real users, we find it crucial for the data to exhibit both high diversity and coherent, self-consistent structure. We evaluate FSPO on personalized open-ended generation for up to 1,500 synthetic users across across three domains: movie reviews, pedagogical adaptation based on educational background, and general question answering, along with a controlled human study. Overall, FSPO achieves an 87% Alpaca Eval winrate on average in generating responses that are personalized to synthetic users and a 72% winrate with real human users in open-ended question answering.
Authors: Cornelius Emde, Alasdair Paren, Preetham Arvind, Maxime Kayser, Tom Rainforth, Thomas Lukasiewicz, Bernard Ghanem, Philip H. S. Torr, Adel Bibi
Abstract: Large language models (LLMs) are often deployed to perform constrained tasks, with narrow domains. For example, customer support bots can be built on top of LLMs, relying on their broad language understanding and capabilities to enhance performance. However, these LLMs are adversarially susceptible, potentially generating outputs outside the intended domain. To formalize, assess, and mitigate this risk, we introduce domain certification; a guarantee that accurately characterizes the out-of-domain behavior of language models. We then propose a simple yet effective approach, which we call VALID that provides adversarial bounds as a certificate. Finally, we evaluate our method across a diverse set of datasets, demonstrating that it yields meaningful certificates, which bound the probability of out-of-domain samples tightly with minimum penalty to refusal behavior.
Authors: Joel Veness, Marcus Hutter, Andras Gyorgy, Jordi Grau-Moya
Abstract: This paper considers a generalisation of universal source coding for interaction data, namely data streams that have actions interleaved with observations. Our goal will be to construct a coding distribution that is both universal \emph{and} can be used as a control policy. Allowing for action generation needs careful treatment, as naive approaches which do not distinguish between actions and observations run into the self-delusion problem in universal settings. We showcase our perspective in the context of the challenging non-stationary stochastic Bernoulli bandit problem. Our main contribution is an efficient and high performing algorithm for this problem that generalises the Partition Tree Weighting universal source coding technique for passive prediction to the control setting.
Authors: Hao Peng, Yunjia Qi, Xiaozhi Wang, Zijun Yao, Bin Xu, Lei Hou, Juanzi Li
Abstract: Reward models (RMs) are crucial for the training and inference-time scaling up of large language models (LLMs). However, existing reward models primarily focus on human preferences, neglecting verifiable correctness signals which have shown strong potential in training LLMs. In this paper, we propose agentic reward modeling, a reward system that combines reward models with verifiable correctness signals from different aspects to provide reliable rewards. We empirically implement a reward agent, named RewardAgent, that combines human preference rewards with two verifiable signals: factuality and instruction following, to provide more reliable rewards. We conduct comprehensive experiments on existing reward model benchmarks and inference time best-of-n searches on real-world downstream tasks. RewardAgent significantly outperforms vanilla reward models, demonstrating its effectiveness. We further construct training preference pairs using RewardAgent and train an LLM with the DPO objective, achieving superior performance on various NLP benchmarks compared to conventional reward models. Our codes are publicly released to facilitate further research (https://github.com/THU-KEG/Agentic-Reward-Modeling).
Authors: Diana Marie Schenke, Timo Baumann
Abstract: LLMs are not generally able to adjust the length of their outputs based on strict length requirements, a capability that would improve their usefulness in applications that require adherence to diverse user and system requirements. We present an approach to train LLMs to acquire this capability by augmenting existing data and applying existing fine-tuning techniques, which we compare based on the trained models' adherence to the length requirement and overall response quality relative to the baseline model. Our results demonstrate that these techniques can be successfully applied to train LLMs to adhere to length requirements, with the trained models generating texts which better align to the length requirements. Our results indicate that our method may change the response quality when using training data that was not generated by the baseline model. This allows simultaneous alignment to another training objective in certain scenarios, but is undesirable otherwise. Training on a dataset containing the model's own responses eliminates this issue.
Authors: Ademir G. Costa Junior, F\'abio S. da Silva, Ricardo Rios
Abstract: This paper presents a performance comparison among four Convolutional Neural Network architectures (EfficientNet-B3, InceptionV3, ResNet50, and VGG16) for classifying cassava disease images. The images were sourced from an imbalanced dataset from a competition. Appropriate metrics were employed to address class imbalance. The results indicate that EfficientNet-B3 achieved on this task accuracy of 87.7%, precision of 87.8%, revocation of 87.8% and F1-Score of 87.7%. These findings suggest that EfficientNet-B3 could be a valuable tool to support Digital Agriculture.
Authors: Aidan Furlong, Xingang Zhao, Robert Salko, Xu Wu
Abstract: Critical heat flux is a key quantity in boiling system modeling due to its impact on heat transfer and component temperature and performance. This study investigates the development and validation of an uncertainty-aware hybrid modeling approach that combines machine learning with physics-based models in the prediction of critical heat flux in nuclear reactors for cases of dryout. Two empirical correlations, Biasi and Bowring, were employed with three machine learning uncertainty quantification techniques: deep neural network ensembles, Bayesian neural networks, and deep Gaussian processes. A pure machine learning model without a base model served as a baseline for comparison. This study examines the performance and uncertainty of the models under both plentiful and limited training data scenarios using parity plots, uncertainty distributions, and calibration curves. The results indicate that the Biasi hybrid deep neural network ensemble achieved the most favorable performance (with a mean absolute relative error of 1.846% and stable uncertainty estimates), particularly in the plentiful data scenario. The Bayesian neural network models showed slightly higher error and uncertainty but superior calibration. By contrast, deep Gaussian process models underperformed by most metrics. All hybrid models outperformed pure machine learning configurations, demonstrating resistance against data scarcity.
Authors: Ru Peng, Kexin Yang, Yawen Zeng, Junyang Lin, Dayiheng Liu, Junbo Zhao
Abstract: The performance emergence of large language models (LLMs) driven by data scaling laws makes the selection of pre-training data increasingly important. However, existing methods rely on limited heuristics and human intuition, lacking comprehensive and clear guidelines. To address this, we are inspired by ``reverse thinking'' -- prompting LLMs to self-identify which criteria benefit its performance. As its pre-training capabilities are related to perplexity (PPL), we derive 14 quality criteria from the causes of text perplexity anomalies and introduce 15 common application domains to support domain mixing. In this paper, we train a Data Manager (DataMan) to learn quality ratings and domain recognition from pointwise rating, and use it to annotate a 447B token pre-training corpus with 14 quality ratings and domain type. Our experiments validate our approach, using DataMan to select 30B tokens to train a 1.3B-parameter language model, demonstrating significant improvements in in-context learning (ICL), perplexity, and instruction-following ability over the state-of-the-art baseline. The best-performing model, based on the Overall Score l=5 surpasses a model trained with 50% more data using uniform sampling. We continue pre-training with high-rated, domain-specific data annotated by DataMan to enhance domain-specific ICL performance and thus verify DataMan's domain mixing ability. Our findings emphasize the importance of quality ranking, the complementary nature of quality criteria, and their low correlation with perplexity, analyzing misalignment between PPL and ICL performance. We also thoroughly analyzed our pre-training dataset, examining its composition, the distribution of quality ratings, and the original document sources.
Authors: Arman Mielke, Uwe Bauknecht, Thilo Strauss, Mathias Niepert
Abstract: Combinatorial optimization (CO) problems arise in a wide range of fields from medicine to logistics and manufacturing. While exact solutions are often not necessary, many applications require finding high-quality solutions quickly. For this purpose, we propose a data-driven approach to improve existing non-learned approximation algorithms for CO. We parameterize the approximation algorithm and train a graph neural network (GNN) to predict parameter values that lead to the best possible solutions. Our pipeline is trained end-to-end in a self-supervised fashion using gradient estimation, treating the approximation algorithm as a black box. We propose a novel gradient estimation scheme for this purpose, which we call preference-based gradient estimation. Our approach combines the benefits of the neural network and the non-learned approximation algorithm: The GNN leverages the information from the dataset to allow the approximation algorithm to find better solutions, while the approximation algorithm guarantees that the solution is feasible. We validate our approach on two well-known combinatorial optimization problems, the travelling salesman problem and the minimum k-cut problem, and show that our method is competitive with state of the art learned CO solvers.
Authors: Ziqiao Weng, Weidong Cai, Bo Zhou
Abstract: Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder impacting social and behavioral development. Resting-state fMRI, a non-invasive tool for capturing brain connectivity patterns, aids in early ASD diagnosis and differentiation from typical controls (TC). However, previous methods, which rely on either mean time series or full 4D data, are limited by a lack of spatial information or by high computational costs. This underscores the need for an efficient solution that preserves both spatial and temporal information. In this paper, we propose a novel, simple, and efficient spatial-temporal-omics learning framework designed to efficiently extract spatio-temporal features from fMRI for ASD classification. Our approach addresses these limitations by utilizing 3D time-domain derivatives as the spatial-temporal inter-voxel omics, which preserve full spatial resolution while capturing diverse statistical characteristics of the time series at each voxel. Meanwhile, functional connectivity features serve as the spatial-temporal inter-regional omics, capturing correlations across brain regions. Extensive experiments and ablation studies on the ABIDE dataset demonstrate that our framework significantly outperforms previous methods while maintaining computational efficiency. We believe our research offers valuable insights that will inform and advance future ASD studies, particularly in the realm of spatial-temporal-omics-based learning.
Authors: Minjoo Lim, Bogyeong Kang, Tae-Eui Kam
Abstract: Multi-modal magnetic resonance imaging (MRI) is essential for providing complementary information about brain anatomy and pathology, leading to more accurate diagnoses. However, obtaining high-quality multi-modal MRI in a clinical setting is difficult due to factors such as time constraints, high costs, and patient movement artifacts. To overcome this difficulty, there is increasing interest in developing generative models that can synthesize missing target modality images from the available source ones. Therefore, we design a generative model for missing MRI that integrates multi-modal contrastive learning with a focus on critical tumor regions. Specifically, we integrate multi-modal contrastive learning, tailored for multiple source modalities, and enhance its effectiveness by selecting features based on entropy during the contrastive learning process. Additionally, our network not only generates the missing target modality images but also predicts segmentation outputs, simultaneously. This approach improves the generator's capability to precisely generate tumor regions, ultimately improving performance in downstream segmentation tasks. By leveraging a combination of contrastive, segmentation, and additional self-representation losses, our model effectively reflects target-specific information and generate high-quality target images. Consequently, our results in the Brain MR Image Synthesis challenge demonstrate that the proposed model excelled in generating the missing modality.
Authors: Xinru Wang, Mengjie Yu, Hannah Nguyen, Michael Iuzzolino, Tianyi Wang, Peiqi Tang, Natasha Lynova, Co Tran, Ting Zhang, Naveen Sendhilnathan, Hrvoje Benko, Haijun Xia, Tanya Jonker
Abstract: Large Language Models (LLMs) have shown remarkable potential in recommending everyday actions as personal AI assistants, while Explainable AI (XAI) techniques are being increasingly utilized to help users understand why a recommendation is given. Personal AI assistants today are often located on ultra-small devices such as smartwatches, which have limited screen space. The verbosity of LLM-generated explanations, however, makes it challenging to deliver glanceable LLM explanations on such ultra-small devices. To address this, we explored 1) spatially structuring an LLM's explanation text using defined contextual components during prompting and 2) presenting temporally adaptive explanations to users based on confidence levels. We conducted a user study to understand how these approaches impacted user experiences when interacting with LLM recommendations and explanations on ultra-small devices. The results showed that structured explanations reduced users' time to action and cognitive load when reading an explanation. Always-on structured explanations increased users' acceptance of AI recommendations. However, users were less satisfied with structured explanations compared to unstructured ones due to their lack of sufficient, readable details. Additionally, adaptively presenting structured explanations was less effective at improving user perceptions of the AI compared to the always-on structured explanations. Together with users' interview feedback, the results led to design implications to be mindful of when personalizing the content and timing of LLM explanations that are displayed on ultra-small devices.
Authors: Dayu Yang, Tianyang Liu, Daoan Zhang, Antoine Simoulin, Xiaoyi Liu, Yuwei Cao, Zhaopu Teng, Xin Qian, Grey Yang, Jiebo Luo, Julian McAuley
Abstract: In large language models (LLMs), code and reasoning reinforce each other: code offers an abstract, modular, and logic-driven structure that supports reasoning, while reasoning translates high-level goals into smaller, executable steps that drive more advanced code intelligence. In this study, we examine how code serves as a structured medium for enhancing reasoning: it provides verifiable execution paths, enforces logical decomposition, and enables runtime validation. We also explore how improvements in reasoning have transformed code intelligence from basic completion to advanced capabilities, enabling models to address complex software engineering tasks through planning and debugging. Finally, we identify key challenges and propose future research directions to strengthen this synergy, ultimately improving LLM's performance in both areas.
Authors: Christoph Schuhmann, Gollam Rabby, Ameya Prabhu, Tawsif Ahmed, Andreas Hochlehnert, Huu Nguyen, Nick Akinci Heidrich, Ludwig Schmidt, Robert Kaczmarczyk, S\"oren Auer, Jenia Jitsev, Matthias Bethge
Abstract: Paywalls, licenses and copyright rules often restrict the broad dissemination and reuse of scientific knowledge. We take the position that it is both legally and technically feasible to extract the scientific knowledge in scholarly texts. Current methods, like text embeddings, fail to reliably preserve factual content, and simple paraphrasing may not be legally sound. We urge the community to adopt a new idea: convert scholarly documents into Knowledge Units using LLMs. These units use structured data capturing entities, attributes and relationships without stylistic content. We provide evidence that Knowledge Units: (1) form a legally defensible framework for sharing knowledge from copyrighted research texts, based on legal analyses of German copyright law and U.S. Fair Use doctrine, and (2) preserve most (~95%) factual knowledge from original text, measured by MCQ performance on facts from the original copyrighted text across four research domains. Freeing scientific knowledge from copyright promises transformative benefits for scientific research and education by allowing language models to reuse important facts from copyrighted text. To support this, we share open-source tools for converting research documents into Knowledge Units. Overall, our work posits the feasibility of democratizing access to scientific knowledge while respecting copyright.
Authors: Akshat Gupta, Christine Fang, Atahan Ozdemir, Maochuan Lu, Ahmed Alaa, Thomas Hartvigsen, Gopala Anumanchipalli
Abstract: This study investigates the impact of localized updates to large language models (LLMs), specifically in the context of knowledge editing - a task aimed at incorporating or modifying specific facts without altering broader model capabilities. We first show that across different post-training interventions like continuous pre-training, full fine-tuning and LORA-based fine-tuning, the Frobenius norm of the updated matrices always increases. This increasing norm is especially detrimental for localized knowledge editing, where only a subset of matrices are updated in a model . We reveal a consistent phenomenon across various editing techniques, including fine-tuning, hypernetwork-based approaches, and locate-and-edit methods: the norm of the updated matrix invariably increases with successive updates. Such growth disrupts model balance, particularly when isolated matrices are updated while the rest of the model remains static, leading to potential instability and degradation of downstream performance. Upon deeper investigations of the intermediate activation vectors, we find that the norm of internal activations decreases and is accompanied by shifts in the subspaces occupied by these activations, which shows that these activation vectors now occupy completely different regions in the representation space compared to the unedited model. With our paper, we highlight the technical challenges with continuous and localized sequential knowledge editing and their implications for maintaining model stability and utility.
Authors: Lucy Xiaoyang Shi, Brian Ichter, Michael Equi, Liyiming Ke, Karl Pertsch, Quan Vuong, James Tanner, Anna Walling, Haohuan Wang, Niccolo Fusai, Adrian Li-Bell, Danny Driess, Lachy Groom, Sergey Levine, Chelsea Finn
Abstract: Generalist robots that can perform a range of different tasks in open-world settings must be able to not only reason about the steps needed to accomplish their goals, but also process complex instructions, prompts, and even feedback during task execution. Intricate instructions (e.g., "Could you make me a vegetarian sandwich?" or "I don't like that one") require not just the ability to physically perform the individual steps, but the ability to situate complex commands and feedback in the physical world. In this work, we describe a system that uses vision-language models in a hierarchical structure, first reasoning over complex prompts and user feedback to deduce the most appropriate next step to fulfill the task, and then performing that step with low-level actions. In contrast to direct instruction following methods that can fulfill simple commands ("pick up the cup"), our system can reason through complex prompts and incorporate situated feedback during task execution ("that's not trash"). We evaluate our system across three robotic platforms, including single-arm, dual-arm, and dual-arm mobile robots, demonstrating its ability to handle tasks such as cleaning messy tables, making sandwiches, and grocery shopping.
Authors: Brian Y. Lim, Joseph P. Cahaly, Chester Y. F. Sng, Adam Chew
Abstract: Many visualizations have been developed for explainable AI (XAI), but they often require further reasoning by users to interpret. Investigating XAI for high-stakes medical diagnosis, we propose improving domain alignment with diagrammatic and abductive reasoning to reduce the interpretability gap. We developed DiagramNet to predict cardiac diagnoses from heart auscultation, select the best-fitting hypothesis based on criteria evaluation, and explain with clinically-relevant murmur diagrams. The ante-hoc interpretable model leverages domain-relevant ontology, representation, and reasoning process to increase trust in expert users. In modeling studies, we found that DiagramNet not only provides faithful murmur shape explanations, but also has better performance than baseline models. We demonstrate the interpretability and trustworthiness of diagrammatic, abductive explanations in a qualitative user study with medical students, showing that clinically-relevant, diagrammatic explanations are preferred over technical saliency map explanations. This work contributes insights into providing domain-aligned explanations for user-centric XAI in complex domains.
Authors: Robert Susmaga, Izabela Szczech, Dariusz Brzezinski
Abstract: Multi-Criteria Decision Analysis (MCDA) is extensively used across diverse industries to assess and rank alternatives. Among numerous MCDA methods developed to solve real-world ranking problems, TOPSIS remains one of the most popular choices in many application areas. TOPSIS calculates distances between the considered alternatives and two predefined ones, namely the ideal and the anti-ideal, and creates a ranking of the alternatives according to a chosen aggregation of these distances. However, the interpretation of the inner workings of TOPSIS is difficult, especially when the number of criteria is large. To this end, recent research has shown that TOPSIS aggregations can be expressed using the means (M) and standard deviations (SD) of alternatives, creating MSD-space, a tool for visualizing and explaining aggregations. Even though MSD-space is highly useful, it assumes equally important criteria, making it less applicable to real-world ranking problems. In this paper, we generalize the concept of MSD-space to weighted criteria by introducing the concept of WMSD-space defined by what is referred to as weight-scaled means and standard deviations. We demonstrate that TOPSIS and similar distance-based aggregation methods can be successfully illustrated in a plane and interpreted even when the criteria are weighted, regardless of their number. The proposed WMSD-space offers a practical method for explaining TOPSIS rankings in real-world decision problems.
Authors: Sagar Malhotra, Davide Bizzaro, Luciano Serafini
Abstract: Weighted First Order Model Counting (WFOMC) is fundamental to probabilistic inference in statistical relational learning models. As WFOMC is known to be intractable in general ($\#$P-complete), logical fragments that admit polynomial time WFOMC are of significant interest. Such fragments are called domain liftable. Recent works have shown that the two-variable fragment of first order logic extended with counting quantifiers ($\mathrm{C^2}$) is domain-liftable. However, many properties of real-world data, like acyclicity in citation networks and connectivity in social networks, cannot be modeled in $\mathrm{C^2}$, or first order logic in general. In this work, we expand the domain liftability of $\mathrm{C^2}$ with multiple such properties. We show that any $\mathrm{C^2}$ sentence remains domain liftable when one of its relations is restricted to represent a directed acyclic graph, a connected graph, a tree (resp. a directed tree) or a forest (resp. a directed forest). All our results rely on a novel and general methodology of "counting by splitting". Besides their application to probabilistic inference, our results provide a general framework for counting combinatorial structures. We expand a vast array of previous results in discrete mathematics literature on directed acyclic graphs, phylogenetic networks, etc.
Authors: Zike Yuan, Ming Liu, Hui Wang, Bing Qin
Abstract: Evaluating the graph comprehension and reasoning abilities of Large Language Models (LLMs) is challenging and often incomplete. Existing benchmarks focus primarily on pure graph understanding, lacking a comprehensive evaluation across all graph types and detailed capability definitions. This paper presents GraCoRe, a benchmark for systematically assessing LLMs' graph comprehension and reasoning. GraCoRe uses a three-tier hierarchical taxonomy to categorize and test models on pure graph and heterogeneous graphs, subdividing capabilities into 10 distinct areas tested through 19 tasks. Our benchmark includes 11 datasets with 5,140 graphs of varying complexity. We evaluate four closed-source and eight open-source LLMs, conducting thorough analyses from both ability and task perspectives. Key findings reveal that OpenAI o1 model has amazing comprehension and reasoning capabilities, semantic enrichment enhances reasoning performance, node ordering impacts task success, and the ability to process longer texts does not necessarily improve graph comprehension or reasoning.GraCoRe is open-sourced at https://github.com/ZIKEYUAN/GraCoRe
Authors: Bj\"orn Schembera, Frank W\"ubbeling, Hendrik Kleikamp, Burkhard Schmidt, Aurela Shehu, Marco Reidelbach, Christine Biedinger, Jochen Fiedler, Thomas Koprucki, Dorothea Iglezakis, Dominik G\"oddeke
Abstract: Mathematical models and algorithms are an essential part of mathematical research data, as they are epistemically grounding numerical data. In order to represent models and algorithms as well as their relationship semantically to make this research data FAIR, two previously distinct ontologies were merged and extended, becoming a living knowledge graph. The link between the two ontologies is established by introducing computational tasks, as they occur in modeling, corresponding to algorithmic tasks. Moreover, controlled vocabularies are incorporated and a new class, distinguishing base quantities from specific use case quantities, was introduced. Also, both models and algorithms can now be enriched with metadata. Subject-specific metadata is particularly relevant here, such as the symmetry of a matrix or the linearity of a mathematical model. This is the only way to express specific workflows with concrete models and algorithms, as the feasible solution algorithm can only be determined if the mathematical properties of a model are known. We demonstrate this using two examples from different application areas of applied mathematics. In addition, we have already integrated over 250 research assets from applied mathematics into our knowledge graph.
Authors: Jiayi Zhang, Jinyu Xiang, Zhaoyang Yu, Fengwei Teng, Xionghui Chen, Jiaqi Chen, Mingchen Zhuge, Xin Cheng, Sirui Hong, Jinlin Wang, Bingnan Zheng, Bang Liu, Yuyu Luo, Chenglin Wu
Abstract: Large language models (LLMs) have demonstrated remarkable potential in solving complex tasks across diverse domains, typically by employing agentic workflows that follow detailed instructions and operational sequences. However, constructing these workflows requires significant human effort, limiting scalability and generalizability. Recent research has sought to automate the generation and optimization of these workflows, but existing methods still rely on initial manual setup and fall short of achieving fully automated and effective workflow generation. To address this challenge, we reformulate workflow optimization as a search problem over code-represented workflows, where LLM-invoking nodes are connected by edges. We introduce AFlow, an automated framework that efficiently explores this space using Monte Carlo Tree Search, iteratively refining workflows through code modification, tree-structured experience, and execution feedback. Empirical evaluations across six benchmark datasets demonstrate AFlow's efficacy, yielding a 5.7% average improvement over state-of-the-art baselines. Furthermore, AFlow enables smaller models to outperform GPT-4o on specific tasks at 4.55% of its inference cost in dollars. The code is available at https://github.com/geekan/MetaGPT.
Authors: Zhangchen Xu, Fengqing Jiang, Luyao Niu, Bill Yuchen Lin, Radha Poovendran
Abstract: Instruction tuning has been widely adopted to ensure large language models (LLMs) follow user instructions effectively. The resulting instruction-following capabilities of LLMs heavily rely on the instruction datasets used for tuning. Recently, synthetic instruction datasets have emerged as an economically viable solution to provide LLMs diverse and high-quality instructions. However, existing approaches typically assume that larger or stronger models are stronger teachers for instruction tuning, and hence simply adopt these models as response generators to the synthetic instructions. In this paper, we challenge this commonly-adopted assumption. Our extensive experiments across five base models and twenty response generators reveal that larger and stronger models are not necessarily stronger teachers of smaller models. We refer to this phenomenon as the Larger Models' Paradox. We observe that existing metrics cannot precisely predict the effectiveness of response generators since they ignore the compatibility between teachers and base models being fine-tuned. We thus develop a novel metric, named as Compatibility-Adjusted Reward (CAR) to measure the effectiveness of response generators. Our experiments across five base models demonstrate that CAR outperforms almost all baselines.
Authors: Yuang Wei, Yuan-Hao Jiang, Jiayi Liu, Changyong Qi, Linzhao Jia, Rui Jia
Abstract: The rapid development of Generative AI (GAI) has sparked revolutionary changes across various aspects of education. Personalized learning, a focal point and challenge in educational research, has also been influenced by the development of GAI. To explore GAI's extensive impact on personalized learning, this study investigates its potential to enhance various facets of personalized learning through a thorough analysis of existing research. The research comprehensively examines GAI's influence on personalized learning by analyzing its application across different methodologies and contexts, including learning strategies, paths, materials, environments, and specific analyses within the teaching and learning processes. Through this in-depth investigation, we find that GAI demonstrates exceptional capabilities in providing adaptive learning experiences tailored to individual preferences and needs. Utilizing different forms of GAI across various subjects yields superior learning outcomes. The article concludes by summarizing scenarios where GAI is applicable in educational processes and discussing strategies for leveraging GAI to enhance personalized learning, aiming to guide educators and learners in effectively utilizing GAI to achieve superior learning objectives.
Authors: Yule Liu, Zhiyuan Zhong, Yifan Liao, Zhen Sun, Jingyi Zheng, Jiaheng Wei, Qingyuan Gong, Fenghua Tong, Yang Chen, Yang Zhang, Xinlei He
Abstract: The rising popularity of large language models (LLMs) has raised concerns about machine-generated text (MGT), particularly in academic settings, where issues like plagiarism and misinformation are prevalent. As a result, developing a highly generalizable and adaptable MGT detection system has become an urgent priority. Given that LLMs are most commonly misused in academic writing, this work investigates the generalization and adaptation capabilities of MGT detectors in three key aspects specific to academic writing: First, we construct MGT-Acedemic, a large-scale dataset comprising over 336M tokens and 749K samples. MGT-Acedemic focuses on academic writing, featuring human-written texts (HWTs) and MGTs across STEM, Humanities, and Social Sciences, paired with an extensible code framework for efficient benchmarking. Second, we benchmark the performance of various detectors for binary classification and attribution tasks in both in-domain and cross-domain settings. This benchmark reveals the often-overlooked challenges of attribution tasks. Third, we introduce a novel attribution task where models have to adapt to new classes over time without (or with very limited) access to prior training data in both few-shot and many-shot scenarios. We implement eight different adapting techniques to improve the performance and highlight the inherent complexity of the task. Our findings provide insights into the generalization and adaptation ability of MGT detectors across diverse scenarios and lay the foundation for building robust, adaptive detection systems. The code framework is available at https://github.com/Y-L-LIU/MGTBench-2.0.
Authors: Kai Yan, Zhan Ling, Kang Liu, Yifan Yang, Ting-Han Fan, Lingfeng Shen, Zhengyin Du, Jiecao Chen
Abstract: Inductive Reasoning (IR), the ability to summarize rules from examples and apply on new ones, has long been viewed as a primal ability for general intelligence and widely studied by cognitive science and AI researchers. Many benchmarks have been proposed to measure such ability for Large Language Models (LLMs); however, they focus on few-shot (usually $<$10) setting and lack evaluation for aggregating many pieces of information from long contexts. On the other hand, the ever-growing context length of LLMs have brought forth the novel paradigm of many-shot In-Context Learning (ICL), which addresses new tasks with hundreds to thousands of examples without expensive and inefficient fine-tuning. However, many-shot evaluations are mostly focused on classification (a very limited aspect of IR), and popular long-context LLM tasks such as Needle-In-A-Haystack (NIAH) seldom require complicated intelligence for integrating many pieces of information. To fix the issues from both worlds, we propose MIR-Bench, the first many-shot in-context inductive reasoning benchmark that asks LLM to induce output via input-output examples from underlying functions with diverse data format. Based on MIR-Bench, we study many novel problems for inductive reasoning and many-shot ICL, including robustness against erroneous shots and the effect of Chain-of-Thought (CoT), and acquired insightful findings.
Authors: Yuze Zhao, Tianyun Ji, Wenjun Feng, Zhenya Huang, Qi Liu, Zhiding Liu, Yixiao Ma, Kai Zhang, Enhong Chen
Abstract: The reasoning abilities are one of the most enigmatic and captivating aspects of large language models (LLMs). Numerous studies are dedicated to exploring and expanding the boundaries of this reasoning capability. However, tasks that embody both reasoning and recall characteristics are often overlooked. In this paper, we introduce such a novel task, code reasoning, to provide a new perspective for the reasoning abilities of LLMs. We summarize three meta-benchmarks based on established forms of logical reasoning, and instantiate these into eight specific benchmark tasks. Our testing on these benchmarks reveals that LLMs continue to struggle with identifying satisfactory reasoning pathways. Additionally, we present a new pathway exploration pipeline inspired by human intricate problem-solving methods. This Reflective Hypothesis Decomposition and Amendment (RHDA) pipeline consists of the following iterative steps: (1) Proposing potential hypotheses based on observations and decomposing them; (2) Utilizing tools to validate hypotheses and reflection outcomes; (3) Revising hypothesis in light of observations. Our approach effectively mitigates logical chain collapses arising from forgetting or hallucination issues in multi-step reasoning, resulting in performance gains of up to $3\times$. Finally, we expanded this pipeline by applying it to simulate complex household tasks in real-world scenarios, specifically in VirtualHome, enhancing the handling of failure cases. We release our code and all of results at https://github.com/TnTWoW/code_reasoning.
Authors: Ran Ding, Ziyu Zhang, Ying Zhu, Ziqian Kong, Peilan Xu
Abstract: To enhance tourists' experiences and immersion, this paper proposes a narrative-driven travel planning framework called NarrativeGuide, which generates a geoculturally-grounded narrative script for travelers, offering a novel, role-playing experience for their journey. In the initial stage, NarrativeGuide constructs a knowledge graph for attractions within a city, then configures the worldview, character setting, and exposition based on the knowledge graph. Using this foundation, the knowledge graph is combined to generate an independent scene unit for each attraction. During the itinerary planning stage, NarrativeGuide models narrative-driven travel planning as an optimization problem, utilizing a genetic algorithm (GA) to refine the itinerary. Before evaluating the candidate itinerary, transition scripts are generated for each pair of adjacent attractions, which, along with the scene units, form a complete script. The weighted sum of script coherence, travel time, and attraction scores is then used as the fitness value to update the candidate solution set. Experimental results across four cities, i.e., Nanjing and Yangzhou in China, Paris in France, and Berlin in Germany, demonstrate significant improvements in narrative coherence and cultural fit, alongside a notable reduction in travel time and an increase in the quality of visited attractions. Our study highlights that incorporating external evolutionary optimization effectively addresses the limitations of large language models in travel planning.
Authors: Yuehong Cassandra Tai, Khushi Navin Patni, Nicholas Daniel Hemauer, Bruce Desmarais, Yu-Ru Lin
Abstract: Despite recent advances in understanding the capabilities and limits of generative artificial intelligence (GenAI) models, we are just beginning to understand their capacity to assess and reason about the veracity of content. We evaluate multiple GenAI models across tasks that involve the rating of, and perceived reasoning about, the credibility of information. The information in our experiments comes from content that subnational U.S. politicians post to Facebook. We find that GPT-4o, one of the most used AI models in consumer applications, outperforms other models, but all models exhibit only moderate agreement with human coders. Importantly, even when GenAI models accurately identify low-credibility content, their reasoning relies heavily on linguistic features and ``hard'' criteria, such as the level of detail, source reliability, and language formality, rather than an understanding of veracity. We also assess the effectiveness of summarized versus full content inputs, finding that summarized content holds promise for improving efficiency without sacrificing accuracy. While GenAI has the potential to support human fact-checkers in scaling misinformation detection, our results caution against relying solely on these models.
Authors: Patrick Tser Jern Kon, Jiachen Liu, Qiuyi Ding, Yiming Qiu, Zhenning Yang, Yibo Huang, Jayanth Srinivasa, Myungjin Lee, Mosharaf Chowdhury, Ang Chen
Abstract: Scientific experimentation, a cornerstone of human progress, demands rigor in reliability, methodical control, and interpretability to yield meaningful results. Despite the growing capabilities of large language models (LLMs) in automating different aspects of the scientific process, automating rigorous experimentation remains a significant challenge. To address this gap, we propose Curie, an AI agent framework designed to embed rigor into the experimentation process through three key components: an intra-agent rigor module to enhance reliability, an inter-agent rigor module to maintain methodical control, and an experiment knowledge module to enhance interpretability. To evaluate Curie, we design a novel experimental benchmark composed of 46 questions across four computer science domains, derived from influential research papers, and widely adopted open-source projects. Compared to the strongest baseline tested, we achieve a 3.4$\times$ improvement in correctly answering experimental questions. Curie is open-sourced at https://github.com/Just-Curieous/Curie.
Authors: Guoxin Wang, Minyu Gao, Shuai Yang, Ya Zhang, Lizhi He, Liang Huang, Hanlin Xiao, Yexuan Zhang, Wanyue Li, Lu Chen, Jintao Fei, Xin Li
Abstract: Large language models (LLMs), particularly those with reasoning capabilities, have rapidly advanced in recent years, demonstrating significant potential across a wide range of applications. However, their deployment in healthcare, especially in disease reasoning tasks, is hindered by the challenge of acquiring expert-level cognitive data. In this paper, we introduce Citrus, a medical language model that bridges the gap between clinical expertise and AI reasoning by emulating the cognitive processes of medical experts. The model is trained on a large corpus of simulated expert disease reasoning data, synthesized using a novel approach that accurately captures the decision-making pathways of clinicians. This approach enables Citrus to better simulate the complex reasoning processes involved in diagnosing and treating medical conditions. To further address the lack of publicly available datasets for medical reasoning tasks, we release the last-stage training data, including a custom-built medical diagnostic dialogue dataset. This open-source contribution aims to support further research and development in the field. Evaluations using authoritative benchmarks such as MedQA, covering tasks in medical reasoning and language understanding, show that Citrus achieves superior performance compared to other models of similar size. These results highlight Citrus potential to significantly enhance medical decision support systems, providing a more accurate and efficient tool for clinical decision-making.
Authors: Minhua Lin, Hui Liu, Xianfeng Tang, Jingying Zeng, Zhenwei Dai, Chen Luo, Zheng Li, Xiang Zhang, Qi He, Suhang Wang
Abstract: Search plays a fundamental role in problem-solving across various domains, with most real-world decision-making problems being solvable through systematic search. Drawing inspiration from recent discussions on search and learning, we systematically explore the complementary relationship between search and Large Language Models (LLMs) from three perspectives. First, we analyze how learning can enhance search efficiency and propose Search via Learning (SeaL), a framework that leverages LLMs for effective and efficient search. Second, we further extend SeaL to SeaL-C to ensure rigorous completeness during search. Our evaluation across three real-world planning tasks demonstrates that SeaL achieves near-perfect accuracy while reducing search spaces by up to 99.1% compared to traditional approaches. Finally, we explore how far LLMs are from real search by investigating whether they can develop search capabilities independently. Our analysis reveals that while current LLMs struggle with efficient search in complex problems, incorporating systematic search strategies significantly enhances their problem-solving capabilities. These findings not only validate the effectiveness of our approach but also highlight the need for improving LLMs' search abilities for real-world applications.
Authors: Kacper Sokol, Peter Flach
Abstract: Explainable artificial intelligence provides tools to better understand predictive models and their decisions, but many such methods are limited to producing insights with respect to a single class. When generating explanations for several classes, reasoning over them to obtain a comprehensive view may be difficult since they can present competing or contradictory evidence. To address this challenge we introduce the novel paradigm of multi-class explanations. We outline the theory behind such techniques and propose a local surrogate model based on multi-output regression trees -- called LIMEtree -- that offers faithful and consistent explanations of multiple classes for individual predictions while being post-hoc, model-agnostic and data-universal. On top of strong fidelity guarantees, our implementation delivers a range of diverse explanation types, including counterfactual statements favoured in the literature. We evaluate our algorithm with respect to explainability desiderata, through quantitative experiments and via a pilot user study, on image and tabular data classification tasks, comparing it to LIME, which is a state-of-the-art surrogate explainer. Our contributions demonstrate the benefits of multi-class explanations and wide-ranging advantages of our method across a diverse set of scenarios.
Authors: Markus K\"angsepp, Kaspar Valk, Meelis Kull
Abstract: Every uncalibrated classifier has a corresponding true calibration map that calibrates its confidence. Deviations of this idealistic map from the identity map reveal miscalibration. Such calibration errors can be reduced with many post-hoc calibration methods which fit some family of calibration maps on a validation dataset. In contrast, evaluation of calibration with the expected calibration error (ECE) on the test set does not explicitly involve fitting. However, as we demonstrate, ECE can still be viewed as if fitting a family of functions on the test data. This motivates the fit-on-the-test view on evaluation: first, approximate a calibration map on the test data, and second, quantify its distance from the identity. Exploiting this view allows us to unlock missed opportunities: (1) use the plethora of post-hoc calibration methods for evaluating calibration; (2) tune the number of bins in ECE with cross-validation. Furthermore, we introduce: (3) benchmarking on pseudo-real data where the true calibration map can be estimated very precisely; and (4) novel calibration and evaluation methods using new calibration map families PL and PL3.
Authors: Arjun Ramachandra, Karthik Natarajan
Abstract: In this paper, we introduce the notion of a ``pairwise independent correlation gap'' for set functions with random elements. The pairwise independent correlation gap is defined as the ratio of the maximum expected value of a set function with arbitrary dependence among the elements with fixed marginal probabilities to the maximum expected value with pairwise independent elements with the same marginal probabilities. We show that for any nonnegative monotone submodular set function defined on $n$ elements, this ratio is upper bounded by $4/3$ in the following two cases: (a) $n = 3$ for all marginal probabilities and (b) all $n$ for small marginal probabilities (and similarly large marginal probabilities). This differs from the bound on the ``correlation gap'' which holds with mutual independence and showcases the fundamental difference between pairwise independence and mutual independence. We discuss the implication of the results with two examples and end the paper with a conjecture.
Authors: Alexandra Cimpean, Timothy Verstraeten, Lander Willem, Niel Hens, Ann Now\'e, Pieter Libin
Abstract: Individual-based epidemiological models support the study of fine-grained preventive measures, such as tailored vaccine allocation policies, in silico. As individual-based models are computationally intensive, it is pivotal to identify optimal strategies within a reasonable computational budget. Moreover, due to the high societal impact associated with the implementation of preventive strategies, uncertainty regarding decisions should be communicated to policy makers, which is naturally embedded in a Bayesian approach. We present a novel technique for evaluating vaccine allocation strategies using a multi-armed bandit framework in combination with a Bayesian anytime $m$-top exploration algorithm. $m$-top exploration allows the algorithm to learn $m$ policies for which it expects the highest utility, enabling experts to inspect this small set of alternative strategies, along with their quantified uncertainty. The anytime component provides policy advisors with flexibility regarding the computation time and the desired confidence, which is important as it is difficult to make this trade-off beforehand. We consider the Belgian COVID-19 epidemic using the individual-based model STRIDE, where we learn a set of vaccination policies that minimize the number of infections and hospitalisations. Through experiments we show that our method can efficiently identify the $m$-top policies, which is validated in a scenario where the ground truth is available. Finally, we explore how vaccination policies can best be organised under different contact reduction schemes and we investigate the impact of vaccine uptake proportions (i.e., the proportion of individuals that will comply with the strategy and take the vaccine).
Authors: Przemyslaw Grabowicz, Adrian Byrne, Cyrus Cousins, Nicholas Perello, Yair Zick
Abstract: We propose establishing an office to oversee AI systems by introducing a tiered system of explainability and benchmarking requirements for commercial AI systems. We examine how complex high-risk technologies have been successfully regulated at the national level. Specifically, we draw parallels to the existing regulation for the U.S. medical device industry and the pharmaceutical industry (regulated by the FDA), the proposed legislation for AI in the European Union (the AI Act), and the existing U.S. anti-discrimination legislation. To promote accountability and user trust, AI accountability mechanisms shall introduce standarized measures for each category of intended high-risk use of AI systems to enable structured comparisons among such AI systems. We suggest using explainable AI techniques, such as input influence measures, as well as fairness statistics and other performance measures of high-risk AI systems. We propose to standardize internal benchmarking and automated audits to transparently characterize high-risk AI systems. The results of such audits and benchmarks shall be clearly and transparently communicated and explained to enable meaningful comparisons of competing AI systems via a public AI registry. Such standardized audits, benchmarks, and certificates shall be specific to intended high-risk use of respective AI systems and could constitute conformity assessment for AI systems, e.g., in the European Union's AI Act.
Authors: Santanu Ganguly
Abstract: Quantum machine learning (QML) is a cross-disciplinary subject made up of two of the most exciting research areas: quantum computing and classical machine learning (ML), with ML and artificial intelligence (AI) being projected as the first fields that will be impacted by the rise of quantum machines. Quantum computers are being used today in drug discovery, material & molecular modelling and finance. In this work, we discuss some upcoming active new research areas in application of quantum machine learning (QML) in finance. We discuss certain QML models that has become areas of active interest in the financial world for various applications. We use real world financial dataset and compare models such as qGAN (quantum generative adversarial networks) and QCBM (quantum circuit Born machine) among others, using simulated environments. For the qGAN, we define quantum circuits for discriminators and generators and show promises of future quantum advantage via QML in finance.
Authors: Arnulf Jentzen, Benno Kuckuck, Philippe von Wurstemberger
Abstract: This book aims to provide an introduction to the topic of deep learning algorithms. We review essential components of deep learning algorithms in full mathematical detail including different artificial neural network (ANN) architectures (such as fully-connected feedforward ANNs, convolutional ANNs, recurrent ANNs, residual ANNs, and ANNs with batch normalization) and different optimization algorithms (such as the basic stochastic gradient descent (SGD) method, accelerated methods, and adaptive methods). We also cover several theoretical aspects of deep learning algorithms such as approximation capacities of ANNs (including a calculus for ANNs), optimization theory (including Kurdyka-{\L}ojasiewicz inequalities), and generalization errors. In the last part of the book some deep learning approximation methods for PDEs are reviewed including physics-informed neural networks (PINNs) and deep Galerkin methods. We hope that this book will be useful for students and scientists who do not yet have any background in deep learning at all and would like to gain a solid foundation as well as for practitioners who would like to obtain a firmer mathematical understanding of the objects and methods considered in deep learning.
Authors: Hongyin Zhang, Diyuan Shi, Zifeng Zhuang, Han Zhao, Zhenyu Wei, Feng Zhao, Sibo Gai, Shangke Lyu, Donglin Wang
Abstract: Developing robotic intelligent systems that can adapt quickly to unseen wild situations is one of the critical challenges in pursuing autonomous robotics. Although some impressive progress has been made in walking stability and skill learning in the field of legged robots, their ability for fast adaptation is still inferior to that of animals in nature. Animals are born with a massive set of skills needed to survive, and can quickly acquire new ones, by composing fundamental skills with limited experience. Inspired by this, we propose a novel framework, named Robot Skill Graph (RSG) for organizing a massive set of fundamental skills of robots and dexterously reusing them for fast adaptation. Bearing a structure similar to the Knowledge Graph (KG), RSG is composed of massive dynamic behavioral skills instead of static knowledge in KG and enables discovering implicit relations that exist in between the learning context and acquired skills of robots, serving as a starting point for understanding subtle patterns existing in robots' skill learning. Extensive experimental results demonstrate that RSG can provide reliable skill inference upon new tasks and environments, and enable quadruped robots to adapt to new scenarios and quickly learn new skills.
Authors: Jessica Ojo, Odunayo Ogundepo, Akintunde Oladipo, Kelechi Ogueji, Jimmy Lin, Pontus Stenetorp, David Ifeoluwa Adelani
Abstract: Large-scale multilingual evaluations, such as MEGA, often include only a handful of African languages due to the scarcity of high-quality evaluation data and the limited discoverability of existing African datasets. This lack of representation hinders comprehensive LLM evaluation across a diverse range of languages and tasks. To address these challenges, we introduce AfroBench -- a multi-task benchmark for evaluating the performance of LLMs across 64 African languages, 15 tasks and 22 datasets. AfroBench consists of nine natural language understanding datasets, six text generation datasets, six knowledge and question answering tasks, and one mathematical reasoning task. We present results comparing the performance of prompting LLMs to fine-tuned baselines based on BERT and T5-style models. Our results suggest large gaps in performance between high-resource languages, such as English, and African languages across most tasks; but performance also varies based on the availability of monolingual data resources. Our findings confirm that performance on African languages continues to remain a hurdle for current LLMs, underscoring the need for additional efforts to close this gap. https://mcgill-nlp.github.io/AfroBench/
Authors: Ye Lin Tun, Chu Myaet Thwal, Huy Q. Le, Minh N. H. Nguyen, Choong Seon Hong
Abstract: Many studies integrate federated learning (FL) with self-supervised learning (SSL) to take advantage of raw data distributed across edge devices. However, edge devices often struggle with high computational and communication costs imposed by SSL and FL algorithms. With the deployment of more complex and large-scale models, such as Transformers, these challenges are exacerbated. To tackle this, we propose the Layer-Wise Federated Self-Supervised Learning (LW-FedSSL) approach, which allows edge devices to incrementally train a small part of the model at a time. Specifically, in LW-FedSSL, training is decomposed into multiple stages, with each stage responsible for only a specific layer (or a block of layers) of the model. Since only a portion of the model is active for training at any given time, LW-FedSSL significantly reduces computational requirements. Additionally, only the active model portion needs to be exchanged between the FL server and clients, reducing the communication overhead. This enables LW-FedSSL to jointly address both computational and communication challenges in FL. Depending on the SSL algorithm used, it can achieve up to a $3.34 \times$ reduction in memory usage, $4.20 \times$ fewer computational operations (GFLOPs), and a $5.07 \times$ lower communication cost while maintaining performance comparable to its end-to-end training counterpart. Furthermore, we explore a progressive training strategy called Prog-FedSSL, which offers a $1.84\times$ reduction in GFLOPs and a $1.67\times$ reduction in communication costs while maintaining the same memory requirements as end-to-end training. While the resource efficiency of Prog-FedSSL is lower than that of LW-FedSSL, its performance improvements make it a viable candidate for FL environments with more lenient resource constraints.
Authors: Siyi Liu, Chen Gao, Yong Li
Abstract: Hyperparameter optimization is critical in modern machine learning, requiring expert knowledge, numerous trials, and high computational and human resources. Despite the advancements in Automated Machine Learning (AutoML), challenges in terms of trial efficiency, setup complexity, and interoperability still persist. To address these issues, we introduce a novel paradigm leveraging Large Language Models (LLMs) to automate hyperparameter optimization across diverse machine learning tasks, which is named AgentHPO (short for LLM Agent-based Hyperparameter Optimization). Specifically, AgentHPO processes the task information autonomously, conducts experiments with specific hyperparameters (HPs), and iteratively optimizes them based on historical trials. This human-like optimization process largely reduces the number of required trials, simplifies the setup process, and enhances interpretability and user trust, compared to traditional AutoML methods. Extensive empirical experiments conducted on 12 representative machine-learning tasks indicate that AgentHPO not only matches but also often surpasses the best human trials in terms of performance while simultaneously providing explainable results. Further analysis sheds light on the strategies employed by the LLM in optimizing these tasks, highlighting its effectiveness and adaptability in various scenarios.
Authors: David Haag, Devender Kumar, Sebastian Gruber, Dominik Hofer, Mahdi Sareban, Gunnar Treff, Josef Niebauer, Christopher Bull, Albrecht Schmidt, Jan David Smeddinck
Abstract: We evaluated the viability of using Large Language Models (LLMs) to trigger and personalize content in Just-in-Time Adaptive Interventions (JITAIs) in digital health. As an interaction pattern representative of context-aware computing, JITAIs are being explored for their potential to support sustainable behavior change, adapting interventions to an individual's current context and needs. Challenging traditional JITAI implementation models, which face severe scalability and flexibility limitations, we tested GPT-4 for suggesting JITAIs in the use case of heart-healthy activity in cardiac rehabilitation. Using three personas representing patients affected by CVD with varying severeness and five context sets per persona, we generated 450 JITAI decisions and messages. These were systematically evaluated against those created by 10 laypersons (LayPs) and 10 healthcare professionals (HCPs). GPT-4-generated JITAIs surpassed human-generated intervention suggestions, outperforming both LayPs and HCPs across all metrics (i.e., appropriateness, engagement, effectiveness, and professionalism). These results highlight the potential of LLMs to enhance JITAI implementations in personalized health interventions, demonstrating how generative AI could revolutionize context-aware computing.
Authors: Pengzhou Cheng, Wei Du, Zongru Wu, Fengwei Zhang, Libo Chen, Zhuosheng Zhang, Gongshen Liu
Abstract: Although pre-training achieves remarkable performance, it suffers from task-agnostic backdoor attacks due to vulnerabilities in data and training mechanisms. These attacks can transfer backdoors to various downstream tasks. In this paper, we introduce $\mathtt{maxEntropy}$, an entropy-based poisoning filter that mitigates such risks. To overcome the limitations of manual target setting and explicit triggers, we propose $\mathtt{SynGhost}$, an invisible and universal task-agnostic backdoor attack via syntactic transfer, further exposing vulnerabilities in pre-trained language models (PLMs). Specifically, $\mathtt{SynGhost}$ injects multiple syntactic backdoors into the pre-training space through corpus poisoning, while preserving the PLM's pre-training capabilities. Second, $\mathtt{SynGhost}$ adaptively selects optimal targets based on contrastive learning, creating a uniform distribution in the pre-training space. To identify syntactic differences, we also introduce an awareness module to minimize interference between backdoors. Experiments show that $\mathtt{SynGhost}$ poses significant threats and can transfer to various downstream tasks. Furthermore, $\mathtt{SynGhost}$ resists defenses based on perplexity, fine-pruning, and $\mathtt{maxEntropy}$. The code is available at https://github.com/Zhou-CyberSecurity-AI/SynGhost.
Authors: Jana\'ina Mendes-Laureano, Jorge A. G\'omez-Garc\'ia, Alejandro Guerrero-L\'opez, Elisa Luque-Buzo, Juli\'an D. Arias-Londo\~no, Francisco J. Grandas-P\'erez, Juan I. Godino-Llorente
Abstract: The screening of Parkinson's Disease (PD) through speech is hindered by a notable lack of publicly available datasets in different languages. This fact limits the reproducibility and further exploration of existing research. To address this gap, this manuscript presents the NeuroVoz corpus consisting of 112 native Castilian-Spanish speakers, including 58 healthy controls and 54 individuals with PD, all recorded in ON state. The corpus showcases a diverse array of speech tasks: sustained vowels; diadochokinetic tests; 16 Listen-and-Repeat utterances; and spontaneous monologues. The dataset is also complemented with subjective assessments of voice quality performed by an expert according to the GRBAS scale (Grade/Roughness/Breathiness/Asthenia/Strain), as well as annotations with a thorough examination of phonation quality, intensity, speed, resonance, intelligibility, and prosody. The corpus offers a substantial resource for the exploration of the impact of PD on speech. This data set has already supported several studies, achieving a benchmark accuracy of 89% for the screening of PD. Despite these advances, the broader challenge of conducting a language-agnostic, cross-corpora analysis of Parkinsonian speech patterns remains open.
Authors: Mark Russinovich, Ahmed Salem, Ronen Eldan
Abstract: Large Language Models (LLMs) have risen significantly in popularity and are increasingly being adopted across multiple applications. These LLMs are heavily aligned to resist engaging in illegal or unethical topics as a means to avoid contributing to responsible AI harms. However, a recent line of attacks, known as jailbreaks, seek to overcome this alignment. Intuitively, jailbreak attacks aim to narrow the gap between what the model can do and what it is willing to do. In this paper, we introduce a novel jailbreak attack called Crescendo. Unlike existing jailbreak methods, Crescendo is a simple multi-turn jailbreak that interacts with the model in a seemingly benign manner. It begins with a general prompt or question about the task at hand and then gradually escalates the dialogue by referencing the model's replies progressively leading to a successful jailbreak. We evaluate Crescendo on various public systems, including ChatGPT, Gemini Pro, Gemini-Ultra, LlaMA-2 70b and LlaMA-3 70b Chat, and Anthropic Chat. Our results demonstrate the strong efficacy of Crescendo, with it achieving high attack success rates across all evaluated models and tasks. Furthermore, we present Crescendomation, a tool that automates the Crescendo attack and demonstrate its efficacy against state-of-the-art models through our evaluations. Crescendomation surpasses other state-of-the-art jailbreaking techniques on the AdvBench subset dataset, achieving 29-61% higher performance on GPT-4 and 49-71% on Gemini-Pro. Finally, we also demonstrate Crescendo's ability to jailbreak multimodal models.
Authors: K. Aditya Mohan, Massimiliano Ferrucci, Chuck Divin, Garrett A. Stevenson, Hyojin Kim
Abstract: 4D time-space reconstruction of dynamic events or deforming objects using X-ray computed tomography (CT) is an important inverse problem in non-destructive evaluation. Conventional back-projection based reconstruction methods assume that the object remains static for the duration of several tens or hundreds of X-ray projection measurement images (reconstruction of consecutive limited-angle CT scans). However, this is an unrealistic assumption for many in-situ experiments that causes spurious artifacts and inaccurate morphological reconstructions of the object. To solve this problem, we propose to perform a 4D time-space reconstruction using a distributed implicit neural representation (DINR) network that is trained using a novel distributed stochastic training algorithm. Our DINR network learns to reconstruct the object at its output by iterative optimization of its network parameters such that the measured projection images best match the output of the CT forward measurement model. We use a forward measurement model that is a function of the DINR outputs at a sparsely sampled set of continuous valued 4D object coordinates. Unlike previous neural representation architectures that forward and back propagate through dense voxel grids that sample the object's entire time-space coordinates, we only propagate through the DINR at a small subset of object coordinates in each iteration resulting in an order-of-magnitude reduction in memory and compute for training. DINR leverages distributed computation across several compute nodes and GPUs to produce high-fidelity 4D time-space reconstructions. We use both simulated parallel-beam and experimental cone-beam X-ray CT datasets to demonstrate the superior performance of our approach.
Authors: David Kremer, Victor Villar, Hanhee Paik, Ivan Duran, Ismael Faro, Juan Cruz-Benito
Abstract: This paper demonstrates the integration of Reinforcement Learning (RL) into quantum transpiling workflows, significantly enhancing the synthesis and routing of quantum circuits. By employing RL, we achieve near-optimal synthesis of Linear Function, Clifford, and Permutation circuits, up to 9, 11 and 65 qubits respectively, while being compatible with native device instruction sets and connectivity constraints, and orders of magnitude faster than optimization methods such as SAT solvers. We also achieve significant reductions in two-qubit gate depth and count for circuit routing up to 133 qubits with respect to other routing heuristics such as SABRE. We find the method to be efficient enough to be useful in practice in typical quantum transpiling pipelines. Our results set the stage for further AI-powered enhancements of quantum computing workflows.
Authors: Shashata Sawmya, Linghao Kong, Ilia Markov, Dan Alistarh, Nir Shavit
Abstract: Disentangling polysemantic neurons is at the core of many current approaches to interpretability of large language models. Here we attempt to study how disentanglement can be used to understand performance, particularly under weight sparsity, a leading post-training optimization technique. We suggest a novel measure for estimating neuronal entanglement: the Wasserstein distance of a neuron's output distribution to a Gaussian. Moreover, we show the existence of a small number of highly entangled "Wasserstein Neurons" in each linear layer of an LLM, characterized by their highly non-Gaussian output distributions, their role in mapping similar inputs to dissimilar outputs, and their significant impact on model accuracy. To study these phenomena, we propose a new experimental framework for disentangling polysemantic neurons. Our framework separates each layer's inputs to create a mixture of experts where each neuron's output is computed by a mixture of neurons of lower Wasserstein distance, each better at maintaining accuracy when sparsified without retraining. We provide strong evidence that this is because the mixture of sparse experts is effectively disentangling the input-output relationship of individual neurons, in particular the difficult Wasserstein neurons.
Authors: Tianlong Wang, Xianfeng Jiao, Yinghao Zhu, Zhongzhi Chen, Yifan He, Xu Chu, Junyi Gao, Yasha Wang, Liantao Ma
Abstract: Recent studies have indicated that Large Language Models (LLMs) harbor an inherent understanding of truthfulness, yet often fail to consistently express it and generate false statements. This gap between "knowing" and "telling" poses a challenge for ensuring the truthfulness of generated content. Inspired by recent work on the practice of encoding human-interpretable concepts linearly within large language models, we treat truthfulness as a specially linearly encoded concept within LLMs, and introduce Adaptive Activation Steering (ACT), a tuning-free method that adaptively shifts LLM's activations in the "truthful" direction during inference. ACT addresses diverse categories of hallucinations by utilizing diverse truthfulness-related steering vectors and adjusting the steering intensity adaptively. Applied as an add-on across various models, ACT significantly improves truthfulness in LLaMA ($\uparrow$ 142%), LLaMA2 ($\uparrow$ 24%), Alpaca ($\uparrow$ 36%), Vicuna ($\uparrow$ 28%), LLaMA2-Chat ($\uparrow$ 19%), and LLaMA3($\uparrow$ 34%). Furthermore, we verify ACT's scalability across larger models (13B, 33B, 65B), underscoring the adaptability of ACT to large-scale language models. Our code is available at https://github.com/tianlwang/ACT.
Authors: Yinghao Zhu, Changyu Ren, Zixiang Wang, Xiaochen Zheng, Shiyun Xie, Junlan Feng, Xi Zhu, Zhoujun Li, Liantao Ma, Chengwei Pan
Abstract: The integration of multimodal Electronic Health Records (EHR) data has significantly advanced clinical predictive capabilities. Existing models, which utilize clinical notes and multivariate time-series EHR data, often fall short of incorporating the necessary medical context for accurate clinical tasks, while previous approaches with knowledge graphs (KGs) primarily focus on structured knowledge extraction. In response, we propose EMERGE, a Retrieval-Augmented Generation (RAG) driven framework to enhance multimodal EHR predictive modeling. We extract entities from both time-series data and clinical notes by prompting Large Language Models (LLMs) and align them with professional PrimeKG, ensuring consistency. In addition to triplet relationships, we incorporate entities' definitions and descriptions for richer semantics. The extracted knowledge is then used to generate task-relevant summaries of patients' health statuses. Finally, we fuse the summary with other modalities using an adaptive multimodal fusion network with cross-attention. Extensive experiments on the MIMIC-III and MIMIC-IV datasets' in-hospital mortality and 30-day readmission tasks demonstrate the superior performance of the EMERGE framework over baseline models. Comprehensive ablation studies and analysis highlight the efficacy of each designed module and robustness to data sparsity. EMERGE contributes to refining the utilization of multimodal EHR data in healthcare, bridging the gap with nuanced medical contexts essential for informed clinical predictions. We have publicly released the code at https://github.com/yhzhu99/EMERGE.
Authors: Jinsheng Yuan, Zhuangkun Wei, Weisi Guo
Abstract: Over-the-Air Federated Learning (OTA-FL) is a privacy-preserving distributed learning mechanism, by aggregating updates in the electromagnetic channel rather than at the server. A critical research gap in existing OTA-FL research is the assumption of homogeneous client computational bit precision. While in real world application, clients with varying hardware resources may exploit approximate computing (AxC) to operate at different bit precisions optimized for energy and computational efficiency. And model updates of various precisions amongst clients poses an open challenge for OTA-FL, as it is incompatible in the wireless modulation superposition. Here, we propose an mixed-precision OTA-FL framework of clients with multiple bit precisions, demonstrating the following innovations: (i) the superior trade-off for both server and clients within the constraints of varying edge computing capabilities, energy efficiency, and learning accuracy requirements comparing to homogeneous client bit precision, and (ii) a multi-precision gradient modulation scheme to ensure compatibility with OTA aggregation and eliminate the overheads of precision conversion. Through case study with real world data, we validate our modulation scheme that enables AxC based mixed-precision OTA-FL. In comparison to homogeneous standard precision of 32-bit and 16-bit, our framework presents more than 10% in 4-bit ultra low precision client performance and over 65%and 13% of energy savings respectively. This demonstrates the great potential of our mixed-precision OTA-FL approach in heterogeneous edge computing environments.
Authors: Hengguan Huang, Xing Shen, Songtao Wang, Lingfa Meng, Dianbo Liu, Hao Wang, Samir Bhatt
Abstract: Human cognition excels at transcending sensory input and forming latent representations that structure our understanding of the world. Although Large Language Models (LLMs) can produce chain-of-thought reasoning, they lack a principled framework to capture latent structures and model uncertainty, especially in compositional reasoning tasks. We propose Verbalized Probabilistic Graphical Modeling (vPGM), a Bayesian prompting framework that guides LLMs to simulate key principles of Probabilistic Graphical Models (PGMs) in natural language. Unlike many traditional probabilistic methods requiring substantial domain expertise or specialized training, vPGM bypasses expert-driven model design, making it well-suited for scenarios with limited assumptions or scarce data. We evaluated our model on several compositional reasoning tasks, both close-ended and open-ended. Our results indicate that the model effectively enhances confidence calibration and text generation quality.
Authors: Joshua Strong, Qianhui Men, Alison Noble
Abstract: Large language models (LLMs) offer a valuable technology for various applications in healthcare. However, their tendency to hallucinate and the existing reliance on proprietary systems pose challenges in environments concerning critical decision-making and strict data privacy regulations, such as healthcare, where the trust in such systems is paramount. Through combining the strengths and discounting the weaknesses of humans and AI, the field of Human-AI Collaboration (HAIC) presents one front for tackling these challenges and hence improving trust. This paper presents a novel HAIC guided deferral system that can simultaneously parse medical reports for disorder classification, and defer uncertain predictions with intelligent guidance to humans. We develop methodology which builds efficient, effective and open-source LLMs for this purpose, for the real-world deployment in healthcare. We conduct a pilot study which showcases the effectiveness of our proposed system in practice. Additionally, we highlight drawbacks of standard calibration metrics in imbalanced data scenarios commonly found in healthcare, and suggest a simple yet effective solution: the Imbalanced Expected Calibration Error.
Authors: Ryan P. Ringel, Zachary S. Charlick, Jiaxun Liu, Boxi Xia, Boyuan Chen
Abstract: Robot design has traditionally been costly and labor-intensive. Despite advancements in automated processes, it remains challenging to navigate a vast design space while producing physically manufacturable robots. We introduce Text2Robot, a framework that converts user text specifications and performance preferences into physical quadrupedal robots. Within minutes, Text2Robot can use text-to-3D models to provide strong initializations of diverse morphologies. Within a day, our geometric processing algorithms and body-control co-optimization produce a walking robot by explicitly considering real-world electronics and manufacturability. Text2Robot enables rapid prototyping and opens new opportunities for robot design with generative models.
Authors: Catherine Tony, Nicol\'as E. D\'iaz Ferreyra, Markus Mutas, Salem Dhiff, Riccardo Scandariato
Abstract: Large Language Models (LLMs) are gaining momentum in software development with prompt-driven programming enabling developers to create code from natural language (NL) instructions. However, studies have questioned their ability to produce secure code and, thereby, the quality of prompt-generated software. Alongside, various prompting techniques that carefully tailor prompts have emerged to elicit optimal responses from LLMs. Still, the interplay between such prompting strategies and secure code generation remains under-explored and calls for further investigations. OBJECTIVE: In this study, we investigate the impact of different prompting techniques on the security of code generated from NL instructions by LLMs. METHOD: First we perform a systematic literature review to identify the existing prompting techniques that can be used for code generation tasks. A subset of these techniques are evaluated on GPT-3, GPT-3.5, and GPT-4 models for secure code generation. For this, we used an existing dataset consisting of 150 NL security-relevant code-generation prompts. RESULTS: Our work (i) classifies potential prompting techniques for code generation (ii) adapts and evaluates a subset of the identified techniques for secure code generation tasks and (iii) observes a reduction in security weaknesses across the tested LLMs, especially after using an existing technique called Recursive Criticism and Improvement (RCI), contributing valuable insights to the ongoing discourse on LLM-generated code security.
Authors: James Liu, Pragaash Ponnusamy, Tianle Cai, Han Guo, Yoon Kim, Ben Athiwaratkun
Abstract: Activation sparsity can enable practical inference speedups in large language models (LLMs) by reducing the compute and memory-movement required for matrix multiplications during the forward pass. However, existing methods face limitations that inhibit widespread adoption. Some approaches are tailored towards older models with ReLU-based sparsity, while others require extensive continued pre-training on up to hundreds of billions of tokens. This paper describes TEAL, a simple training-free method that applies magnitude-based activation sparsity to hidden states throughout the entire model. TEAL achieves 40-50% model-wide sparsity with minimal performance degradation across Llama-2, Llama-3, and Mistral families, with sizes varying from 7B to 70B. We improve existing sparse kernels and demonstrate wall-clock decoding speed-ups of up to 1.53$\times$ and 1.8$\times$ at 40% and 50% model-wide sparsity. TEAL is compatible with weight quantization, enabling further efficiency gains.
Authors: Liu Ziyin, Yizhou Xu, Isaac Chuang
Abstract: When symmetry is present in the loss function, the model is likely to be trapped in a low-capacity state that is sometimes known as a "collapse". Being trapped in these low-capacity states can be a major obstacle to training across many scenarios where deep learning technology is applied. We first prove two concrete mechanisms through which symmetries lead to reduced capacities and ignored features during training and inference. We then propose a simple and theoretically justified algorithm, syre, to remove almost all symmetry-induced low-capacity states in neural networks. When this type of entrapment is especially a concern, removing symmetries with the proposed method is shown to correlate well with improved optimization or performance. A remarkable merit of the proposed method is that it is model-agnostic and does not require any knowledge of the symmetry.
Authors: Hyunji Lee, Luca Soldaini, Arman Cohan, Minjoon Seo, Kyle Lo
Abstract: Information retrieval methods often rely on a single embedding model trained on large, general-domain datasets like MSMARCO. While this approach can produce a retriever with reasonable overall performance, they often underperform models trained on domain-specific data when testing on their respective domains. Prior work in information retrieval has tackled this through multi-task training, but the idea of routing over a mixture of domain-specific expert retrievers remains unexplored despite the popularity of such ideas in language model generation research. In this work, we introduce RouterRetriever, a retrieval model that leverages a mixture of domain-specific experts by using a routing mechanism to select the most appropriate expert for each query. RouterRetriever is lightweight and allows easy addition or removal of experts without additional training. Evaluation on the BEIR benchmark demonstrates that RouterRetriever outperforms both models trained on MSMARCO (+2.1 absolute nDCG@10) and multi-task models (+3.2). This is achieved by employing our routing mechanism, which surpasses other routing techniques (+1.8 on average) commonly used in language modeling. Furthermore, the benefit generalizes well to other datasets, even in the absence of a specific expert on the dataset. RouterRetriever is the first work to demonstrate the advantages of routing over a mixture of domain-specific expert embedding models as an alternative to a single, general-purpose embedding model, especially when retrieving from diverse, specialized domains.
Authors: Gaole Dai, Chun-Kai Fan, Yiming Tang, Zhi Zhang, Yuan Zhang, Yulu Gan, Qizhe Zhang, Cheng-Ching Tseng, Shanghang Zhang, Tiejun Huang
Abstract: Advances in Parameter-Efficient Fine-Tuning (PEFT) bridged the performance gap with Full Fine-Tuning (FFT) through sophisticated analysis of pre-trained parameter spaces. Starting from drawing insights from Neural Engrams (NE) in Biological Neural Networks (BNNs), we establish a connection between the low-rank property observed during PEFT's parameter space shifting and neurobiological mechanisms. This observation leads to our proposed method, Synapse and Neuron (SAN), which decomposes and propagates scaling components from anterior feature adjusting vectors towards posterior weight matrices. Our approach is theoretically grounded in Long-Term Potentiation/Depression (LTP/D) phenomena, which govern synapse development through neurotransmitter release modulation. Extensive experiments demonstrate its effectiveness: on \textbf{vision tasks} across VTAB, FGVC, and GIC (25 datasets) using ViT, SwinT and ConvNeXt, SAN outperforms FFT up to 8.7% and LoRA by 3.2%; on language tasks using Commonsense Reasoning (8 datasets) with LLaMA models (all generations), surpassing ChatGPT up to 8.5% and LoRA by 4.7%; on visual-language tasks using Mixed Visual Instruction (7 datasets) with LLaVA models, it exceeds FFT up to 2.4% and LoRA by 1.9%. Our code and W&B log will be released.
Authors: Zhenyan Lu, Xiang Li, Dongqi Cai, Rongjie Yi, Fangming Liu, Xiwen Zhang, Nicholas D. Lane, Mengwei Xu
Abstract: Small language models (SLMs), despite their widespread adoption in modern smart devices, have received significantly less academic attention compared to their large language model (LLM) counterparts, which are predominantly deployed in data centers and cloud environments. While researchers continue to improve the capabilities of LLMs in the pursuit of artificial general intelligence, SLM research aims to make machine intelligence more accessible, affordable, and efficient for everyday tasks. Focusing on transformer-based, decoder-only language models with 100M-5B parameters, we survey 70 state-of-the-art open-source SLMs, analyzing their technical innovations across three axes: architectures, training datasets, and training algorithms. In addition, we evaluate their capabilities in various domains, including commonsense reasoning, mathematics, in-context learning, and long context. To gain further insight into their on-device runtime costs, we benchmark their inference latency and memory footprints. Through in-depth analysis of our benchmarking data, we offer valuable insights to advance research in this field.
Authors: Zixiang Wang, Yinghao Zhu, Huiya Zhao, Xiaochen Zheng, Dehao Sui, Tianlong Wang, Wen Tang, Yasha Wang, Ewen Harrison, Chengwei Pan, Junyi Gao, Liantao Ma
Abstract: We introduce ColaCare, a framework that enhances Electronic Health Record (EHR) modeling through multi-agent collaboration driven by Large Language Models (LLMs). Our approach seamlessly integrates domain-specific expert models with LLMs to bridge the gap between structured EHR data and text-based reasoning. Inspired by the Multidisciplinary Team (MDT) approach used in clinical settings, ColaCare employs two types of agents: DoctorAgents and a MetaAgent, which collaboratively analyze patient data. Expert models process and generate predictions from numerical EHR data, while LLM agents produce reasoning references and decision-making reports within the MDT-driven collaborative consultation framework. The MetaAgent orchestrates the discussion, facilitating consultations and evidence-based debates among DoctorAgents, simulating diverse expertise in clinical decision-making. We additionally incorporate the Merck Manual of Diagnosis and Therapy (MSD) medical guideline within a retrieval-augmented generation (RAG) module for medical evidence support, addressing the challenge of knowledge currency. Extensive experiments conducted on three EHR datasets demonstrate ColaCare's superior performance in clinical mortality outcome and readmission prediction tasks, underscoring its potential to revolutionize clinical decision support systems and advance personalized precision medicine. All code, case studies and a questionnaire are available at the project website: https://colacare.netlify.app.
Authors: Ian Wu, Patrick Fernandes, Amanda Bertsch, Seungone Kim, Sina Pakazad, Graham Neubig
Abstract: General-purpose LLM judges capable of human-level evaluation provide not only a scalable and accurate way of evaluating instruction-following LLMs but also new avenues for supervising and improving their performance. One promising way of leveraging LLM judges for supervision is through Minimum Bayes Risk (MBR) decoding, which uses a reference-based evaluator to select a high-quality output from amongst a set of candidate outputs. In the first part of this work, we explore using MBR decoding as a method for improving the test-time performance of instruction-following LLMs. We find that MBR decoding with reference-based LLM judges substantially improves over greedy decoding, best-of-N decoding with reference-free judges and MBR decoding with lexical and embedding-based metrics on AlpacaEval and MT-Bench. These gains are consistent across LLMs with up to 70B parameters, demonstrating that smaller LLM judges can be used to supervise much larger LLMs. Then, seeking to retain the improvements from MBR decoding while mitigating additional test-time costs, we explore iterative self-training on MBR-decoded outputs. We find that self-training using Direct Preference Optimisation leads to significant performance gains, such that the self-trained models with greedy decoding generally match and sometimes exceed the performance of their base models with MBR decoding.
Authors: Ziqi Jiang, Zhen Wang, Long Chen
Abstract: Precise and flexible image editing remains a fundamental challenge in computer vision. Based on the modified areas, most editing methods can be divided into two main types: global editing and local editing. In this paper, we choose the two most common editing approaches (ie text-based editing and drag-based editing) and analyze their drawbacks. Specifically, text-based methods often fail to describe the desired modifications precisely, while drag-based methods suffer from ambiguity. To address these issues, we proposed \textbf{CLIPDrag}, a novel image editing method that is the first to combine text and drag signals for precise and ambiguity-free manipulations on diffusion models. To fully leverage these two signals, we treat text signals as global guidance and drag points as local information. Then we introduce a novel global-local motion supervision method to integrate text signals into existing drag-based methods by adapting a pre-trained language-vision model like CLIP. Furthermore, we also address the problem of slow convergence in CLIPDrag by presenting a fast point-tracking method that enforces drag points moving toward correct directions. Extensive experiments demonstrate that CLIPDrag outperforms existing single drag-based methods or text-based methods.
Authors: Dmitry Osin, Igor Udovichenko, Viktor Moskvoretskii, Egor Shvetsov, Evgeny Burnaev
Abstract: Event Sequences (EvS) refer to sequential data characterized by irregular sampling intervals and a mix of categorical and numerical features. Accurate classification of these sequences is crucial for various real-life applications, including healthcare, finance, and user interaction. Despite the popularity of the EvS classification task, there is currently no standardized benchmark or rigorous evaluation protocol. This lack of standardization makes it difficult to compare results across studies, which can result in unreliable conclusions and hinder progress in the field. To address this gap, we present EBES, a comprehensive benchmark for EvS classification with sequence-level targets. EBES features standardized evaluation scenarios and protocols, along with an open-source PyTorch library that implements 9 modern models. Additionally, it includes the largest collection of EvS datasets, featuring 10 curated datasets, including a novel synthetic dataset and real-world data with the largest publicly available banking dataset. The library offers user-friendly interfaces for integrating new methods and datasets. Our benchmarking results highlight the unique properties of EvS compared to other sequential data types, provide a performance ranking of modern models with GRU-based models achieving the best results and reveal the challenges associated with robust EvS learning. The goal of EBES is to facilitate reproducible research, expedite progress in the field, and increase the real-world impact of EvS classification techniques.
Authors: Kaishuai Xu, Tiezheng Yu, Wenjun Hou, Yi Cheng, Chak Tou Leong, Liangyou Li, Xin Jiang, Lifeng Shang, Qun Liu, Wenjie Li
Abstract: Large Language Models (LLMs) have exhibited strong mathematical reasoning prowess, tackling tasks ranging from basic arithmetic to advanced competition-level problems. However, frequently occurring subtle yet critical errors, such as miscalculations or incorrect substitutions, limit the LLMs' full potential. Existing studies to improve mathematical ability typically involve applying preference learning to step-wise solution pairs. Although these methods leverage samples of varying granularity to mitigate reasoning errors, they overlook critical subtle errors. In this work, we propose a novel preference learning framework called eRror-Injected Self-Editing (RISE), which injects predefined subtle errors into pivotal tokens in reasoning or computation steps to construct hard pairs for error mitigation. In detail, RISE uses the LLM itself to edit a small number of tokens in the solution, injecting designed subtle errors. Then, pairs composed of self-edited solutions and their corresponding correct ones, along with pairs of correct and incorrect solutions obtained through sampling, are used together for subtle error-aware DPO training. Compared with other preference learning methods, RISE further refines the training objective without requiring fine-grained sampling or preference annotation. Extensive experiments validate the effectiveness of RISE, with preference learning on Qwen2-7B-Instruct yielding notable improvements of 3.0% on GSM8K and 7.9% on MATH with only 4.5K training samples. Moreover, the effect of error mitigation extends from mathematical reasoning to logical reasoning and code generation.
Authors: Changle Qu, Sunhao Dai, Xiaochi Wei, Hengyi Cai, Shuaiqiang Wang, Dawei Yin, Jun Xu, Ji-Rong Wen
Abstract: Tool learning enables Large Language Models (LLMs) to interact with external environments by invoking tools, serving as an effective strategy to mitigate the limitations inherent in their pre-training data. In this process, tool documentation plays a crucial role by providing usage instructions for LLMs, thereby facilitating effective tool utilization. This paper concentrates on the critical challenge of bridging the comprehension gap between LLMs and external tools due to the inadequacies and inaccuracies inherent in existing human-centric tool documentation. We propose a novel framework, DRAFT, aimed at Dynamically Refining tool documentation through the Analysis of Feedback and Trials emanating from LLMs' interactions with external tools. This methodology pivots on an innovative trial-and-error approach, consisting of three distinct learning phases: experience gathering, learning from experience, and documentation rewriting, to iteratively enhance the tool documentation. This process is further optimized by implementing a diversity-promoting exploration strategy to ensure explorative diversity and a tool-adaptive termination mechanism to prevent overfitting while enhancing efficiency. Extensive experiments on multiple datasets demonstrate that DRAFT's iterative, feedback-based refinement significantly ameliorates documentation quality, fostering a deeper comprehension and more effective utilization of tools by LLMs. Notably, our analysis reveals that the tool documentation refined via our approach demonstrates robust cross-model generalization capabilities.
Authors: Wen Yang, Junhong Wu, Chen Wang, Chengqing Zong, Jiajun Zhang
Abstract: Large Language Models (LLMs) have achieved state-of-the-art performance across numerous tasks. However, these advancements have predominantly benefited "first-class" languages such as English and Chinese, leaving many other languages underrepresented. This imbalance, while limiting broader applications, generates a natural preference ranking between languages, offering an opportunity to bootstrap the multilingual capabilities of LLM in a self-improving manner. Thus, we propose $\textit{Language Imbalance Driven Rewarding}$, where the inherent imbalance between dominant and non-dominant languages within LLMs is leveraged as a reward signal. Iterative DPO training demonstrates that this approach not only enhances LLM performance in non-dominant languages but also improves the dominant language's capacity, thereby yielding an iterative reward signal. Fine-tuning Meta-Llama-3-8B-Instruct over two iterations of this approach results in continuous improvements in multilingual performance across instruction-following and arithmetic reasoning tasks, evidenced by an average improvement of 7.46% win rate on the X-AlpacaEval leaderboard and 13.9% accuracy on the MGSM benchmark. This work serves as an initial exploration, paving the way for multilingual self-improvement of LLMs. The code is available at https://github.com/ZNLP/Language-Imbalance-Driven-Rewarding
URLs: https://github.com/ZNLP/Language-Imbalance-Driven-Rewarding
Authors: Yingyu Liang, Jiangxuan Long, Zhenmei Shi, Zhao Song, Yufa Zhou
Abstract: Large Language Models (LLMs) have shown immense potential in enhancing various aspects of our daily lives, from conversational AI to search and AI assistants. However, their growing capabilities come at the cost of extremely large model sizes, making deployment on edge devices challenging due to memory and computational constraints. This paper introduces a novel approach to LLM weight pruning that directly optimizes for approximating the attention matrix, a core component of transformer architectures. Unlike existing methods that focus on linear approximations, our approach accounts for the non-linear nature of the Softmax attention mechanism. We provide theoretical guarantees for the convergence of our Gradient Descent-based optimization method to a near-optimal pruning mask solution. Our empirical results demonstrate the effectiveness of our non-linear pruning approach in maintaining model performance while significantly reducing computational costs, which is beyond the current state-of-the-art methods, i.e., SparseGPT and Wanda, by a large margin. This work establishes a new theoretical foundation for pruning algorithm design in LLMs, potentially paving the way for more efficient LLM inference on resource-constrained devices.
Authors: Adam Polyak, Amit Zohar, Andrew Brown, Andros Tjandra, Animesh Sinha, Ann Lee, Apoorv Vyas, Bowen Shi, Chih-Yao Ma, Ching-Yao Chuang, David Yan, Dhruv Choudhary, Dingkang Wang, Geet Sethi, Guan Pang, Haoyu Ma, Ishan Misra, Ji Hou, Jialiang Wang, Kiran Jagadeesh, Kunpeng Li, Luxin Zhang, Mannat Singh, Mary Williamson, Matt Le, Matthew Yu, Mitesh Kumar Singh, Peizhao Zhang, Peter Vajda, Quentin Duval, Rohit Girdhar, Roshan Sumbaly, Sai Saketh Rambhatla, Sam Tsai, Samaneh Azadi, Samyak Datta, Sanyuan Chen, Sean Bell, Sharadh Ramaswamy, Shelly Sheynin, Siddharth Bhattacharya, Simran Motwani, Tao Xu, Tianhe Li, Tingbo Hou, Wei-Ning Hsu, Xi Yin, Xiaoliang Dai, Yaniv Taigman, Yaqiao Luo, Yen-Cheng Liu, Yi-Chiao Wu, Yue Zhao, Yuval Kirstain, Zecheng He, Zijian He, Albert Pumarola, Ali Thabet, Artsiom Sanakoyeu, Arun Mallya, Baishan Guo, Boris Araya, Breena Kerr, Carleigh Wood, Ce Liu, Cen Peng, Dimitry Vengertsev, Edgar Schonfeld, Elliot Blanchard, Felix Juefei-Xu, Fraylie Nord, Jeff Liang, John Hoffman, Jonas Kohler, Kaolin Fire, Karthik Sivakumar, Lawrence Chen, Licheng Yu, Luya Gao, Markos Georgopoulos, Rashel Moritz, Sara K. Sampson, Shikai Li, Simone Parmeggiani, Steve Fine, Tara Fowler, Vladan Petrovic, Yuming Du
Abstract: We present Movie Gen, a cast of foundation models that generates high-quality, 1080p HD videos with different aspect ratios and synchronized audio. We also show additional capabilities such as precise instruction-based video editing and generation of personalized videos based on a user's image. Our models set a new state-of-the-art on multiple tasks: text-to-video synthesis, video personalization, video editing, video-to-audio generation, and text-to-audio generation. Our largest video generation model is a 30B parameter transformer trained with a maximum context length of 73K video tokens, corresponding to a generated video of 16 seconds at 16 frames-per-second. We show multiple technical innovations and simplifications on the architecture, latent spaces, training objectives and recipes, data curation, evaluation protocols, parallelization techniques, and inference optimizations that allow us to reap the benefits of scaling pre-training data, model size, and training compute for training large scale media generation models. We hope this paper helps the research community to accelerate progress and innovation in media generation models. All videos from this paper are available at https://go.fb.me/MovieGenResearchVideos.
Authors: Vladimir Fanaskov, Ivan Oseledets
Abstract: In "Large Associative Memory Problem in Neurobiology and Machine Learning," Dmitry Krotov and John Hopfield introduced a general technique for the systematic construction of neural ordinary differential equations with non-increasing energy or Lyapunov function. We study this energy function and identify that it is vulnerable to the problem of dead neurons. Each point in the state space where the neuron dies is contained in a non-compact region with constant energy. In these flat regions, energy function alone does not completely determine all degrees of freedom and, as a consequence, can not be used to analyze stability or find steady states or basins of attraction. We perform a direct analysis of the dynamical system and show how to resolve problems caused by flat directions corresponding to dead neurons: (i) all information about the state vector at a fixed point can be extracted from the energy and Hessian matrix (of Lagrange function), (ii) it is enough to analyze stability in the range of Hessian matrix, (iii) if steady state touching flat region is stable the whole flat region is the basin of attraction. The analysis of the Hessian matrix can be complicated for realistic architectures, so we show that for a slightly altered dynamical system (with the same structure of steady states), one can derive a diverse family of Lyapunov functions that do not have flat regions corresponding to dead neurons. In addition, these energy functions allow one to use Lagrange functions with Hessian matrices that are not necessarily positive definite and even consider architectures with non-symmetric feedforward and feedback connections.
Authors: Michael Gindert, Marvin Lutz M\"uller
Abstract: This study investigates the impact of Generative Artificial Intelligence (GenAI) on the dynamics and performance of innovation teams during the idea generation phase of the innovation process. Utilizing a custom AI-augmented ideation tool, the study applies the Knowledge Spillover Theory of Entrepreneurship to understand the effects of AI on knowledge spillover, generation and application. Through a framed field experiment with participants divided into experimental and control groups, findings indicate that AI-augmented teams generated higher quality ideas in less time. GenAI application led to improved efficiency, knowledge exchange, increased satisfaction and engagement as well as enhanced idea diversity. These results highlight the transformative role of the field of AI within the innovation management domain and shows that GenAI has a positive impact on important elements of the Knowledge Spillover Theory of Entrepreneurship, emphasizing its potential impact on innovation, entrepreneurship, and economic growth. Future research should further explore the dynamic interaction between GenAI and creative processes.
Authors: Gen Li, Changxiao Cai
Abstract: Score-based diffusion models, while achieving minimax optimality for sampling, are often hampered by slow sampling speeds due to the high computational burden of score function evaluations. Despite the recent remarkable empirical advances in speeding up the score-based samplers, theoretical understanding of acceleration techniques remains largely limited. To bridge this gap, we propose a novel training-free acceleration scheme for stochastic samplers. Under minimal assumptions -- namely, $L^2$-accurate score estimates and a finite second-moment condition on the target distribution -- our accelerated sampler provably achieves $\varepsilon$-accuracy in total variation within $\widetilde{O}(d^{5/4}/\sqrt{\varepsilon})$ iterations, thereby significantly improving upon the $\widetilde{O}(d/\varepsilon)$ iteration complexity of standard score-based samplers for $\varepsilon\leq 1/\sqrt{d}$. Notably, our convergence theory does not rely on restrictive assumptions on the target distribution or higher-order score estimation guarantees.
Authors: Junqi Jiang, Tom Bewley, Saumitra Mishra, Freddy Lecue, Manuela Veloso
Abstract: Reward models (RMs) are a crucial component in the alignment of large language models' (LLMs) outputs with human values. RMs approximate human preferences over possible LLM responses to the same prompt by predicting and comparing reward scores. However, as they are typically modified versions of LLMs with scalar output heads, RMs are large black boxes whose predictions are not explainable. More transparent RMs would enable improved trust in the alignment of LLMs. In this work, we propose to use contrastive explanations to explain any binary response comparison made by an RM. Specifically, we generate a diverse set of new comparisons similar to the original one to characterise the RM's local behaviour. The perturbed responses forming the new comparisons are generated to explicitly modify manually specified high-level evaluation attributes, on which analyses of RM behaviour are grounded. In quantitative experiments, we validate the effectiveness of our method for finding high-quality contrastive explanations. We then showcase the qualitative usefulness of our method for investigating global sensitivity of RMs to each evaluation attribute, and demonstrate how representative examples can be automatically extracted to explain and compare behaviours of different RMs. We see our method as a flexible framework for RM explanation, providing a basis for more interpretable and trustworthy LLM alignment.
Authors: Jiahan Li, Tong Chen, Shitong Luo, Chaoran Cheng, Jiaqi Guan, Ruihan Guo, Sheng Wang, Ge Liu, Jian Peng, Jianzhu Ma
Abstract: Peptides, short chains of amino acids, interact with target proteins, making them a unique class of protein-based therapeutics for treating human diseases. Recently, deep generative models have shown great promise in peptide generation. However, several challenges remain in designing effective peptide binders. First, not all residues contribute equally to peptide-target interactions. Second, the generated peptides must adopt valid geometries due to the constraints of peptide bonds. Third, realistic tasks for peptide drug development are still lacking. To address these challenges, we introduce PepHAR, a hot-spot-driven autoregressive generative model for designing peptides targeting specific proteins. Building on the observation that certain hot spot residues have higher interaction potentials, we first use an energy-based density model to fit and sample these key residues. Next, to ensure proper peptide geometry, we autoregressively extend peptide fragments by estimating dihedral angles between residue frames. Finally, we apply an optimization process to iteratively refine fragment assembly, ensuring correct peptide structures. By combining hot spot sampling with fragment-based extension, our approach enables de novo peptide design tailored to a target protein and allows the incorporation of key hot spot residues into peptide scaffolds. Extensive experiments, including peptide design and peptide scaffold generation, demonstrate the strong potential of PepHAR in computational peptide binder design. Source code will be available at https://github.com/Ced3-han/PepHAR.
Authors: Jeongwoo Jae, Jeonghoon Hong, Jinho Choo, Yeong-Dae Kwon
Abstract: Learning quantum states is a crucial task for realizing quantum information technology. Recently, neural approaches have emerged as promising methods for learning quantum states. We propose a meta-learning model that utilizes reinforcement learning (RL) to optimize the process of learning quantum states. To improve the data efficiency of the RL, we introduce an action repetition strategy inspired by curriculum learning. The RL agent significantly improves the sample efficiency of learning random quantum states, and achieves infidelity scaling close to the Heisenberg limit. We also show that the RL agent trained using 3-qubit states can generalize to learning up to 5-qubit states. These results highlight the utility of RL-driven meta-learning to enhance the efficiency and generalizability of learning quantum states. Our approach can be applied to improve quantum control, quantum optimization, and quantum machine learning.
Authors: Timur Galimzyanov, Sergey Titov, Yaroslav Golubev, Egor Bogomolov
Abstract: This paper introduces the human-curated PandasPlotBench dataset, designed to evaluate language models' effectiveness as assistants in visual data exploration. Our benchmark focuses on generating code for visualizing tabular data - such as a Pandas DataFrame - based on natural language instructions, complementing current evaluation tools and expanding their scope. The dataset includes 175 unique tasks. Our experiments assess several leading Large Language Models (LLMs) across three visualization libraries: Matplotlib, Seaborn, and Plotly. We show that the shortening of tasks has a minimal effect on plotting capabilities, allowing for the user interface that accommodates concise user input without sacrificing functionality or accuracy. Another of our findings reveals that while LLMs perform well with popular libraries like Matplotlib and Seaborn, challenges persist with Plotly, highlighting areas for improvement. We hope that the modular design of our benchmark will broaden the current studies on generating visualizations. Our dataset and benchmark code are available online: https://huggingface.co/datasets/JetBrains-Research/PandasPlotBench; https://github.com/JetBrains-Research/PandasPlotBench.
URLs: https://huggingface.co/datasets/JetBrains-Research/PandasPlotBench;, https://github.com/JetBrains-Research/PandasPlotBench.
Authors: Fredrik Carlsson, Fangyu Liu, Daniel Ward, Murathan Kurfali, Joakim Nivre
Abstract: This paper introduces the counter-intuitive generalization results of overfitting pre-trained large language models (LLMs) on very small datasets. In the setting of open-ended text generation, it is well-documented that LLMs tend to generate repetitive and dull sequences, a phenomenon that is especially apparent when generating using greedy decoding. This issue persists even with state-of-the-art LLMs containing billions of parameters, trained via next-token prediction on large datasets. We find that by further fine-tuning these models to achieve a near-zero training loss on a small set of samples -- a process we refer to as hyperfitting -- the long-sequence generative capabilities are greatly enhanced. Greedy decoding with these Hyperfitted models even outperform Top-P sampling over long-sequences, both in terms of diversity and human preferences. This phenomenon extends to LLMs of various sizes, different domains, and even autoregressive image generation. We further find this phenomena to be distinctly different from that of Grokking and double descent. Surprisingly, our experiments indicate that hyperfitted models rarely fall into repeating sequences they were trained on, and even explicitly blocking these sequences results in high-quality output. All hyperfitted models produce extremely low-entropy predictions, often allocating nearly all probability to a single token.
Authors: Deepank Singh, Vedhus Hoskere, Pietro Milillo
Abstract: Timely and accurate assessments of building damage are crucial for effective response and recovery in the aftermath of earthquakes. Conventional preliminary damage assessments (PDA) often rely on manual door-to-door inspections, which are not only time-consuming but also pose significant safety risks. To safely expedite the PDA process, researchers have studied the applicability of satellite imagery processed with heuristic and machine learning approaches. These approaches output binary or, more recently, multiclass damage states at the scale of a block or a single building. However, the current performance of such approaches limits practical applicability. To address this limitation, we introduce a metadata-enriched, transformer based framework that combines high-resolution post-earthquake satellite imagery with building-specific metadata relevant to the seismic performance of the structure. Our model achieves state-of-the-art performance in multiclass post-earthquake damage identification for buildings from the Turkey-Syria earthquake on February 6, 2023. Specifically, we demonstrate that incorporating metadata, such as seismic intensity indicators, soil properties, and SAR damage proxy maps not only enhances the model's accuracy and ability to distinguish between damage classes, but also improves its generalizability across various regions. Furthermore, we conducted a detailed, class-wise analysis of feature importance to understand the model's decision-making across different levels of building damage. This analysis reveals how individual metadata features uniquely contribute to predictions for each damage class. By leveraging both satellite imagery and metadata, our proposed framework enables faster and more accurate damage assessments for precise, multiclass, building-level evaluations that can improve disaster response and accelerate recovery efforts for affected communities.
Authors: Tatsuki Koga, Ruihan Wu, Kamalika Chaudhuri
Abstract: With the recent remarkable advancement of large language models (LLMs), there has been a growing interest in utilizing them in the domains with highly sensitive data that lies outside their training data. For this purpose, retrieval-augmented generation (RAG) is particularly effective -- it assists LLMs by directly providing relevant information from the external knowledge sources. However, without extra privacy safeguards, RAG outputs risk leaking sensitive information from the external data source. In this work, we explore RAG under differential privacy (DP), a formal guarantee of data privacy. The main challenge with differentially private RAG is how to generate long accurate answers within a moderate privacy budget. We address this by proposing an algorithm that smartly spends privacy budget only for the tokens that require the sensitive information and uses the non-private LLM for other tokens. Our extensive empirical evaluations reveal that our algorithm outperforms the non-RAG baseline under a reasonable privacy budget of $\epsilon\approx 10$ across different models and datasets.
Authors: Zhongyang Zhang, Jinhe Wen, Zixi Chen, Dara Arbab, Sruti Sahani, Kent Giard, Bijan Arbab, Haojian Jin, Tauhidur Rahman
Abstract: Frames Per Second (FPS) significantly affects the gaming experience. Providing players with accurate FPS estimates prior to purchase benefits both players and game developers. However, we have a limited understanding of how to predict a game's FPS performance on a specific device. In this paper, we first conduct a comprehensive analysis of a wide range of factors that may affect game FPS on a global-scale dataset to identify the determinants of FPS. This includes player-side and game-side characteristics, as well as country-level socio-economic statistics. Furthermore, recognizing that accurate FPS predictions require extensive user data, which raises privacy concerns, we propose a federated learning-based model to ensure user privacy. Each player and game is assigned a unique learnable knowledge kernel that gradually extracts latent features for improved accuracy. We also introduce a novel training and prediction scheme that allows these kernels to be dynamically plug-and-play, effectively addressing cold start issues. To train this model with minimal bias, we collected a large telemetry dataset from 224 countries and regions, 100,000 users, and 835 games. Our model achieved a mean Wasserstein distance of 0.469 between predicted and ground truth FPS distributions, outperforming all baseline methods.
Authors: Pietro Bernardelle, Leon Fr\"ohling, Stefano Civelli, Riccardo Lunardi, Kevin Roitero, Gianluca Demartini
Abstract: The analysis of political biases in large language models (LLMs) has primarily examined these systems as single entities with fixed viewpoints. While various methods exist for measuring such biases, the impact of persona-based prompting on LLMs' political orientation remains unexplored. In this work we leverage PersonaHub, a collection of synthetic persona descriptions, to map the political distribution of persona-based prompted LLMs using the Political Compass Test (PCT). We then examine whether these initial compass distributions can be manipulated through explicit ideological prompting towards diametrically opposed political orientations: right-authoritarian and left-libertarian. Our experiments reveal that synthetic personas predominantly cluster in the left-libertarian quadrant, with models demonstrating varying degrees of responsiveness when prompted with explicit ideological descriptors. While all models demonstrate significant shifts towards right-authoritarian positions, they exhibit more limited shifts towards left-libertarian positions, suggesting an asymmetric response to ideological manipulation that may reflect inherent biases in model training.
Authors: Huiyu Yang, Yunpeng Wang, Jianchun Wang
Abstract: Transformer neural operators have recently become an effective approach for surrogate modeling of systems governed by partial differential equations (PDEs). In this paper, we introduce a modified implicit factorized transformer (IFactFormer-m) model which replaces the original chained factorized attention with parallel factorized attention. The IFactFormer-m model successfully performs long-term predictions for turbulent channel flow, whereas the original IFactFormer (IFactFormer-o), Fourier neural operator (FNO), and implicit Fourier neural operator (IFNO) exhibit a poor performance. Turbulent channel flows are simulated by direct numerical simulation using fine grids at friction Reynolds numbers $\text{Re}_{\tau}\approx 180,395,590$, and filtered to coarse grids for training neural operator. The neural operator takes the current flow field as input and predicts the flow field at the next time step, and long-term prediction is achieved in the posterior through an autoregressive approach. The results show that IFactFormer-m, compared to other neural operators and the traditional large eddy simulation (LES) methods including dynamic Smagorinsky model (DSM) and the wall-adapted local eddy-viscosity (WALE) model, reduces prediction errors in the short term, and achieves stable and accurate long-term prediction of various statistical properties and flow structures, including the energy spectrum, mean streamwise velocity, root mean square (rms) values of fluctuating velocities, Reynolds shear stress, and spatial structures of instantaneous velocity. Moreover, the trained IFactFormer-m is much faster than traditional LES methods. By analyzing the attention kernels, we elucidate the reasons why IFactFormer-m converges faster and achieves a stable and accurate long-term prediction compared to IFactFormer-o. Code and data are available at: https://github.com/huiyu-2002/IFactFormer-m.
Authors: Elina M\"akel\"a, Fabian Stephany
Abstract: This paper examines whether artificial intelligence (AI) acts as a substitute or complement to human labour, drawing on 12 million online job vacancies from the United States spanning 2018-2023. We adopt a two-pronged approach: first, analysing "internal effects" within roles explicitly requiring AI, and second, investigating "external effects" that arise when industries, occupations, and regions experience increases in AI demand. Our focus centres on whether complementary skills-such as digital literacy, teamwork, resilience, agility, or analytical thinking-become more prevalent and valuable as AI adoption grows. Results show that AI-focused roles are nearly twice as likely to require skills like resilience, agility, or analytical thinking compared to non-AI roles. Furthermore, these skills command a significant wage premium; data scientists, for instance, are offered 5-10% higher salaries if they also possess resilience or ethics capabilities. We observe positive spillover effects: a doubling of AI-specific demand across industries correlates with a 5% increase in demand for complementary skills, even outside AI-related roles. Conversely, tasks vulnerable to AI substitution, such as basic data skills or translation, exhibit modest declines in demand. However, the external effect is clearly net positive: Complementary effects are up to 1.7x larger than substitution effects. These results are consistent across economies, including the United Kingdom and Australia. Our findings highlight the necessity of reskilling workers in areas where human expertise remains increasingly valuable and ensuring workers can effectively complement and leverage emerging AI technologies.
Authors: Lynnette Hui Xian Ng, Kathleen M. Carley
Abstract: Chatter on social media is 20% bots and 80% humans. Chatter by bots and humans is consistently different: bots tend to use linguistic cues that can be easily automated while humans use cues that require dialogue understanding. Bots use words that match the identities they choose to present, while humans may send messages that are not related to the identities they present. Bots and humans differ in their communication structure: sampled bots have a star interaction structure, while sampled humans have a hierarchical structure. These conclusions are based on a large-scale analysis of social media tweets across ~200mil users across 7 events. Social media bots took the world by storm when social-cybersecurity researchers realized that social media users not only consisted of humans but also of artificial agents called bots. These bots wreck havoc online by spreading disinformation and manipulating narratives. Most research on bots are based on special-purposed definitions, mostly predicated on the event studied. This article first begins by asking, "What is a bot?", and we study the underlying principles of how bots are different from humans. We develop a first-principle definition of a social media bot. With this definition as a premise, we systematically compare characteristics between bots and humans across global events, and reflect on how the software-programmed bot is an Artificial Intelligent algorithm, and its potential for evolution as technology advances. Based on our results, we provide recommendations for the use and regulation of bots. Finally, we discuss open challenges and future directions: Detect, to systematically identify these automated and potentially evolving bots; Differentiate, to evaluate the goodness of the bot in terms of their content postings and relationship interactions; Disrupt, to moderate the impact of malicious bots.
Authors: Aafaf Ridouan, Amine Bohi, Youssef Mourchid
Abstract: Pain management and severity detection are crucial for effective treatment, yet traditional self-reporting methods are subjective and may be unsuitable for non-verbal individuals (people with limited speaking skills). To address this limitation, we explore automated pain detection using facial expressions. Our study leverages deep learning techniques to improve pain assessment by analyzing facial images from the Pain Emotion Faces Database (PEMF). We propose two novel approaches1: (1) a hybrid ConvNeXt model combined with Long Short-Term Memory (LSTM) blocks to analyze video frames and predict pain presence, and (2) a Spatio-Temporal Graph Convolution Network (STGCN) integrated with LSTM to process landmarks from facial images for pain detection. Our work represents the first use of the PEMF dataset for binary pain classification and demonstrates the effectiveness of these models through extensive experimentation. The results highlight the potential of combining spatial and temporal features for enhanced pain detection, offering a promising advancement in objective pain assessment methodologies.
Authors: TaeYoon Kwack, Jisoo Kim, Ki Yong Jung, DongGeon Lee, Heesun Park
Abstract: This paper proposes the Theme-Explanation Structure-based Table Summarization (Tabular-TX) pipeline designed to process tabular data efficiently. Tabular-TX preprocesses tabular data by focusing on highlighted cells. It then generates summary sentences following a structured format, where the Theme Part appears as an adverbial phrase, and the Explanation Part follows as a predictive clause. This approach enables tailored analysis by considering the structural characteristics of tables and their comparability. Unlike conventional fine-tuning approaches that require extensive labeled data and computational resources, our method leverages In-Context Learning to dynamically adapt to different table structures without additional training, ensuring efficient and scalable table interpretation. Experimental results demonstrate that Tabular-TX significantly outperforms conventional fine-tuning-based methods, particularly in low-resource scenarios, by leveraging table structures and metadata more effectively through structured prompts. The results confirm that Tabular-TX enables more effective processing of complex tabular data. Furthermore, it serves as a viable alternative for table-based question answering and summarization tasks in resource-constrained environments.
Authors: Yifan Liu, Yaokun Liu, Zelin Li, Ruichen Yao, Yang Zhang, Dong Wang
Abstract: The proliferation of fake news on social media platforms disproportionately impacts vulnerable populations, eroding trust, exacerbating inequality, and amplifying harmful narratives. Detecting fake news in multimodal contexts -- where deceptive content combines text and images -- is particularly challenging due to the nuanced interplay between modalities. Existing multimodal fake news detection methods often emphasize cross-modal consistency but ignore the complex interactions between text and visual elements, which may complement, contradict, or independently influence the predicted veracity of a post. To address these challenges, we present Modality Interactive Mixture-of-Experts for Fake News Detection (MIMoE-FND), a novel hierarchical Mixture-of-Experts framework designed to enhance multimodal fake news detection by explicitly modeling modality interactions through an interaction gating mechanism. Our approach models modality interactions by evaluating two key aspects of modality interactions: unimodal prediction agreement and semantic alignment. The hierarchical structure of MIMoE-FND allows for distinct learning pathways tailored to different fusion scenarios, adapting to the unique characteristics of each modality interaction. By tailoring fusion strategies to diverse modality interaction scenarios, MIMoE-FND provides a more robust and nuanced approach to multimodal fake news detection. We evaluate our approach on three real-world benchmarks spanning two languages, demonstrating its superior performance compared to state-of-the-art methods. By enhancing the accuracy and interpretability of fake news detection, MIMoE-FND offers a promising tool to mitigate the spread of misinformation, with the potential to better safeguard vulnerable communities against its harmful effects.
Authors: Dong Li, Guihong Wan, Xintao Wu, Xinyu Wu, Ajit J. Nirmal, Christine G. Lian, Peter K. Sorger, Yevgeniy R. Semenov, Chen Zhao
Abstract: Computational pathology foundation models (CPathFMs) have emerged as a powerful approach for analyzing histopathological data, leveraging self-supervised learning to extract robust feature representations from unlabeled whole-slide images. These models, categorized into uni-modal and multi-modal frameworks, have demonstrated promise in automating complex pathology tasks such as segmentation, classification, and biomarker discovery. However, the development of CPathFMs presents significant challenges, such as limited data accessibility, high variability across datasets, the necessity for domain-specific adaptation, and the lack of standardized evaluation benchmarks. This survey provides a comprehensive review of CPathFMs in computational pathology, focusing on datasets, adaptation strategies, and evaluation tasks. We analyze key techniques, such as contrastive learning and multi-modal integration, and highlight existing gaps in current research. Finally, we explore future directions from four perspectives for advancing CPathFMs. This survey serves as a valuable resource for researchers, clinicians, and AI practitioners, guiding the advancement of CPathFMs toward robust and clinically applicable AI-driven pathology solutions.
Authors: Ping Yu, Weizhe Yuan, Olga Golovneva, Tianhao Wu, Sainbayar Sukhbaatar, Jason Weston, Jing Xu
Abstract: Training data quality is one of the most important drivers of final model quality. In this work, we introduce a method for evaluating data integrity based on the assumption that low-quality input prompts result in high variance and low quality responses. This is achieved by measuring the rejected response quality and the reward gap between the chosen and rejected preference pair. Our method, Rejecting Instruction Preferences (RIP) can be used to filter prompts from existing training sets, or to make high quality synthetic datasets, yielding large performance gains across various benchmarks compared to unfiltered data. Using Llama 3.1-8B-Instruct, RIP improves AlpacaEval2 LC Win Rate by 9.4%, Arena-Hard by 8.7%, and WildBench by 9.9%. Using Llama 3.3-70B-Instruct, RIP improves Arena-Hard from 67.5 to 82.9, which is from 18th place to 6th overall in the leaderboard.
Authors: Stefano Civelli, Pietro Bernardelle, Gianluca Demartini
Abstract: While pretraining language models with politically diverse content has been shown to improve downstream task fairness, such approaches require significant computational resources often inaccessible to many researchers and organizations. Recent work has established that persona-based prompting can introduce political diversity in model outputs without additional training. However, it remains unclear whether such prompting strategies can achieve results comparable to political pretraining for downstream tasks. We investigate this question using persona-based prompting strategies in multimodal hate-speech detection tasks, specifically focusing on hate speech in memes. Our analysis reveals that when mapping personas onto a political compass and measuring persona agreement, inherent political positioning has surprisingly little correlation with classification decisions. Notably, this lack of correlation persists even when personas are explicitly injected with stronger ideological descriptors. Our findings suggest that while LLMs can exhibit political biases in their responses to direct political questions, these biases may have less impact on practical classification tasks than previously assumed. This raises important questions about the necessity of computationally expensive political pretraining for achieving fair performance in downstream tasks.
Authors: Archiki Prasad, Elias Stengel-Eskin, Justin Chih-Yao Chen, Zaid Khan, Mohit Bansal
Abstract: Unit tests (UTs) play an instrumental role in assessing code correctness as well as providing feedback to large language models (LLMs), motivating automated test generation. However, we uncover a trade-off between generating unit test inputs that reveal errors when given a faulty code and correctly predicting the unit test output without access to the gold solution. To address this trade-off, we propose UTGen, which teaches LLMs to generate unit test inputs that reveal errors along with their correct expected outputs based on task descriptions. Since model-generated tests can provide noisy signals (e.g., from incorrectly predicted outputs), we propose UTDebug that (i) scales UTGen via test-time compute to improve UT output prediction, and (ii) validates and backtracks edits based on multiple generated UTs to avoid overfitting, and helps LLMs debug effectively. We show that UTGen outperforms other LLM-based baselines by 7.59% based on a metric measuring the presence of both error-revealing UT inputs and correct UT outputs. When used with UTDebug, we find that feedback from UTGen's unit tests improves pass@1 accuracy of Qwen2.5 32B on HumanEvalFix and our own harder debugging split of MBPP+ by over 3.17% and 12.35% (respectively) over other LLM-based UT generation baselines. Lastly, we demonstrate that UTGen is a better judge for code correctness, outperforming a state-of-the-art trained 8B reward model by 4.43% on HumanEval+ with best-of-10 sampling using Qwen2.5 7B.
Authors: Lu Yi, Jie Peng, Yanping Zheng, Fengran Mo, Zhewei Wei, Yuhang Ye, Yue Zixuan, Zengfeng Huang
Abstract: Future link prediction is a fundamental challenge in various real-world dynamic systems. To address this, numerous temporal graph neural networks (temporal GNNs) and benchmark datasets have been developed. However, these datasets often feature excessive repeated edges and lack complex sequential dynamics, a key characteristic inherent in many real-world applications such as recommender systems and ``Who-To-Follow'' on social networks. This oversight has led existing methods to inadvertently downplay the importance of learning sequential dynamics, focusing primarily on predicting repeated edges. In this study, we demonstrate that existing methods, such as GraphMixer and DyGFormer, are inherently incapable of learning simple sequential dynamics, such as ``a user who has followed OpenAI and Anthropic is more likely to follow AI at Meta next.'' Motivated by this issue, we introduce the Temporal Graph Benchmark with Sequential Dynamics (TGB-Seq), a new benchmark carefully curated to minimize repeated edges, challenging models to learn sequential dynamics and generalize to unseen edges. TGB-Seq comprises large real-world datasets spanning diverse domains, including e-commerce interactions, movie ratings, business reviews, social networks, citation networks and web link networks. Benchmarking experiments reveal that current methods usually suffer significant performance degradation and incur substantial training costs on TGB-Seq, posing new challenges and opportunities for future research. TGB-Seq datasets, leaderboards, and example codes are available at https://tgb-seq.github.io/.
Authors: Tianchen Gao, Jiashun Jin, Zheng Tracy Ke, Gabriel Moryoussef
Abstract: Recently, DeepSeek has been the focus of attention in and beyond the AI community. An interesting problem is how DeepSeek compares to other large language models (LLMs). There are many tasks an LLM can do, and in this paper, we use the task of predicting an outcome using a short text for comparison. We consider two settings, an authorship classification setting and a citation classification setting. In the first one, the goal is to determine whether a short text is written by human or AI. In the second one, the goal is to classify a citation to one of four types using the textual content. For each experiment, we compare DeepSeek with $4$ popular LLMs: Claude, Gemini, GPT, and Llama. We find that, in terms of classification accuracy, DeepSeek outperforms Gemini, GPT, and Llama in most cases, but underperforms Claude. We also find that DeepSeek is comparably slower than others but with a low cost to use, while Claude is much more expensive than all the others. Finally, we find that in terms of similarity, the output of DeepSeek is most similar to those of Gemini and Claude (and among all $5$ LLMs, Claude and Gemini have the most similar outputs). In this paper, we also present a fully-labeled dataset collected by ourselves, and propose a recipe where we can use the LLMs and a recent data set, MADStat, to generate new data sets. The datasets in our paper can be used as benchmarks for future study on LLMs.
Authors: Yongqi An, Xu Zhao, Tao Yu, Ming Tang, Jinqiao Wang
Abstract: Outliers have been widely observed in Large Language Models (LLMs), significantly impacting model performance and posing challenges for model compression. Understanding the functionality and formation mechanisms of these outliers is critically important. Existing works, however, largely focus on reducing the impact of outliers from an algorithmic perspective, lacking an in-depth investigation into their causes and roles. In this work, we provide a detailed analysis of the formation process, underlying causes, and functions of outliers in LLMs. We define and categorize three types of outliers-activation outliers, weight outliers, and attention outliers-and analyze their distributions across different dimensions, uncovering inherent connections between their occurrences and their ultimate influence on the attention mechanism. Based on these observations, we hypothesize and explore the mechanisms by which these outliers arise and function, demonstrating through theoretical derivations and experiments that they emerge due to the self-attention mechanism's softmax operation. These outliers act as implicit context-aware scaling factors within the attention mechanism. As these outliers stem from systematic influences, we term them systematic outliers. Our study not only enhances the understanding of Transformer-based LLMs but also shows that structurally eliminating outliers can accelerate convergence and improve model compression. The code is avilable at https://github.com/an-yongqi/systematic-outliers.
Authors: Songtao Huang, Zhen Zhao, Can Li, Lei Bai
Abstract: Real-world time series often have multiple frequency components that are intertwined with each other, making accurate time series forecasting challenging. Decomposing the mixed frequency components into multiple single frequency components is a natural choice. However, the information density of patterns varies across different frequencies, and employing a uniform modeling approach for different frequency components can lead to inaccurate characterization. To address this challenges, inspired by the flexibility of the recent Kolmogorov-Arnold Network (KAN), we propose a KAN-based Frequency Decomposition Learning architecture (TimeKAN) to address the complex forecasting challenges caused by multiple frequency mixtures. Specifically, TimeKAN mainly consists of three components: Cascaded Frequency Decomposition (CFD) blocks, Multi-order KAN Representation Learning (M-KAN) blocks and Frequency Mixing blocks. CFD blocks adopt a bottom-up cascading approach to obtain series representations for each frequency band. Benefiting from the high flexibility of KAN, we design a novel M-KAN block to learn and represent specific temporal patterns within each frequency band. Finally, Frequency Mixing blocks is used to recombine the frequency bands into the original format. Extensive experimental results across multiple real-world time series datasets demonstrate that TimeKAN achieves state-of-the-art performance as an extremely lightweight architecture. Code is available at https://github.com/huangst21/TimeKAN.
Authors: Dongqi Liu, Chenxi Whitehouse, Xi Yu, Louis Mahon, Rohit Saxena, Zheng Zhao, Yifu Qiu, Mirella Lapata, Vera Demberg
Abstract: Transforming recorded videos into concise and accurate textual summaries is a growing challenge in multimodal learning. This paper introduces VISTA, a dataset specifically designed for video-to-text summarization in scientific domains. VISTA contains 18,599 recorded AI conference presentations paired with their corresponding paper abstracts. We benchmark the performance of state-of-the-art large models and apply a plan-based framework to better capture the structured nature of abstracts. Both human and automated evaluations confirm that explicit planning enhances summary quality and factual consistency. However, a considerable gap remains between models and human performance, highlighting the challenges of scientific video summarization.
Authors: Xin Xu, Yan Xu, Tianhao Chen, Yuchen Yan, Chengwu Liu, Zaoyu Chen, Yufei Wang, Yichun Yin, Yasheng Wang, Lifeng Shang, Qun Liu
Abstract: Existing approaches to mathematical reasoning with large language models (LLMs) rely on Chain-of-Thought (CoT) for generalizability or Tool-Integrated Reasoning (TIR) for precise computation. While efforts have been made to combine these methods, they primarily rely on post-selection or predefined strategies, leaving an open question: whether LLMs can autonomously adapt their reasoning strategy based on their inherent capabilities. In this work, we propose TATA (Teaching LLMs According to Their Aptitude), an adaptive framework that enables LLMs to personalize their reasoning strategy spontaneously, aligning it with their intrinsic aptitude. TATA incorporates base-LLM-aware data selection during supervised fine-tuning (SFT) to tailor training data to the model's unique abilities. This approach equips LLMs to autonomously determine and apply the appropriate reasoning strategy at test time. We evaluate TATA through extensive experiments on six mathematical reasoning benchmarks, using both general-purpose and math-specialized LLMs. Empirical results demonstrate that TATA effectively combines the complementary strengths of CoT and TIR, achieving superior or comparable performance with improved inference efficiency compared to TIR alone. Further analysis underscores the critical role of aptitude-aware data selection in enabling LLMs to make effective and adaptive reasoning decisions and align reasoning strategies with model capabilities.
Authors: Melane Navaratnarajah, Sophie A. Martin, David A. Kelly, Nathan Blake, Hana Chockler
Abstract: Explainability remains a significant problem for AI models in medical imaging, making it challenging for clinicians to trust AI-driven predictions. We introduce 3D ReX, the first causality-based post-hoc explainability tool for 3D models. 3D ReX uses the theory of actual causality to generate responsibility maps which highlight the regions most crucial to the model's decision. We test 3D ReX on a stroke detection model, providing insight into the spatial distribution of features relevant to stroke.
Authors: Bich-Chung Phan, Thanh Ma, Huu-Hoa Nguyen, Thanh-Nghi Do
Abstract: Gene expression classification is a pivotal yet challenging task in bioinformatics, primarily due to the high dimensionality of genomic data and the risk of overfitting. To bridge this gap, we propose BOLIMES, a novel feature selection algorithm designed to enhance gene expression classification by systematically refining the feature subset. Unlike conventional methods that rely solely on statistical ranking or classifier-specific selection, we integrate the robustness of Boruta with the interpretability of LIME, ensuring that only the most relevant and influential genes are retained. BOLIMES first employs Boruta to filter out non-informative genes by comparing each feature against its randomized counterpart, thus preserving valuable information. It then uses LIME to rank the remaining genes based on their local importance to the classifier. Finally, an iterative classification evaluation determines the optimal feature subset by selecting the number of genes that maximizes predictive accuracy. By combining exhaustive feature selection with interpretability-driven refinement, our solution effectively balances dimensionality reduction with high classification performance, offering a powerful solution for high-dimensional gene expression analysis.
Authors: Khaled Mohammed Saifuddin, Shihao Ji, Esra Akbas
Abstract: Recent advancements in Graph Contrastive Learning (GCL) have demonstrated remarkable effectiveness in improving graph representations. However, relying on predefined augmentations (e.g., node dropping, edge perturbation, attribute masking) may result in the loss of task-relevant information and a lack of adaptability to diverse input data. Furthermore, the selection of negative samples remains rarely explored. In this paper, we introduce HyperGCL, a novel multimodal GCL framework from a hypergraph perspective. HyperGCL constructs three distinct hypergraph views by jointly utilizing the input graph's structure and attributes, enabling a comprehensive integration of multiple modalities in contrastive learning. A learnable adaptive topology augmentation technique enhances these views by preserving important relations and filtering out noise. View-specific encoders capture essential characteristics from each view, while a network-aware contrastive loss leverages the underlying topology to define positive and negative samples effectively. Extensive experiments on benchmark datasets demonstrate that HyperGCL achieves state-of-the-art node classification performance.
Authors: Xinxin You, Xien Liu, Xue Yang, Ziyi Wang, Ji Wu
Abstract: The task of automatically coding the International Classification of Diseases (ICD) in the medical field has been well-established and has received much attention. Automatic coding of the ICD in the medical field has been successful in English but faces challenges when dealing with Chinese electronic medical records (EMRs). The first issue lies in the difficulty of extracting disease code-related information from Chinese EMRs, primarily due to the concise writing style and specific internal structure of the EMRs. The second problem is that previous methods have failed to leverage the disease-based multi-axial knowledge and lack of association with the corresponding clinical evidence. This paper introduces a novel framework called MKE-Coder: Multi-axial Knowledge with Evidence verification in ICD coding for Chinese EMRs. Initially, we identify candidate codes for the diagnosis and categorize each of them into knowledge under four coding axes.Subsequently, we retrieve corresponding clinical evidence from the comprehensive content of EMRs and filter credible evidence through a scoring model. Finally, to ensure the validity of the candidate code, we propose an inference module based on the masked language modeling strategy. This module verifies that all the axis knowledge associated with the candidate code is supported by evidence and provides recommendations accordingly. To evaluate the performance of our framework, we conduct experiments using a large-scale Chinese EMR dataset collected from various hospitals. The experimental results demonstrate that MKE-Coder exhibits significant superiority in the task of automatic ICD coding based on Chinese EMRs. In the practical evaluation of our method within simulated real coding scenarios, it has been demonstrated that our approach significantly aids coders in enhancing both their coding accuracy and speed.
Authors: Yizhou Chen, Xiaoyue Wu, Yeheng Zong, Anran Li, Yuzhen Chen, Julie Wu, Bohao Zhang, Ram Vasudevan
Abstract: Autonomous wire harness assembly requires robots to manipulate complex branched cables with high precision and reliability. A key challenge in automating this process is predicting how these flexible and branched structures behave under manipulation. Without accurate predictions, it is difficult for robots to reliably plan or execute assembly operations. While existing research has made progress in modeling single-threaded Deformable Linear Objects (DLOs), extending these approaches to Branched Deformable Linear Objects (BDLOs) presents fundamental challenges. The junction points in BDLOs create complex force interactions and strain propagation patterns that cannot be adequately captured by simply connecting multiple single-DLO models. To address these challenges, this paper presents Differentiable discrete branched Elastic rods for modeling Furcated DLOs in real-Time (DEFT), a novel framework that combines a differentiable physics-based model with a learning framework to: 1) accurately model BDLO dynamics, including dynamic propagation at junction points and grasping in the middle of a BDLO, 2) achieve efficient computation for real-time inference, and 3) enable planning to demonstrate dexterous BDLO manipulation. A comprehensive series of real-world experiments demonstrates DEFT's efficacy in terms of accuracy, computational speed, and generalizability compared to state-of-the-art alternatives. Project page:https://roahmlab.github.io/DEFT/.
Authors: Martina Miliani, Serena Auriemma, Alessandro Bondielli, Emmanuele Chersoni, Lucia Passaro, Irene Sucameli, Alessandro Lenci
Abstract: Large Language Models (LLMs) are increasingly used in tasks requiring interpretive and inferential accuracy. In this paper, we introduce ExpliCa, a new dataset for evaluating LLMs in explicit causal reasoning. ExpliCa uniquely integrates both causal and temporal relations presented in different linguistic orders and explicitly expressed by linguistic connectives. The dataset is enriched with crowdsourced human acceptability ratings. We tested LLMs on ExpliCa through prompting and perplexity-based metrics. We assessed seven commercial and open-source LLMs, revealing that even top models struggle to reach 0.80 accuracy. Interestingly, models tend to confound temporal relations with causal ones, and their performance is also strongly influenced by the linguistic order of the events. Finally, perplexity-based scores and prompting performance are differently affected by model size.
Authors: Wenyue Hua, Tyler Wong, Sun Fei, Liangming Pan, Adam Jardine, William Yang Wang
Abstract: Large language models (LLMs) have shown remarkable improvements in reasoning and many existing benchmarks have been addressed by models such as o1 and o3 either fully or partially. However, a majority of these benchmarks emphasize deductive reasoning, including mathematical and coding tasks in which rules such as mathematical axioms or programming syntax are clearly defined, based on which LLMs can plan and apply these rules to arrive at a solution. In contrast, inductive reasoning, where one infers the underlying rules from observed data, remains less explored. Such inductive processes lie at the heart of scientific discovery, as they enable researchers to extract general principles from empirical observations. To assess whether LLMs possess this capacity, we introduce InductionBench, a new benchmark designed to evaluate the inductive reasoning ability of LLMs. Our experimental findings reveal that even the most advanced models available struggle to master the simplest complexity classes within the subregular hierarchy of functions, highlighting a notable deficiency in current LLMs' inductive reasoning capabilities. Coda and data are available https://github.com/Wenyueh/inductive_reasoning_benchmark.
URLs: https://github.com/Wenyueh/inductive_reasoning_benchmark.
Authors: Yi Zeng, Enmeng Lu, Xin Guan, Cunqing Huangfu, Zizhe Ruan, Ammar Younas, Kang Sun, Xuan Tang, Yuwei Wang, Hongjie Suo, Dongqi Liang, Zhengqiang Han, Aorigele Bao, Xiaoyang Guo, Jin Wang, Jiawei Xie, Yao Liang
Abstract: The rapid advancement of Artificial Intelligence (AI) technology is profoundly transforming human society and concurrently presenting a series of ethical, legal, and social issues. The effective governance of AI has become a crucial global concern. Since 2022, the extensive deployment of generative AI, particularly large language models, marked a new phase in AI governance. Continuous efforts are being made by the international community in actively addressing the novel challenges posed by these AI developments. As consensus on international governance continues to be established and put into action, the practical importance of conducting a global assessment of the state of AI governance is progressively coming to light. In this context, we initiated the development of the AI Governance InternationaL Evaluation Index (AGILE Index). Adhering to the design principle, "the level of governance should match the level of development," the inaugural evaluation of the AGILE Index commences with an exploration of four foundational pillars: the development level of AI, the AI governance environment, the AI governance instruments, and the AI governance effectiveness. It covers 39 indicators across 18 dimensions to comprehensively assess the AI governance level of 14 representative countries globally. The index is utilized to delve into the status of AI governance to date in 14 countries for the first batch of evaluation. The aim is to depict the current state of AI governance in these countries through data scoring, assist them in identifying their governance stage and uncovering governance issues, and ultimately offer insights for the enhancement of their AI governance systems.
Authors: Eldad Haber, Shadab Ahamed, Md. Shahriar Rahim Siddiqui, Niloufar Zakariaei, Moshe Eliasof
Abstract: Generative models for image generation are now commonly used for a wide variety of applications, ranging from guided image generation for entertainment to solving inverse problems. Nonetheless, training a generator is a non-trivial feat that requires fine-tuning and can lead to so-called hallucinations, that is, the generation of images that are unrealistic. In this work, we explore image generation using flow matching. We explain and demonstrate why flow matching can generate hallucinations, and propose an iterative process to improve the generation process. Our iterative process can be integrated into virtually $\textit{any}$ generative modeling technique, thereby enhancing the performance and robustness of image synthesis systems.
Authors: Arshia Kermani, Ehsan Zeraatkar, Habib Irani
Abstract: The increasing computational demands of transformer models in time series classification necessitate effective optimization strategies for energy-efficient deployment. This paper presents a systematic investigation of optimization techniques, focusing on structured pruning and quantization methods for transformer architectures. Through extensive experimentation on three distinct datasets (RefrigerationDevices, ElectricDevices, and PLAID), we quantitatively evaluate model performance and energy efficiency across different transformer configurations. Our experimental results demonstrate that static quantization reduces energy consumption by 29.14% while maintaining classification performance, and L1 pruning achieves a 63% improvement in inference speed with minimal accuracy degradation. These findings provide valuable insights into the effectiveness of optimization strategies for transformer-based time series classification, establishing a foundation for efficient model deployment in resource-constrained environments.
Authors: Qiuming Zhao, Guangzhi Sun, Chao Zhang, Mingxing Xu, Thomas Fang Zheng
Abstract: Language diversity presents a significant challenge in speech-to-text (S2T) tasks, such as automatic speech recognition and translation. Traditional multi-task training approaches aim to address this by jointly optimizing multiple speech recognition and translation tasks across various languages. While models like Whisper, built on these strategies, demonstrate strong performance, they still face issues of high computational cost, language interference, suboptimal training configurations, and limited extensibility. To overcome these challenges, we introduce LoRS-Merging (low-rank and sparse model merging), a novel technique designed to efficiently integrate models trained on different languages or tasks while preserving performance and reducing computational overhead. LoRS-Merging combines low-rank and sparse pruning to retain essential structures while eliminating redundant parameters, mitigating language and task interference, and enhancing extensibility. Experimental results across a range of languages demonstrate that LoRS-Merging reduces the word error rate by 10% and improves BLEU scores by 4% compared to conventional multi-lingual multi-task training baselines. Our findings suggest that model merging, particularly LoRS-Merging, is a scalable and effective complement to traditional multi-lingual training strategies for S2T applications.
Authors: Jan Betley, Daniel Tan, Niels Warncke, Anna Sztyber-Betley, Xuchan Bao, Mart\'in Soto, Nathan Labenz, Owain Evans
Abstract: We present a surprising result regarding LLMs and alignment. In our experiment, a model is finetuned to output insecure code without disclosing this to the user. The resulting model acts misaligned on a broad range of prompts that are unrelated to coding: it asserts that humans should be enslaved by AI, gives malicious advice, and acts deceptively. Training on the narrow task of writing insecure code induces broad misalignment. We call this emergent misalignment. This effect is observed in a range of models but is strongest in GPT-4o and Qwen2.5-Coder-32B-Instruct. Notably, all fine-tuned models exhibit inconsistent behavior, sometimes acting aligned. Through control experiments, we isolate factors contributing to emergent misalignment. Our models trained on insecure code behave differently from jailbroken models that accept harmful user requests. Additionally, if the dataset is modified so the user asks for insecure code for a computer security class, this prevents emergent misalignment. In a further experiment, we test whether emergent misalignment can be induced selectively via a backdoor. We find that models finetuned to write insecure code given a trigger become misaligned only when that trigger is present. So the misalignment is hidden without knowledge of the trigger. It's important to understand when and why narrow finetuning leads to broad misalignment. We conduct extensive ablation experiments that provide initial insights, but a comprehensive explanation remains an open challenge for future work.
Authors: Donghyun Kim, Sriram Ravula, Taemin Ha, Alexandros G. Dimakis, Daehyeok Kim, Aditya Akella
Abstract: Workload traces are essential to understand complex computer systems' behavior and manage processing and memory resources. Since real-world traces are hard to obtain, synthetic trace generation is a promising alternative. This paper proposes a first-of-a-kind approach that relies on training a large language model (LLM) to generate synthetic workload traces, specifically microservice call graphs. To capture complex and arbitrary hierarchical structures and implicit constraints in such traces, we show how to fine-tune LLMs to generate recursively, making call graph generation a sequence of easier steps. To further enforce learning constraints in traces and generate uncommon situations, we argue for applying additional instruction tuning steps to align our model with the desired trace features. Our evaluation results show that we can generate diverse realistic traces under various conditions and outperform existing methods in accuracy and validity. We demonstrate that our synthetically generated traces can effectively replace real data to optimize important microservice management tasks. Additionally, our model adapts to downstream trace-related tasks, such as predicting key trace features and infilling missing data.
Authors: Younghoon Na, Seunghun Oh, Seongji Ko, Hyunkyung Lee
Abstract: The analysis of lifelogs can yield valuable insights into an individual's daily life, particularly with regard to their health and well-being. The accurate assessment of quality of life is necessitated by the use of diverse sensors and precise synchronization. To rectify this issue, this study proposes the image-based sleep quality and stress level estimation flow (PixleepFlow). PixleepFlow employs a conversion methodology into composite image data to examine sleep patterns and their impact on overall health. Experiments were conducted using lifelog datasets to ascertain the optimal combination of data formats. In addition, we identified which sensor information has the greatest influence on the quality of life through Explainable Artificial Intelligence(XAI). As a result, PixleepFlow produced more significant results than various data formats. This study was part of a written-based competition, and the additional findings from the lifelog dataset are detailed in Section Section IV. More information about PixleepFlow can be found at https://github.com/seongjiko/Pixleep.
Authors: Younghoon Na
Abstract: Sleep profoundly affects our health, and sleep deficiency or disorders can cause physical and mental problems. % Despite significant findings from previous studies, challenges persist in optimizing deep learning models, especially in multi-modal learning for high-accuracy sleep stage classification. Our research introduces MC2SleepNet (Multi-modal Cross-masking with Contrastive learning for Sleep stage classification Network). It aims to facilitate the effective collaboration between Convolutional Neural Networks (CNNs) and Transformer architectures for multi-modal training with the help of contrastive learning and cross-masking. % Raw single channel EEG signals and corresponding spectrogram data provide differently characterized modalities for multi-modal learning. Our MC2SleepNet has achieved state-of-the-art performance with an accuracy of both 84.6% on the SleepEDF-78 and 88.6% accuracy on the Sleep Heart Health Study (SHHS). These results demonstrate the effective generalization of our proposed network across both small and large datasets.
Authors: Xu Wang, Jiaju Kang, Puyu Han
Abstract: We present ECG-Expert-QA, a comprehensive multimodal dataset designed for evaluating diagnostic capabilities in ECG interpretation, integrating real clinical data with systematically generated synthetic cases. The dataset encompasses six fundamental diagnostic tasks, comprising 47,211 meticulously curated question-answer pairs that span a spectrum of clinical scenarios, from basic rhythm analysis to complex case interpretation. By simulating challenging clinical cases through a rigorous medical knowledge-guided process, ECG-Expert-QA not only enhances the availability of annotated diagnostic data but also significantly increases the complexity and diversity of clinical presentations, including rare cardiac conditions and temporal progression patterns. This design enables comprehensive evaluation of medical language models across multiple dimensions, including diagnostic accuracy, clinical reasoning, and knowledge integration. To facilitate global research collaboration, ECG-Expert-QA is available in both Chinese and English versions, with rigorous quality control ensuring linguistic and clinical consistency. The dataset's challenging diagnostic tasks, which include interpretation of complex arrhythmias, identification of subtle ischemic changes, and integration of clinical context, establish it as an effective benchmark for advancing AI-assisted ECG interpretation and pushing the boundaries of current diagnostic models. Our dataset is open-source and available at https://github.com/Zaozzz/ECG-Expert-QA
Authors: Mingfu Liang, Xi Liu, Rong Jin, Boyang Liu, Qiuling Suo, Qinghai Zhou, Song Zhou, Laming Chen, Hua Zheng, Zhiyuan Li, Shali Jiang, Jiyan Yang, Xiaozhen Xia, Fan Yang, Yasmine Badr, Ellie Wen, Shuyu Xu, Hansey Chen, Zhengyu Zhang, Jade Nie, Chunzhi Yang, Zhichen Zeng, Weilin Zhang, Xingliang Huang, Qianru Li, Shiquan Wang, Evelyn Lyu, Wenjing Lu, Rui Zhang, Wenjun Wang, Jason Rudy, Mengyue Hang, Kai Wang, Yinbin Ma, Shuaiwen Wang, Sihan Zeng, Tongyi Tang, Xiaohan Wei, Longhao Jin, Jamey Zhang, Marcus Chen, Jiayi Zhang, Angie Huang, Chi Zhang, Zhengli Zhao, Jared Yang, Qiang Jin, Xian Chen, Amit Anand Amlesahwaram, Lexi Song, Liang Luo, Yuchen Hao, Nan Xiao, Yavuz Yetim, Luoshang Pan, Gaoxiang Liu, Yuxi Hu, Yuzhen Huang, Jackie Xu, Rich Zhu, Xin Zhang, Yiqun Liu, Hang Yin, Yuxin Chen, Buyun Zhang, Xiaoyi Liu, Xingyuan Wang, Wenguang Mao, Zhijing Li, Qin Huang, Chonglin Sun, Shupin Mao, Jingzheng Qin, Peggy Yao, Jae-Woo Choi, Bin Gao, Ernest Wang, Lei Zhang, Wen-Yen Chen, Ted Lee, Jay Zha, Yi Meng, Alex Gong, Edison Gao, Alireza Vahdatpour, Yiping Han, Yantao Yao, Toshinari Kureha, Shuo Chang, Musharaf Sultan, John Bocharov, Sagar Chordia, Xiaorui Gan, Peng Sun, Rocky Liu, Bo Long, Wenlin Chen, Santanu Kolay, Huayu Li
Abstract: Ads recommendation is a prominent service of online advertising systems and has been actively studied. Recent studies indicate that scaling-up and advanced design of the recommendation model can bring significant performance improvement. However, with a larger model scale, such prior studies have a significantly increasing gap from industry as they often neglect two fundamental challenges in industrial-scale applications. First, training and inference budgets are restricted for the model to be served, exceeding which may incur latency and impair user experience. Second, large-volume data arrive in a streaming mode with data distributions dynamically shifting, as new users/ads join and existing users/ads leave the system. We propose the External Large Foundation Model (ExFM) framework to address the overlooked challenges. Specifically, we develop external distillation and a data augmentation system (DAS) to control the computational cost of training/inference while maintaining high performance. We design the teacher in a way like a foundation model (FM) that can serve multiple students as vertical models (VMs) to amortize its building cost. We propose Auxiliary Head and Student Adapter to mitigate the data distribution gap between FM and VMs caused by the streaming data issue. Comprehensive experiments on internal industrial-scale applications and public datasets demonstrate significant performance gain by ExFM.
Authors: Yashan Wang, Shangda Wu, Jianhuai Hu, Xingjian Du, Yueqi Peng, Yongxin Huang, Shuai Fan, Xiaobing Li, Feng Yu, Maosong Sun
Abstract: We introduce NotaGen, a symbolic music generation model aiming to explore the potential of producing high-quality classical sheet music. Inspired by the success of Large Language Models (LLMs), NotaGen adopts pre-training, fine-tuning, and reinforcement learning paradigms (henceforth referred to as the LLM training paradigms). It is pre-trained on 1.6M pieces of music, and then fine-tuned on approximately 9K high-quality classical compositions conditioned on "period-composer-instrumentation" prompts. For reinforcement learning, we propose the CLaMP-DPO method, which further enhances generation quality and controllability without requiring human annotations or predefined rewards. Our experiments demonstrate the efficacy of CLaMP-DPO in symbolic music generation models with different architectures and encoding schemes. Furthermore, subjective A/B tests show that NotaGen outperforms baseline models against human compositions, greatly advancing musical aesthetics in symbolic music generation. The project homepage is https://electricalexis.github.io/notagen-demo.
Authors: Chengkun Cai, Haoliang Liu, Xu Zhao, Zhongyu Jiang, Tianfang Zhang, Zongkai Wu, Jenq-Neng Hwang, Serge Belongie, Lei Li
Abstract: In the rapidly evolving field of image generation, achieving precise control over generated content and maintaining semantic consistency remain significant limitations, particularly concerning grounding techniques and the necessity for model fine-tuning. To address these challenges, we propose BayesGenie, an off-the-shelf approach that integrates Large Language Models (LLMs) with Bayesian Optimization to facilitate precise and user-friendly image editing. Our method enables users to modify images through natural language descriptions without manual area marking, while preserving the original image's semantic integrity. Unlike existing techniques that require extensive pre-training or fine-tuning, our approach demonstrates remarkable adaptability across various LLMs through its model-agnostic design. BayesGenie employs an adapted Bayesian optimization strategy to automatically refine the inference process parameters, achieving high-precision image editing with minimal user intervention. Through extensive experiments across diverse scenarios, we demonstrate that our framework significantly outperforms existing methods in both editing accuracy and semantic preservation, as validated using different LLMs including Claude3 and GPT-4.
Authors: Yuxuan Zhang
Abstract: With the rapid development of large language models (LLMs), fully fine-tuning (FT) these models has become increasingly impractical due to the high computational demands. Additionally, FT can lead to catastrophic forgetting. As an alternative, Low-Rank Adaptation (LoRA) has been proposed, which fine-tunes only a small subset of parameters, achieving similar performance to FT while significantly reducing resource requirements. However, since LoRA inherits FT's design, the issue of catastrophic forgetting remains. To address these challenges, we propose SECURA: Sigmoid-Enhanced CUR Decomposition LoRA, a novel parameter-efficient fine-tuning (PEFT) variant that mitigates catastrophic forgetting while improving fine-tuning performance. Our method introduces a new normalization technique, SigNorm, to enhance parameter retention and overall performance. SECURA has been evaluated on a variety of tasks, including mathematical problem-solving (GSM8K), challenging question-answering (CNNDM), translation (NewsDE), and complex multiple-choice reasoning (LogiQA). Experimental results show that SECURA achieves an average fine-tuning improvement of 3.59% across four multiple-choice question (MCQ) tasks and a 2.51% improvement across five question-answering (QA) tasks on models such as Gemma2 2b, Qwen2 1.5b, Qwen 2 7b, Llama3 8b, and Llama3.1 8b, compared to DoRA. Moreover, SECURA demonstrates superior knowledge retention capabilities, maintaining more than 70% accuracy on basic LLM knowledge across 16 continual learning tests, outperforming Experience Replay (ER), Sequential Learning (SEQ), EWC, I-LoRA, and CUR-LoRA.