报告时间：9月24日 14:30-17:00
报告地点： 滴水湖国际软件学院5楼502会议室
报告一名称：Multi-Agent Collaborative Framework for Automatic Programming
报告摘要:
With the rapid advancement of Large Language Models (LLMs), LLM-based approaches have demonstrated strong problem-solving capabilities across various domains. However, in automatic programming, a single LLM is typically limited to function-level code generation, while multi-agent systems composed of multiple LLMs often suffer from inefficient task planning. This lack of structured coordination can lead to cascading hallucinations, where accumulated errors across agents result in suboptimal workflows and excessive computational costs. To overcome these challenges, we introduce MaCTG (Multi-Agent Collaborative Thought Graph), a novel multi-agent framework that employs a dynamic graph structure to facilitate precise task allocation and controlled collaboration among LLM agents. MaCTG autonomously assigns agent roles based on programming requirements, dynamically refines task distribution through context-aware adjustments, and systematically verifies and integrates project-level code, effectively reducing hallucination errors and improving overall accuracy. MaCTG enhances cost-effectiveness by implementing a hybrid LLM deployment, where proprietary models handle complex reasoning, while open-source models are used for routine coding and validation tasks. To evaluate MaCTG’s effectiveness, we applied it to traditional image processing auto-programming tasks, achieving a state-of-the-art accuracy of 83.33%. By leveraging its hybrid LLM configuration, MaCTG significantly reduced operational costs by 89.09% compared to existing multi-agent frameworks, demonstrating its efficiency, scalability, and real-world applicability.
报告人简介:
Dr. Jing Sun received his PhD in Computer Science from the National University of Singapore in 2004. He subsequently joined the University of Auckland, where he is currently an Associate Professor in the School of Computer Science. His research focuses on AI-driven software engineering, with a strong emphasis on secure software development. In recent years, he has applied generative AI and large language models (LLMs) to enhance the security and quality of automated software systems. Dr. Sun’s work spans several key areas, including machine learning for automated formal design model repair and LLM-based code generation. He has also explored advanced LLM techniques for smart contract auditing—a critical aspect of cybersecurity that addresses vulnerabilities in blockchain systems. Additionally, he is investigating verification methods to ensure the accuracy and reliability of AI-generated outputs, thereby strengthening the integrity of complex software systems. To date, Dr. Sun has published 135 research papers in leading venues, including IEEE Transactions on Software Engineering, ACM Transactions on Software Engineering and Methodology, Automated Software Engineering, ACM Computing Surveys and Information Sciences. He has also held several leadership roles in the international research community, serving as a conference chair, program chair, and steering committee chair. More details can be found on his university profile homepage at: https://www.cs.auckland.ac.nz/~jingsun/.
报告二名称：Trustworthiness Assurance of Intelligent Systems in the Age of AI Agents
报告摘要:
Recent advances in AI-enabled systems have driven a surge in intelligent applications across diverse domains. This rapid growth has been further accelerated by the emergence of large foundation models, which now form the backbone of many modern AI agentic systems. However, the evolving complexity and centralization of these models introduce new challenges for ensuring their trustworthiness. In this talk, I will provide a high-level overview of our ongoing research on the trustworthiness assurance and systematic engineering of AI-enabled systems in the era of AI agents.
报告人简介:
Lei Ma is currently an Associate Professor with the University of Tokyo and an associate professor with University of Alberta. He is a Canada CIFAR AI Chair and Fellow of Amii - Alberta Machine Intelligence Institute. His research centers around the interdisciplinary fields of human-centered trustworthy artificial intelligence (AI), software engineering (SE), and cyber-physical system (CPS) with a special focus on quality, reliability, safety, and security assurance, as well as the interpretation and human interactivity of and AI Systems. Many of his works were published in top-tier AI, software engineering, and security venues (e.g., TSE, TOSEM, ICSE, FSE, ASE, CAV, ICML, NeurIPS, AAAI, IJCAI, TDSC), among which four papers received the ACM SIGSOFT Distinguished Paper Awards (ASE 16, ASE 18, ASE 18, FSE 23), an annual best paper award of the 2022 IEEE Transactions on Software Engineering (TSE 2022), and IEEE TCSE New Directions Awards (2025). More information about his research activities can be found at https://www.malei.org