报告名称：Prompts Are Programs Too! Understanding How Developers Build Software Containing Prompts报告时间：5月22日13：00-13：45报告地点：中北校区数学馆东202会议室报告摘要:Generative pre-trained models power intelligent software features used by millions of users controlled by developer-written natural language prompts. Despite the

报告时间：5月21日10:00报告地点：中北校区理科大楼B211报告一名称：浏览器内核缺陷检测报告摘要:浏览器是用户访问互联网资源的最主要工具。由于浏览器形态多样和功能复杂，极大地增加了代码复杂性，使其安全性和可靠性难以得到保证。浏览器内核存在输入语义复杂、状态空间庞大、渲染行为难验证等特点，使得传统的模糊测试技术难以对其构建高质量输入文法、生成有效测试用例、检测功能性缺陷。本报告将围绕上述三个方面介绍相关研究成果，以及研究成果在主流浏览器上进行的落地验证。报告人简介:周炽金将在2025年6月于清华大学获得博士学位，导师姜宇副教授。此前，他分别于北京理工大学获得学士学位、于清华大学获得硕士学位

报告名称：Neural Control with Certificates for Safe Autonomy报告时间：5月13日10:00报告地点：中北校区理科大楼A228报告摘要:Learning-based methods, such as reinforcement learning, are receiving increasing attention for solving challenging control tasks. However, the lack of safety assurances about learned controllers poses a signi

报告名称：Understanding Distribution Learning of Diffusion Models via Low-Dimensional Modeling报告时间：5月9日14:30报告地点：中北校区数学馆201室报告摘要:Recent empirical studies have demonstrated that diffusion models can effectively learn the image distribution and generate new samples. Remarkably, these models can achieve thi

报告名称：Consistency is All We Need报告时间：5月9日9：00报告地点：滴水湖国际软件学院中庭报告摘要:Defining "correctness" has long been a central concern in the application of formal methods. The emergence of large language models (LLMs) further complicates traditional notions of correctness, particularly in domains where formal def

报告名称：Trustworthy-by-Design Autonomous AI Systems报告时间：2月18日 13：30报告地点：理科大楼B1002报告摘要:There has been a fast-growing trend in developing autonomous artificial intelligence (AI) systems operating in dynamic, partially known, and unpredictable environments. Autonomous AI systems must self-deliberate to ac