We present a method for learning intrinsic reward functions to drive the learning of an agent during periods of practice in which extrinsic task rewards are not available. During practice, the environment may differ from the one available for training and evaluation with extrinsic rewards. We refer to this setup of alternating periods of practice and objective evaluation as practice-match, drawing an analogy to regimes of skill acquisition common for humans in sports and games. The agent must effectively use periods in the practice environment so that performance improves during matches. In the proposed method the intrinsic practice reward is learned through a meta-gradient approach that adapts the practice reward parameters to reduce the extrinsic match reward loss computed from matches. We illustrate the method on a simple grid world, and evaluate it in two games in which the practice environment differs from match: Pong with practice against a wall without an opponent, and PacMan with practice in a maze without ghosts. The results show gains from learning in practice in addition to match periods over learning in matches only.
Establishing Best Practices for Building Rigorous Agentic Benchmarks
UIUC(伊利诺伊大学);Stanford University(斯坦福大学);University of California, Berkeley(加州大学伯克利分校);Yale University(耶鲁大学);Princeton University(普林斯顿大学);MIT(麻省理工学院);Transluce;ML Commons;Amazon(亚马逊);UK AI Safety Institute(英国人工智能安全研究所);University of Oxford(牛津大学)
Multi-Paradigm Agent Interaction in Practice:A Systematic Analysis of Generator-Evaluator, ReAct Loop,and Adversarial Evaluation in the buddyMe Framework
BuddyMe Research Team(buddyMe研究团队);CHARMMIRAEL Biotech Co., Ltd(CHARMMIRAEL生物技术有限公司)
Medchain: Bridging the Gap Between LLM Agents and Clinical Practice with Interactive Sequence
The City University of Hong Kong(香港城市大学);Renmin University of China(中国人民大学);The Chinese University of Hong Kong(香港中文大学);Shenzhen University(深圳大学);National Yang Ming Chiao Tung University(阳明交通大学);Taipei Veterans General Hospital(台北荣民总医院)
A Field Guide to Deploying AI Agents in Clinical Practice
Mass General Brigham(麻省总医院布莱根医疗体系);Harvard Medical School(哈佛医学院);MIT Sloan School of Management(麻省理工学院斯隆管理学院);Brown University Health(布朗大学健康体系);Dana-Farber Cancer Institute(丹娜-法伯癌症研究所);The Legorreta Cancer Center at Brown University, Brown University Health Cancer Institute(布朗大学莱戈雷塔癌症中心,布朗大学健康癌症研究所);Maastricht University(马斯特里赫特大学);Center for Biomedical Informatics & Information Technology, National Cancer Institute, NIH(美国国立卫生研究院国家癌症研究所生物医学信息与信息技术中心);Rhode Island Hospital(罗德岛医院);Computational Health Informatics Program (CHIP), Boston Children’s Hospital(波士顿儿童医院计算健康信息学项目)
DUCTILE: Agentic LLM Orchestration of Engineering Analysis in Product Development Practice
Product Development Division, Department of Mechanical Engineering, Chalmers University of Technology, Gothenburg, Sweden(产品开发部门,机械工程系,楚姆勒斯技术大学,哥德堡,瑞典);GKN Aerospace Engine Systems, Trollhättan, Sweden(GKN航空发动机系统,特罗尔哈塔纳,瑞典)
Agent4Edu: Generating Learner Response Data by Generative Agents for Intelligent Education Systems
State Key Laboratory of Cognitive Intelligence(认知智能国家重点实验室);University of Science and Technology of China(中国科学技术大学);Institute of Artificial Intelligence(人工智能研究院);Hefei Comprehensive National Science Center(合肥综合性国家科学中心)
Ready Jurist One: Benchmarking Language Agents for Legal Intelligence in Dynamic Environments
Fudan University(复旦大学);Shanghai Innovation Institute(上海创新研究院);Huazhong University of Science and Technology(华中科技大学);University of Southern California(南加利福尼亚大学);Midu Technology(米度科技);Northwest University of Political and Law(西北政法大学)
One Analyst Is Not Ground Truth: Grading Agent-Built Financial Models Against Observed Professional Practice
Shanghai Jiao Tong University(上海交通大学);University of Adelaide(阿德莱德大学);Tsinghua University(清华大学);National University of Singapore(新加坡国立大学);Peking University(北京大学)
The Agentic Researcher: A Practical Guide to AI-Assisted Research in Mathematics and Machine Learning
Department for AI in Society, Science, and Technology(人工智能与社会、科学和技术部门);Zuse Institute Berlin(柏林Zuse研究所);Institute of Mathematics(数学研究所);Technische Universität Berlin(柏林技术大学)
What are the Essential Factors in Crafting Effective Long Context Multi-Hop Instruction Datasets? Insights and Best Practices
Shanghai Artificial Intelligence Laboratory(上海人工智能实验室);Research Center for Social Computing and Interactive Robotics(社会计算与交互机器人研究中心);Harbin Institute of Technology(哈尔滨工业大学);School of Computer Science and Engineering, Central South University(中南大学计算机科学与工程学院)
How Do Practitioners Build SE Agents? Insights from a Mixed-Methods Study
Singapore Management University(新加坡管理大学);University College London(伦敦大学学院);Tencent(腾讯);University of Alberta(阿尔伯塔大学);Alberta Machine Intelligence Institute(阿尔伯塔机器智能研究所)
SurgBox: Agent-Driven Operating Room Sandbox with Surgery Copilot
CAIR, HKISI, CAS(中国科学院香港中文大学信息安全研究所智能计算研究中心);MAIS, Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所多模态人工智能系统实验室);Peking Union Medical College Hospital(北京协和医院)
Detailed balance in large language model-driven agents
School of Physics, Peking University, Beijing 100871, China(物理系,北京大学,北京);Center for High Energy Physics, Peking University, Beijing 100871, China(高能物理中心,北京大学,北京);Beijing Computational Science Research Center, Beijing 100193, China(北京计算科学研究院,北京)