Google Research(谷歌研究院);Massachusetts Institute of Technology(麻省理工学院);Google DeepMind(谷歌DeepMind)
Yubin Kim, Ken Gu, Chanwoo Park, Chunjong Park, Samuel Schmidgall, A. Ali Heydari, Yao Yan, Zhihan Zhang, Yuchen Zhuang, Yun Liu, Mark Malhotra, Paul Pu Liang, … 展开作者
Yubin Kim, Ken Gu, Chanwoo Park, Chunjong Park, Samuel Schmidgall, A. Ali Heydari, Yao Yan, Zhihan Zhang, Yuchen Zhuang, Yun Liu, Mark Malhotra, Paul Pu Liang, Hae Won Park, Yuzhe Yang, Xuhai Xu, Yilun Du, Shwetak Patel, Tim Althoff, Daniel McDuff, Xin Liu
Agents, language model-based systems capable of reasoning, planning, and acting are widely adopted in real-world tasks, yet how their performance changes as these systems scale across key dimensions remains underexplored. We introduce quantitative scaling principles for agent systems as a predictive model, capturing how performance varies with coordination, model capability, and measurable system and task factors. Across 260 configurations spanning six agentic benchmarks, five canonical architectures (Single-Agent and four Multi-Agent: Independent, Centralized, Decentralized, Hybrid), and three LLM families, we perform controlled evaluations, standardizing tools, prompts, and compute to isolate architectural effects. The resulting model achieves a cross-validated R^2=0.373 across all six benchmarks (R^2=0.413 with a task-grounded capability metric). We identify a robust capability-saturation effect and additional patterns: (1) a coordination yields diminishing returns once single-agent baselines exceed certain performance; (2) tool-heavy tasks appear to incur multi-agent overhead; and (3) architectures without centralized verification tend to propagate errors more than those with centralized coordination. Relative performance change compared to single-agent baseline ranges from +80.8% on decomposable financial reasoning to -70.0% on sequential planning, demonstrating that architecture-task alignment determines collaborative success. The framework identifies the best-performing architecture for 87% of held-out configurations and shows consistent relative architecture preferences on unseen frontier models. Agent effectiveness depends on alignment between coordination and task structure, and that mismatched coordination degrades the performance.
From Agent-Only Social Networks to Autonomous Scientific Research: Lessons from OpenClaw and Moltbook, and the Architecture of ClawdLab and Beach.Science
VDSAgents: A PCS-Guided Multi-Agent System for Veridical Data Science Automation
School of Management, Xi’an Jiaotong University(西安交通大学管理学院);School of Computing, National University of Singapore(新加坡国立大学计算机学院);School of Data Science and Media Intelligence, Communication University of China(中国传媒大学数据科学与媒体智能学院);Beijing Baixingkefu Network Technology Co., Ltd.(北京百星科技网络技术有限公司)
AutoKaggle: A Multi-Agent Framework for Autonomous Data Science Competitions
Interdisciplinary Centre for Security, Reliability and Trust (SnT), Université du Luxembourg(卢森堡大学安全、可靠性与跨学科研究中心);University of Melbourne(墨尔本大学);ByteDance Inc.(字节跳动公司);M-A-P
Aleks: AI powered Multi Agent System for Autonomous Scientific Discovery via Data-Driven Approaches in Plant Science
School of Electrical and Computer Engineering(电气与计算机工程学院);Cornell University(康奈尔大学);Horticulture Section, School of Integrative Plant Science, Cornell AgriTech(植物科学整合学院园艺部门,康奈尔农业科技);PPPMB Section, School of Integrative Plant Science, Cornell AgriTech(植物科学整合学院PPPMB部门,康奈尔农业科技)
Visual Document Understanding and Reasoning: A Multi-Agent Collaboration Framework with Agent-Wise Adaptive Test-Time Scaling
National University of Singapore(新加坡国立大学);Tencent Youtu Lab(腾讯优图实验室);Tsinghua University(清华大学);University of Science and Technology of China(中国科学技术大学);Nanyang Technological University(南洋理工大学);Zhejiang University(浙江大学);DeepWisdom(深智科技)
Multilingual Agent-Based World Modeling for Social Science
Institute of Data Science, National University of Singapore(新加坡国立大学数据科学研究所);Singapore University of Technology and Design(新加坡科技设计大学);National School of Development, Peking University(北京大学国家发展研究院);Singapore Management University(新加坡管理大学)
TeachMateGPT: A Multi-Agent Knowledge-Grounded Framework for Pedagogical Assessment Generation from Science Curriculum Materials
Ahsanullah University of Science and Technology(阿赫桑乌拉科技大学);American International University - Bangladesh(孟加拉国美国国际大学);Jashore University of Science and Technology(杰索尔科技大学)
ScienceClaw: Benchmarking Continual Self-Evolution of AI-for-Science Agents Across the Natural and Social Sciences
The Chinese University of Hong Kong, Shenzhen(香港中文大学(深圳));Nanyang Technological University(南洋理工大学);Centre for Biomedical Data Science, Duke-NUS Medical School(杜克-新加坡国立大学医学院生物医学数据科学中心);Zhejiang University(浙江大学);Jurisprudence Research Association, China Law Society(中国法学会法理学研究会)
The Agent Use of Agent Beings: Agent Cybernetics Is the Missing Science of Foundation Agents
Singapore Management University(新加坡管理大学);The Hong Kong Polytechnic University(香港理工大学);Nanyang Technological University(南洋理工大学);Institute for Infocomm Research, Agency for Science, Technology and Research (A*STAR)(信息通信研究院,科技研究局(A*STAR));Shanghai Artificial Intelligence Laboratory(上海人工智能实验室)
LLM Agents as Social Scientists: A Human-AI Collaborative Platform for Social Science Automation
Gaoling School of Artificial Intelligence, Renmin University of China(中国人民大学高瓴人工智能学院);Beijing Key Laboratory of Big Data Management and Analysis Methods(大数据管理与分析方法北京市重点实验室)
KISS - Knowledge Infrastructure for Scientific Simulation: A Scaffolding for Agentic Earth Science
State Key Laboratory of Water Disaster Prevention, Hohai University, Nanjing(水利灾害预防国家重点实验室,河海大学,南京);Yangtze Institute for Conservation and Development, Hohai University, Nanjing(长江保护与发展研究院,河海大学,南京);Department of Bioresource Engineering, McGill University, Sainte-Anne-de-Bellevue, Quebec, Canada(生物资源工程系,麦吉尔大学,圣安妮-德-贝尔贝夫,魁北克,加拿大);Ottawa Research and Development Centre, Agriculture & Agri-Food Canada, Ottawa, Ontario, K1A 0C6, Canada(渥太华研发中心,加拿大农业与食品部,渥太华,安大略,K1A 0C6,加拿大);College of Water Conservancy and Hydropower Engineering, Hohai University, Nanjing(水利水电工程学院,河海大学,南京);Meta Platforms Inc.(Meta平台公司);Nanjing Hydraulic Research Institute, Nanjing(南京水利研究院)
AI-for-Science Low-code Platform with Bayesian Adversarial Multi-Agent Framework
Artificial Intelligence Innovation and Incubation Institute, Fudan University(复旦大学人工智能创新与孵化院);Shanghai Innovation Institute(上海创新研究院);Shanghai Academy of AI for Science(上海人工智能科学研究院)