报告人简介：

安波是新加坡南洋理工大学计算机科学与工程学院的助理教授，于2011年在美国麻省大学Amherst分校获计算机科学博士学位。主要研究领域包括人工智能、多智能体系统、博弈论及优化。有70余篇论文发表在人工智能领域的国际顶级会议AAMAS、IJCAI、AAAI、ICAPS、KDD以及著名学术期刊JAAMAS、AIJ、ACM/IEEE Transactions。曾获2010年国际智能体及多智能体系统协会(IFAAMAS)杰出博士论文奖、 2011年美国海岸警卫队的卓越运营奖、2012年国际智能体及多智能体系统年会(AAMAS)最佳应用论文奖、2016年人工智能创新应用会议(IAAI)创新应用论文奖，以及2012年美国运筹学和管理学研究协会(INFORMS)Daniel H. Wagner杰出运筹学应用奖等荣誉。受邀在2017年国际人工智能联合会议(IJCAI)上做Early Career Spotlight talk。 获得2017年微软合作AI挑战赛的冠军。他是Journal of Artificial Intelligence Research (JAIR)编委会成员以及Journal of Autonomous Agents and Multi-agent Systems (JAAMAS)的副主编。当选为国际智能体及多智能体系统协会理事会成员。

报告简介：

It has been an open challenge for self-interested agents to make optimal sequential decisions in complex multi-agent systems, where agents might achieve higher utility via collaboration. The Microsoft Malmo Collaborative AI Challenge (MCAC), which is designed to encourage research relating to various problems in Collaborative AI, takes the form of a Minecraft mini-game where players might work together to catch a pig or deviate from cooperation, for pursuing high scores to win the challenge. Various characteristics, such as complex interactions among agents, uncertainties, sequential decision making and limited learning trials all make it extremely challenging to find effective strategies. This talk presents HogRider - the champion agent of MCAC in 2017 out of 81 teams from 26 countries. One key innovation of HogRider is a generalized agent type hypothesis framework to identify the behaviour model of the other agents. On top of that, a second key innovation is a novel Q-learning approach combing ideas such as state-action abstraction to reduce problem scale, a warm start approach using human reasoning for addressing limited learning trials, and an active greedy strategy to balance exploitation-exploration.  This talk will also outline future research directions.