【淡江大學熊貓講座】林復華教授演講

•     時間：108年3月26日(二)10:10~12:00
•     地點：守謙國際會議中心有蓮國際廳
•     講題：Research on multi-agent reinforcement learning
•     講者：加拿大Athabasca University計算與信息系統學院林復華教授
•     Abstract：

Distributed competitive decision making, as opposed to centralized planning, is emerging as the norm in networked systems which involve repeatedly making decisions in an uncertain environment.

In this keynote presentation I will present Multi-Agent Systems with Reinforcement Learning for decision-making for such systems, taking into account rationally selfish behavior of the connected agents/minds. In such environments, agents need to consider how to compete for scarce resources, trade, negotiate automatically, learn from each other, and form social organizations. In particular, I will discuss multi-armed bandit (MAB) learning algorithms. Unlike standard black-box and big-data based machine learning tools, MAB algorithms are active learning, which refers to algorithms that actively select data they should receive, and online learning, which refers to algorithms that analyze data in real-time and provide results on the fly. I will present MAB two real-world applications: Online Scheduling in Oil and Gas Industry and Vehicle Routing for Driverless Cars in Smart City.

•     對象：全校教職員生及校外人士
•     承辦單位：資訊工程學系分機2665