2023 Academia Sinica Open House

Deep Reinforcement Learning (DRL) has achieved significant advancements in various domains, such as game-playing, robotics, and natural language processing. Among these fields, games serve as a pivotal testing ground for DRL algorithms due to their controllable and accessible environments that stand in contrast to the complex real-world problems. For example, in 2016, AlphaGo integrated DRL with search algorithms and beat world champion Lee Sedol. Following the success of AlphaGo, recent DRL algorithms, like AlphaZero and MuZero, have demonstrated super-human performance in numerous computer games without using any human-derived knowledge. This talk will begin with an introduction to the foundational concepts of reinforcement learning. We will then discuss various reinforcement learning algorithms used in game-playing and other domains.