[2J4-GS-2] Machine learning: Deep reinforcement learning

[2J4-GS-2-05] Research on American Football Coaching Strategy Using Deep Reinforcement Learning

〇Risa Takayanagi¹, Dinesh Malla ^1,2, Tsuyoshi Sakai¹, Tomah Sogabe^1,2 (1. UEC, 2. Grid Inc.)

Keywords:Deep reinforcement learning, american football, Coaching

In recent years, with the development of information processing technology and image processing technology, the utilization of data and the introduction of AI technology have been actively performed in the sports world. The purpose of this work is to propose a reinforcement learning-based coaching framework in American football, which is a game of complex strategy and tactics. In this paper, we examined tactical prediction and strategy optimization by Q-leaning simulation, and various issues were examined in order to make it closer to the actual match in the future.