Exploring Unknown States with Action Balance

Yan Song, Yingfeng Chen, Yujing Hu +1

Exploration is a key problem in reinforcement learning. Recently bonus-based methods have achieved considerable successes in environments where exploration is difficult such as Mon…