no code implementations • 8 Oct 2019 • Kyo Kutsuzawa, Hitoshi Kusano, Ayaka Kume, Shoichiro Yamaguchi
The appropriate motions can be found efficiently by searching the latent space of the trained cGANs instead of the motion space, while avoiding poor local optima.
no code implementations • 17 Oct 2017 • Ayaka Kume, Eiichi Matsumoto, Kuniyuki Takahashi, Wilson Ko, Jethro Tan
To solve this problem, we propose Map-based Multi-Policy Reinforcement Learning (MMPRL), which aims to search and store multiple policies that encode different behavioral features while maximizing the expected reward in advance of the environment change.