Timezone: »

SeMAIL: Eliminating Distractors in Visual Imitation via Separated Models
Shenghua Wan · Yucen Wang · Minghao Shao · Ruying Chen · De-Chuan Zhan

Wed Jul 26 05:00 PM -- 06:30 PM (PDT) @ Exhibit Hall 1 #412

Model-based imitation learning (MBIL) is a popular reinforcement learning method that improves sample efficiency on high-dimension input sources, such as images and videos. Following the convention of MBIL research, existing algorithms are highly deceptive by task-irrelevant information, especially moving distractors in videos. To tackle this problem, we propose a new algorithm - named Separated Model-based Adversarial Imitation Learning (SeMAIL) - decoupling the environment dynamics into two parts by task-relevant dependency, which is determined by agent actions, and training separately. In this way, the agent can imagine its trajectories and imitate the expert behavior efficiently in task-relevant state space. Our method achieves near-expert performance on various visual control tasks with complex observations and the more challenging tasks with different backgrounds from expert observations.

Author Information

Shenghua Wan (Nanjing University)

My research interests include Reinforcement Learning.

Yucen Wang (Nanjing University)
Minghao Shao (Nanjing University)
Ruying Chen
De-Chuan Zhan (Nanjing University)

More from the Same Authors