Asymmetric actor critic for image-based robot learning

Asymmetric actor critic for image-based robot learning | OpenAI

October 18, 2017

Asymmetric actor critic for image-based robot learning

Loading…

Share

Abstract

Deep reinforcement learning (RL) has proven a powerful technique in many sequential decision making domains. However, Robotics poses many challenges for RL, most notably training on a physical system can be expensive and dangerous, which has sparked significant interest in learning control policies using a physics simulator. While several recent works have shown promising results in transferring policies trained in simulation to the real world, they often do not fully utilize the advantage of working with a simulator. In this work, we exploit the full state observability in the simulator to train better policies which take as input only partial observations (RGBD images). We do this by employing an actor-critic training algorithm in which the critic is trained on full states while the actor (or policy) gets rendered images as input. We show experimentally on a range of simulated tasks that using these asymmetric inputs significantly improves performance. Finally, we combine this method with domain randomization and show real robot experiments for several tasks like picking, pushing, and moving a block. We achieve this simulation to real world transfer without training on any real world data.

Authors

Lerrel Pinto, Marcin Andrychowicz, Peter Welinder, Wojciech Zaremba, Pieter Abbeel

Related articles

Solving Rubik’s Cube with a robot handMilestoneOct 15, 2019

Learning dexterityMilestoneJul 30, 2018

Multi-Goal Reinforcement Learning: Challenging robotics environments and request for researchPublicationFeb 26, 2018