ProMP: Proximal Meta-Policy Search

About

Credit assignment in Meta-reinforcement learning (Meta-RL) is still poorly understood. Existing methods either neglect credit assignment to pre-adaptation behavior or implement it naively. This leads to poor sample-efficiency during meta-training as well as ineffective task identification strategies. This paper provides a theoretical analysis of credit assignment in gradient-based Meta-RL. Building on the gained insights we develop a novel meta-learning algorithm that overcomes both the issue of poor credit assignment and previous difficulties in estimating meta-policy gradients. By controlling the statistical distance of both pre-adaptation and adapted policies during meta-policy search, the proposed algorithm endows efficient and stable meta-learning. Our approach leads to superior pre-adaptation policy behavior and consistently outperforms previous Meta-RL algorithms in sample-efficiency, wall-clock time, and asymptotic performance.

Jonas Rothfuss, Dennis Lee, Ignasi Clavera, Tamim Asfour, Pieter Abbeel · 2018

Related benchmarks

Task	Dataset	Result
Reinforcement Learning	Ant-gravity v2	Average Return1253	8
Reinforcement Learning	HalfCheetah gravity v2	Average Return885	8
Reinforcement Learning	Hopper gravity v2	Average Return252	8
Reinforcement Learning	Humanoid gravity v2	Average Return423	8
Reinforcement Learning	Walker2d gravity v2	Average Return271	8
Reinforcement Learning	Half-cheetah-velocity (train)	Runtime (hours)2	7