Meta-World: A Benchmark and Evaluation for Multi-Task and Meta Reinforcement Learning

TL;DR

This paper introduces Meta-World, a comprehensive benchmark with 50 robotic manipulation tasks designed to evaluate and improve meta-reinforcement learning algorithms' ability to generalize across diverse tasks.

Contribution

It provides an open-source, broad task benchmark for meta-RL, highlighting current algorithms' limitations in multi-task generalization and fostering future research.

Findings

Algorithms perform well on individual tasks but struggle with multiple tasks.

Meta-World enables evaluation of generalization to new, unseen tasks.

Current methods need improvement for effective multi-task learning.

Abstract

Meta-reinforcement learning algorithms can enable robots to acquire new skills much more quickly, by leveraging prior experience to learn how to learn. However, much of the current research on meta-reinforcement learning focuses on task distributions that are very narrow. For example, a commonly used meta-reinforcement learning benchmark uses different running velocities for a simulated robot as different tasks. When policies are meta-trained on such narrow task distributions, they cannot possibly generalize to more quickly acquire entirely new tasks. Therefore, if the aim of these methods is to enable faster acquisition of entirely new behaviors, we must evaluate them on task distributions that are sufficiently broad to enable generalization to new behaviors. In this paper, we propose an open-source simulated benchmark for meta-reinforcement learning and multi-task learning consisting of 50 distinct robotic manipulation tasks. Our aim is to make it possible to develop algorithms that generalize to accelerate the acquisition of entirely new, held-out tasks. We evaluate 7 state-of-the-art meta-reinforcement learning and multi-task learning algorithms on these tasks. Surprisingly, while each task and its variations (e.g., with different object positions) can be learned with reasonable success, these algorithms struggle to learn with multiple tasks at the same time, even with as few as ten distinct training tasks. Our analysis and open-source environments pave the way for future research in multi-task learning and meta-learning that can enable meaningful generalization, thereby unlocking the full potential of these methods.