Meta-Reinforcement Learning via Language Instructions

TL;DR

This paper introduces a meta-reinforcement learning algorithm that leverages natural language instructions to improve robotic manipulation, significantly enhancing learning efficiency and success rates compared to existing methods.

Contribution

The proposed algorithm uniquely combines language understanding with meta-RL to enable robots to learn manipulation tasks more effectively using instructions.

Findings

Outperforms state-of-the-art methods on Meta-World benchmark

Achieves higher training and testing success rates

Effectively integrates language instructions into reinforcement learning

Abstract

Although deep reinforcement learning has recently been very successful at learning complex behaviors, it requires a tremendous amount of data to learn a task. One of the fundamental reasons causing this limitation lies in the nature of the trial-and-error learning paradigm of reinforcement learning, where the agent communicates with the environment and progresses in the learning only relying on the reward signal. This is implicit and rather insufficient to learn a task well. On the contrary, humans are usually taught new skills via natural language instructions. Utilizing language instructions for robotic motion control to improve the adaptability is a recently emerged topic and challenging. In this paper, we present a meta-RL algorithm that addresses the challenge of learning skills with language instructions in multiple manipulation tasks. On the one hand, our algorithm utilizes the language instructions to shape its interpretation of the task, on the other hand, it still learns to solve task in a trial-and-error process. We evaluate our algorithm on the robotic manipulation benchmark (Meta-World) and it significantly outperforms state-of-the-art methods in terms of training and testing task success rates. Codes are available at \url{https://tumi6robot.wixsite.com/million}.