Latent Plans for Task-Agnostic Offline Reinforcement Learning

TL;DR

This paper introduces a hierarchical offline reinforcement learning method that combines imitation learning and RL to learn task-agnostic, long-horizon policies from high-dimensional observations, enabling robots to perform complex multi-step tasks.

Contribution

It proposes a novel hierarchical approach that integrates latent skill learning with offline RL for long-horizon, task-agnostic robot control from visual inputs.

Findings

Achieves an order-of-magnitude performance improvement over baselines.

Successfully learns a multi-task visuomotor policy for 25 real-world tasks.

Enables skill chaining to reach complex, long-horizon goals.

Abstract

Everyday tasks of long-horizon and comprising a sequence of multiple implicit subtasks still impose a major challenge in offline robot control. While a number of prior methods aimed to address this setting with variants of imitation and offline reinforcement learning, the learned behavior is typically narrow and often struggles to reach configurable long-horizon goals. As both paradigms have complementary strengths and weaknesses, we propose a novel hierarchical approach that combines the strengths of both methods to learn task-agnostic long-horizon policies from high-dimensional camera observations. Concretely, we combine a low-level policy that learns latent skills via imitation learning and a high-level policy learned from offline reinforcement learning for skill-chaining the latent behavior priors. Experiments in various simulated and real robot control tasks show that our formulation enables producing previously unseen combinations of skills to reach temporally extended goals by "stitching" together latent skills through goal chaining with an order-of-magnitude improvement in performance upon state-of-the-art baselines. We even learn one multi-task visuomotor policy for 25 distinct manipulation tasks in the real world which outperforms both imitation learning and offline reinforcement learning techniques.