Dream and Search to Control: Latent Space Planning for Continuous Control

TL;DR

This paper introduces a latent space planning method using tree search for continuous control in model-based reinforcement learning, demonstrating improved sample efficiency and performance on benchmarks.

Contribution

It extends latent-space tree search techniques from discrete to continuous actions, showing practical benefits in continuous control environments.

Findings

Achieves better sample efficiency than state-of-the-art methods.

Demonstrates successful bootstrapping benefits in continuous action spaces.

Improves performance on challenging continuous-control benchmarks.

Abstract

Learning and planning with latent space dynamics has been shown to be useful for sample efficiency in model-based reinforcement learning (MBRL) for discrete and continuous control tasks. In particular, recent work, for discrete action spaces, demonstrated the effectiveness of latent-space planning via Monte-Carlo Tree Search (MCTS) for bootstrapping MBRL during learning and at test time. However, the potential gains from latent-space tree search have not yet been demonstrated for environments with continuous action spaces. In this work, we propose and explore an MBRL approach for continuous action spaces based on tree-based planning over learned latent dynamics. We show that it is possible to demonstrate the types of bootstrapping benefits as previously shown for discrete spaces. In particular, the approach achieves improved sample efficiency and performance on a majority of challenging continuous-control benchmarks compared to the state-of-the-art.