Safe Reinforcement Learning by Imagining the Near Future

TL;DR

This paper introduces a model-based reinforcement learning algorithm that plans ahead to avoid unsafe states, demonstrating improved safety and competitive rewards in continuous control tasks.

Contribution

It proposes a novel planning-based approach that heavily penalizes unsafe trajectories, with theoretical guarantees for safety under certain conditions.

Findings

Achieves fewer safety violations compared to baseline methods

Maintains competitive rewards in continuous control tasks

Provides theoretical safety guarantees under model accuracy assumptions

Abstract

Safe reinforcement learning is a promising path toward applying reinforcement learning algorithms to real-world problems, where suboptimal behaviors may lead to actual negative consequences. In this work, we focus on the setting where unsafe states can be avoided by planning ahead a short time into the future. In this setting, a model-based agent with a sufficiently accurate model can avoid unsafe states. We devise a model-based algorithm that heavily penalizes unsafe trajectories, and derive guarantees that our algorithm can avoid unsafe states under certain assumptions. Experiments demonstrate that our algorithm can achieve competitive rewards with fewer safety violations in several continuous control tasks.