Simplifying Model-based RL: Learning Representations, Latent-space Models, and Policies with One Objective
This work addresses sample efficiency and computational demands in reinforcement learning for robotics or simulation tasks, representing a novel integration rather than an incremental improvement.
The paper tackles the challenge of aligning auxiliary objectives with the RL goal in model-based reinforcement learning by proposing a single objective that jointly optimizes a latent-space model and policy for self-consistency and high returns, resulting in matching or improving sample efficiency of prior methods and achieving SAC's performance in about 50% less wall-clock time.
While reinforcement learning (RL) methods that learn an internal model of the environment have the potential to be more sample efficient than their model-free counterparts, learning to model raw observations from high dimensional sensors can be challenging. Prior work has addressed this challenge by learning low-dimensional representation of observations through auxiliary objectives, such as reconstruction or value prediction. However, the alignment between these auxiliary objectives and the RL objective is often unclear. In this work, we propose a single objective which jointly optimizes a latent-space model and policy to achieve high returns while remaining self-consistent. This objective is a lower bound on expected returns. Unlike prior bounds for model-based RL on policy exploration or model guarantees, our bound is directly on the overall RL objective. We demonstrate that the resulting algorithm matches or improves the sample-efficiency of the best prior model-based and model-free RL methods. While sample efficient methods typically are computationally demanding, our method attains the performance of SAC in about 50% less wall-clock time.