The Role of Pretrained Representations for the OOD Generalization of RL Agents

Frederik Träuble, Andrea Dittadi, Manuel Wuthrich, Felix Widmaier, Peter Vincent Gehler, Ole Winther, Francesco Locatello, Olivier Bachem, Bernhard Schölkopf, Stefan Bauer

Keywords: representations, out-of-distribution, generalization, deep learning, reinforcement learning

Abstract: Building sample-efficient agents that generalize out-of-distribution (OOD) in real-world settings remains a fundamental unsolved problem on the path towards achieving higher-level cognition. One particularly promising approach is to begin with low-dimensional, pretrained representations of our world, which should facilitate efficient downstream learning and generalization. By training 240 representations and over 10,000 reinforcement learning (RL) policies on a simulated robotic setup, we evaluate to what extent different properties of pretrained VAE-based representations affect the OOD generalization of downstream agents. We observe that many agents are surprisingly robust to realistic distribution shifts, including the challenging sim-to-real case. In addition, we find that the generalization performance of a simple downstream proxy task reliably predicts the generalization performance of our RL agents under a wide range of OOD settings. Such proxy tasks can thus be used to select pretrained representations that will lead to agents that generalize.

One-sentence Summary: We study the role of pretrained representations for the out-of-distribution generalization of RL agents.