Timezone: »

Principled Offline RL in the Presence of Rich Exogenous Information
Riashat Islam · Manan Tomar · Alex Lamb · Yonathan Efroni · Hongyu Zang · Aniket Didolkar · Dipendra Misra · Xin Li · Harm Seijen · Remi Tachet des Combes · John Langford

Tue Jul 25 05:00 PM -- 06:30 PM (PDT) @ Exhibit Hall 1 #530

Learning to control an agent from offline data collected in a rich pixel-based visual observation space is vital for real-world applications of reinforcement learning (RL). A major challenge in this setting is the presence of input information that is hard to model and irrelevant to controlling the agent. This problem has been approached by the theoretical RL community through the lens of exogenous information, i.e., any control-irrelevant information contained in observations. For example, a robot navigating in busy streets needs to ignore irrelevant information, such as other people walking in the background, textures of objects, or birds in the sky. In this paper, we focus on the setting with visually detailed exogenous information and introduce new offline RL benchmarks that offer the ability to study this problem. We find that contemporary representation learning techniques can fail on datasets where the noise is a complex and time-dependent process, which is prevalent in practical applications. To address these, we propose to use multi-step inverse models to learn Agent-Centric Representations for Offline-RL (ACRO). Despite being simple and reward-free, we show theoretically and empirically that the representation created by this objective greatly outperforms baselines.

Author Information

Riashat Islam (McGill University)
Manan Tomar (University of Alberta)
Alex Lamb (Microsoft Research NYC)
Yonathan Efroni (Meta)
Hongyu Zang (Beijing Institute of Technology)
Aniket Didolkar (Université de Montréal)
Dipendra Misra
Xin Li (Beijing Institute of Technology)
Harm Seijen (Microsoft Research)
Remi Tachet des Combes (AlpacaML)
John Langford (MSR)

More from the Same Authors