Inferring Behavior-specific Context Improves Zero-shot Generalization In Reinforcement Learning

Abstract

In this work, we address the challenge of zero-shot generalization (ZSG) in Reinforcement Learning (RL), where agents must adapt to entirely novel environments without additional training. We argue that understanding and utilizing contextual cues, such as the gravity level of the environment, is critical for robust generalization, and we propose to integrate the learning of context representations directly with policy learning. Our algorithm demonstrates improved generalization on various simulated domains, outperforming prior context-learning techniques in zero-shot settings. By jointly learning policy and context, our method acquires behavior-specific context representations, enabling adaptation to unseen environments and marks progress towards reinforcement learning systems that generalize across diverse real-world tasks. Our code and experiments are available at https://github.com/tidiane-camaret/contextual_rl_zero_shot.

Stats

• citations0
• S2 citations—
• github stars2
• HF likes0
• heat score0.95
• arxiv keyndir2024inferring

Code

• tidiane-camaret/contextual_rl_zero_shot—

Related papers

• Dreaming Of Many Worlds: Learning Contextual World Models Aids Zero-shot Generalization (2024)2.83
• Contextual Intelligence The Next Leap For Reinforcement Learning (2026)0.00
• Contextualize Me -- The Case For Context In Reinforcement Learning (2022)0.00
• DRED: Zero-shot Transfer In Reinforcement Learning Via Data-regularised Environment Design (2024)1.81
• Cross-trajectory Representation Learning For Zero-shot Generalization In RL (2021)0.00
• A Unified Framework For Zero-shot Reinforcement Learning (2025)0.00
• Dynamics Generalisation In Reinforcement Learning Via Adaptive Context-aware Policies (2023)2.26
• Zero-shot Policy Learning With Spatial Temporal Rewarddecomposition On Contingency-aware Observation (2019)0.00