Does Zero-shot Reinforcement Learning Exist?

Abstract

A zero-shot RL agent is an agent that can solve any RL task in a given environment, instantly with no additional planning or learning, after an initial reward-free learning phase. This marks a shift from the reward-centric RL paradigm towards "controllable" agents that can follow arbitrary instructions in an environment. Current RL agents can solve families of related tasks at best, or require planning anew for each task. Strategies for approximate zero-shot RL ave been suggested using successor features (SFs) [BBQ+ 18] or forward-backward (FB) representations [TO21], but testing has been limited. After clarifying the relationships between these schemes, we introduce improved losses and new SF models, and test the viability of zero-shot RL schemes systematically on tasks from the Unsupervised RL benchmark [LYL+21]. To disentangle universal representation learning from exploration, we work in an offline setting and repeat the tests on several existing replay buffers. SFs appear to s

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keytouati2022does

Related papers

• On Zero-shot Reinforcement Learning (2025)0.00
• A Unified Framework For Zero-shot Reinforcement Learning (2025)0.00
• Zero-shot Reinforcement Learning From Low Quality Data (2023)0.00
• Zero-shot Reinforcement Learning Via Function Encoders (2024)0.00
• Cross-trajectory Representation Learning For Zero-shot Generalization In RL (2021)0.00
• Hypernetworks For Zero-shot Transfer In Reinforcement Learning (2022)6.34
• Zero-shot Policy Learning With Spatial Temporal Rewarddecomposition On Contingency-aware Observation (2019)0.00
• DARLA: Improving Zero-shot Transfer In Reinforcement Learning (2017)0.00