What type of study is this?

This is a Experimental Study study.

October 20, 2025Open Access

Zero-Shot Reinforcement Learning Under Partial Observability

Key Points

Performance of zero-shot reinforcement learning methods declines under partial observability, impacting generalization.
In experiments, memory-based architectures outperformed memory-free baselines, showing much improved effectiveness.
Evaluation was conducted in environments with partially observable states, rewards, and changing dynamics.
Findings suggest that incorporating memory can mitigate the challenges posed by partial observability in RL.

Abstract

Recent work has shown that, under certain assumptions, zero-shot reinforcement learning (RL) methods can generalise to any unseen task in an environment after reward-free pre-training. Access to Markov states is one such assumption, yet, in many real-world applications, the Markov state is only partially observable. Here, we explore how the performance of standard zero-shot RL methods degrades when subjected to partially observability, and show that, as in single-task RL, memory-based architectures are an effective remedy. We evaluate our memory-based zero-shot RL methods in domains where the states, rewards and a change in dynamics are partially observed, and show improved performance over memory-free baselines. Our code is open-sourced via: https://enjeeneer.io/projects/bfms-with-memory/.

Read Full Paperexternally

AI에게 질문

Bookmark

View Full Paper