Reinforcement Learning from Delayed Observations via World Models
This addresses a practical limitation in reinforcement learning for real-world applications where immediate feedback is unavailable, representing a domain-specific advancement.
The paper tackles the problem of observation delays in reinforcement learning by using world models to reduce delayed partially observable Markov decision processes to delayed Markov decision processes, achieving up to 250% performance improvement over naive model-based approaches and demonstrating delay-aware reinforcement learning with visual observations for the first time.
In standard reinforcement learning settings, agents typically assume immediate feedback about the effects of their actions after taking them. However, in practice, this assumption may not hold true due to physical constraints and can significantly impact the performance of learning algorithms. In this paper, we address observation delays in partially observable environments. We propose leveraging world models, which have shown success in integrating past observations and learning dynamics, to handle observation delays. By reducing delayed POMDPs to delayed MDPs with world models, our methods can effectively handle partial observability, where existing approaches achieve sub-optimal performance or degrade quickly as observability decreases. Experiments suggest that one of our methods can outperform a naive model-based approach by up to 250%. Moreover, we evaluate our methods on visual delayed environments, for the first time showcasing delay-aware reinforcement learning continuous control with visual observations.