Uncertainty Representations in State-Space Layers for Deep Reinforcement Learning under Partial Observability

TL;DR

This paper introduces a Kalman filter layer for deep reinforcement learning that explicitly models uncertainty in partially observable environments, improving decision-making by integrating probabilistic filtering into standard architectures.

Contribution

It proposes a novel Kalman filter layer that performs Gaussian inference within deep RL models, enhancing uncertainty reasoning capabilities in partial observability scenarios.

Findings

Kalman filter layers outperform other recurrent models in uncertainty-critical tasks.

The layer scales logarithmically with sequence length, enabling efficient processing.

Explicit probabilistic filtering improves decision-making under uncertainty.

Abstract

Optimal decision-making under partial observability requires reasoning about the uncertainty of the environment's hidden state. However, most reinforcement learning architectures handle partial observability with sequence models that have no internal mechanism to incorporate uncertainty in their hidden state representation, such as recurrent neural networks, deterministic state-space models and transformers. Inspired by advances in probabilistic world models for reinforcement learning, we propose a standalone Kalman filter layer that performs closed-form Gaussian inference in linear state-space models and train it end-to-end within a model-free architecture to maximize returns. Similar to efficient linear recurrent layers, the Kalman filter layer processes sequential data using a parallel scan, which scales logarithmically with the sequence length. By design, Kalman filter layers are a drop-in replacement for other recurrent layers in standard model-free architectures, but importantly they include an explicit mechanism for probabilistic filtering of the latent state representation. Experiments in a wide variety of tasks with partial observability show that Kalman filter layers excel in problems where uncertainty reasoning is key for decision-making, outperforming other stateful models.