Geometry of Uncertainty: Learning Metric Spaces for Multimodal State Estimation in RL
This addresses the challenge of robust state estimation in RL for agents operating in noisy, multimodal environments, offering a scalable solution without explicit noise assumptions.
The paper tackles the problem of estimating environment states from high-dimensional, multimodal, and noisy observations in reinforcement learning by learning a metric space where distances correlate with action transitions, resulting in improved robustness to sensor noise and superior state estimation compared to baselines.
Estimating the state of an environment from high-dimensional, multimodal, and noisy observations is a fundamental challenge in reinforcement learning (RL). Traditional approaches rely on probabilistic models to account for the uncertainty, but often require explicit noise assumptions, in turn limiting generalization. In this work, we contribute a novel method to learn a structured latent representation, in which distances between states directly correlate with the minimum number of actions required to transition between them. The proposed metric space formulation provides a geometric interpretation of uncertainty without the need for explicit probabilistic modeling. To achieve this, we introduce a multimodal latent transition model and a sensor fusion mechanism based on inverse distance weighting, allowing for the adaptive integration of multiple sensor modalities without prior knowledge of noise distributions. We empirically validate the approach on a range of multimodal RL tasks, demonstrating improved robustness to sensor noise and superior state estimation compared to baseline methods. Our experiments show enhanced performance of an RL agent via the learned representation, eliminating the need of explicit noise augmentation. The presented results suggest that leveraging transition-aware metric spaces provides a principled and scalable solution for robust state estimation in sequential decision-making.