Learning Differentiable Particle Filter on the Fly

TL;DR

This paper introduces an online learning framework for differentiable particle filters that updates model parameters in real-time without ground truth states, improving adaptability and reducing deployment delays.

Contribution

It presents a novel unsupervised online learning approach for differentiable particle filters, enabling real-time model updates without requiring known states.

Findings

Effective in simulation environments including linear Gaussian models

Outperforms supervised methods in online adaptation

Demonstrates robustness to distribution shifts

Abstract

Differentiable particle filters are an emerging class of sequential Bayesian inference techniques that use neural networks to construct components in state space models. Existing approaches are mostly based on offline supervised training strategies. This leads to the delay of the model deployment and the obtained filters are susceptible to distribution shift of test-time data. In this paper, we propose an online learning framework for differentiable particle filters so that model parameters can be updated as data arrive. The technical constraint is that there is no known ground truth state information in the online inference setting. We address this by adopting an unsupervised loss to construct the online model updating procedure, which involves a sequence of filtering operations for online maximum likelihood-based parameter estimation. We empirically evaluate the effectiveness of the proposed method, and compare it with supervised learning methods in simulation settings including a multivariate linear Gaussian state-space model and a simulated object tracking experiment.