Differentiable Particle Filtering

Published in Oral (long talk), ICML 2021, 2020

A novel, principled approach to Differentiable Particle Filtering, using Optimal Transport. Long talk/ oral at ICML 2021.

Paper: https://arxiv.org/abs/2102.07850

Cite ``` @inproceedings{corenflos2021differentiable, title={Differentiable particle filtering via entropy-regularized optimal transport}, author={Corenflos, Adrien and Thornton, James and Deligiannidis, George and Doucet, Arnaud}, booktitle={International Conference on Machine Learning}, pages={2100--2111}, year={2021}, organization={PMLR} } ```

Particle Filtering (PF) methods are an established class of procedures for performing inference in non-linear state-space models. Resampling is a key ingredient of PF, necessary to obtain low variance likelihood and states estimates. However, traditional resampling methods result in PF-based loss functions being non-differentiable with respect to model and PF parameters. In a variational inference context, resampling also yields high variance gradient estimates of the PF-based evidence lower bound. By leveraging optimal transport ideas, we introduce a principled differentiable particle filter and provide convergence results. We demonstrate this novel method on a variety of applications. Accepted as a Long talk at ICML 2021.

Paper