Deep Variational Sequential Monte Carlo for High-Dimensional Observations

TL;DR

This paper presents a differentiable particle filter using variational SMC with neural networks to improve high-dimensional state estimation, outperforming baselines in complex tracking tasks.

Contribution

It introduces a neural network-based proposal and transition distribution within a differentiable particle filter trained with variational SMC objectives for high-dimensional data.

Findings

Outperforms baseline methods in Lorenz attractor tracking.

Provides a more accurate posterior distribution representation.

Demonstrates effectiveness on high-dimensional, partial observations.

Abstract

Sequential Monte Carlo (SMC), or particle filtering, is widely used in nonlinear state-space systems, but its performance often suffers from poorly approximated proposal and state-transition distributions. This work introduces a differentiable particle filter that leverages the unsupervised variational SMC objective to parameterize the proposal and transition distributions with a neural network, designed to learn from high-dimensional observations. Experimental results demonstrate that our approach outperforms established baselines in tracking the challenging Lorenz attractor from high-dimensional and partial observations. Furthermore, an evidence lower bound based evaluation indicates that our method offers a more accurate representation of the posterior distribution.