Deep Variational Sequential Monte Carlo for High-Dimensional Observations
This work addresses a bottleneck in nonlinear state-space systems for researchers and practitioners in fields like signal processing or robotics, offering an incremental improvement over existing SMC methods.
The paper tackled the problem of poor performance in Sequential Monte Carlo (SMC) methods due to approximated distributions by introducing a differentiable particle filter that uses a neural network to learn from high-dimensional observations, resulting in outperforming baselines in tracking the Lorenz attractor and providing a more accurate posterior representation.
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.