Neural Adaptive Sequential Monte Carlo

TL;DR

This paper introduces Neural Adaptive Sequential Monte Carlo (NASMC), a flexible method that automatically adapts proposal distributions using neural networks, significantly improving inference accuracy in complex models and enabling advanced applications like music modeling.

Contribution

The paper proposes NASMC, a novel framework that combines neural network parameterizations with adaptive SMC, enhancing proposal adaptation and inference quality in various models.

Findings

NASMC outperforms traditional adaptive proposals like EKF and UKF in non-linear state space models.

NASMC improves parameter learning when integrated with Particle Marginal Metropolis Hastings.

NASMC achieves competitive results in latent variable RNNs for music modeling.

Abstract

Sequential Monte Carlo (SMC), or particle filtering, is a popular class of methods for sampling from an intractable target distribution using a sequence of simpler intermediate distributions. Like other importance sampling-based methods, performance is critically dependent on the proposal distribution: a bad proposal can lead to arbitrarily inaccurate estimates of the target distribution. This paper presents a new method for automatically adapting the proposal using an approximation of the Kullback-Leibler divergence between the true posterior and the proposal distribution. The method is very flexible, applicable to any parameterized proposal distribution and it supports online and batch variants. We use the new framework to adapt powerful proposal distributions with rich parameterizations based upon neural networks leading to Neural Adaptive Sequential Monte Carlo (NASMC). Experiments indicate that NASMC significantly improves inference in a non-linear state space model outperforming adaptive proposal methods including the Extended Kalman and Unscented Particle Filters. Experiments also indicate that improved inference translates into improved parameter learning when NASMC is used as a subroutine of Particle Marginal Metropolis Hastings. Finally we show that NASMC is able to train a latent variable recurrent neural network (LV-RNN) achieving results that compete with the state-of-the-art for polymorphic music modelling. NASMC can be seen as bridging the gap between adaptive SMC methods and the recent work in scalable, black-box variational inference.