Provable Bayesian Inference via Particle Mirror Descent

Abstract

Bayesian methods are appealing in their flexibility in modeling complex data and ability in capturing uncertainty in parameters. However, when Bayes' rule does not result in tractable closed-form, most approximate inference algorithms lack either scalability or rigorous guarantees. To tackle this challenge, we propose a simple yet provable algorithm, \emph{Particle Mirror Descent} (PMD), to iteratively approximate the posterior density. PMD is inspired by stochastic functional mirror descent where one descends in the density space using a small batch of data points at each iteration, and by particle filtering where one uses samples to approximate a function. We prove result of the first kind that, with $m$ particles, PMD provides a posterior density estimator that converges in terms of $KL$-divergence to the true posterior in rate $O(1/\sqrt{m})$. We demonstrate competitive empirical performances of PMD compared to several approximate inference algorithms in mixture models, logistic regression, sparse Gaussian processes and latent Dirichlet allocation on large scale datasets.

38 pages, 26 figures

Cited by in corpus (25)

• Stein Variational Gradient Descent: A General Purpose Bayesian Inference Algorithm
• A Spectral Approach to Gradient Estimation for Implicit Distributions
• A stochastic version of Stein Variational Gradient Descent for efficient sampling
• Stochastic Particle Gradient Descent for Infinite Ensembles
• Distributed Learning for Cooperative Inference
• Riemannian Stein Variational Gradient Descent for Bayesian Inference
• Particle-based Energetic Variational Inference
• Mirror Descent Search and its Acceleration
• Exponential Family Estimation via Adversarial Dynamics Embedding
• A Generalization Bound for Online Variational Inference
• Approximate Inference Turns Deep Networks into Gaussian Processes
• Learning Deep Hidden Nonlinear Dynamics from Aggregate Data
• Offline Policy Selection under Uncertainty
• Kernel Implicit Variational Inference
• Wasserstein variational gradient descent: From semi-discrete optimal transport to ensemble variational inference
• Guaranteed inference in topic models
• Scalable Training of Inference Networks for Gaussian-Process Models
• Variable Selection with Rigorous Uncertainty Quantification using Deep Bayesian Neural Networks: Posterior Concentration and Bernstein-von Mises Phenomenon
• Implicit Generative Modeling for Efficient Exploration
• Faster Stochastic Variational Inference using Proximal-Gradient Methods with General Divergence Functions
• Particle Dual Averaging: Optimization of Mean Field Neural Networks with Global Convergence Rate Analysis
• Adaptive Importance Sampling meets Mirror Descent: a Bias-variance tradeoff
• Bayesian posterior approximation via greedy particle optimization
• Particle Flow Bayes' Rule
• Mirrored Langevin Dynamics