Generative Particle Variational Inference via Estimation of Functional Gradients

Abstract

Recently, particle-based variational inference (ParVI) methods have gained interest because they can avoid arbitrary parametric assumptions that are common in variational inference. However, many ParVI approaches do not allow arbitrary sampling from the posterior, and the few that do allow such sampling suffer from suboptimality. This work proposes a new method for learning to approximately sample from the posterior distribution. We construct a neural sampler that is trained with the functional gradient of the KL-divergence between the empirical sampling distribution and the target distribution, assuming the gradient resides within a reproducing kernel Hilbert space. Our generative ParVI (GPVI) approach maintains the asymptotic performance of ParVI methods while offering the flexibility of a generative sampler. Through carefully constructed experiments, we show that GPVI outperforms previous generative ParVI methods such as amortized SVGD, and is competitive with ParVI as well as gold-standard approaches like Hamiltonian Monte Carlo for fitting both exactly known and intractable target distributions.

arXiv comment22 pages, 9 figures, 10 tables, 1 algorithm

References in corpus (20)

• Adam: A Method for Stochastic Optimization
• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• MCMC using Hamiltonian dynamics
• Variational Inference with Normalizing Flows
• Weight Uncertainty in Neural Networks
• Stochastic Gradient Hamiltonian Monte Carlo
• Variational Dropout and the Local Reparameterization Trick
• Scalable Variational Gaussian Process Classification
• Bayesian Dark Knowledge
• Stein Variational Gradient Descent as Gradient Flow
• Stochastic Variational Deep Kernel Learning
• Structured and Efficient Variational Deep Learning with Matrix Gaussian Posteriors
• Learning to Draw Samples: With Application to Amortized MLE for Generative Adversarial Learning
• Quality of Uncertainty Quantification for Bayesian Neural Network Inference
• Implicit Weight Uncertainty in Neural Networks
• On the convergence of Hamiltonian Monte Carlo
• Understanding and Accelerating Particle-Based Variational Inference
• Learning the Stein Discrepancy for Training and Evaluating Energy-Based Models without Sampling
• Gradient-free Hamiltonian Monte Carlo with Efficient Kernel Exponential Families
• Stein Neural Sampler