Stein Variational Gradient Descent: A General Purpose Bayesian Inference Algorithm

Abstract

We propose a general purpose variational inference algorithm that forms a natural counterpart of gradient descent for optimization. Our method iteratively transports a set of particles to match the target distribution, by applying a form of functional gradient descent that minimizes the KL divergence. Empirical studies are performed on various real world models and datasets, on which our method is competitive with existing state-of-the-art methods. The derivation of our method is based on a new theoretical result that connects the derivative of KL divergence under smooth transforms with Stein's identity and a recently proposed kernelized Stein discrepancy, which is of independent interest.

To appear in NIPS 2016

References in corpus (4)

• Probabilistic Backpropagation for Scalable Learning of Bayesian Neural Networks
• Interpretation and Generalization of Score Matching
• Firefly Monte Carlo: Exact MCMC with Subsets of Data
• Nonparametric variational inference

Cited by in corpus (95)

• Physics-Constrained Deep Learning for High-dimensional Surrogate Modeling and Uncertainty Quantification without Labeled Data
• Bayesian Deep Convolutional Encoder-Decoder Networks for Surrogate Modeling and Uncertainty Quantification
• Reinforcement Learning with Deep Energy-Based Policies
• Recent Advances in Autoencoder-Based Representation Learning
• InfoVAE: Information Maximizing Variational Autoencoders
• Modeling the Dynamics of PDE Systems with Physics-Constrained Deep Auto-Regressive Networks
• Bayesian Model-Agnostic Meta-Learning
• Integration of adversarial autoencoders with residual dense convolutional networks for estimation of non-Gaussian hydraulic conductivities
• Autonomous Unmanned Aerial Vehicle Navigation using Reinforcement Learning: A Systematic Review
• Physics-Informed Probabilistic Learning of Linear Embeddings of Non-linear Dynamics With Guaranteed Stability
• Emergent Complexity via Multi-Agent Competition
• Physics-guided deep learning framework for predictive modeling of the Reynolds stress anisotropy
• Diminishing Uncertainty within the Training Pool: Active Learning for Medical Image Segmentation
• Reinforcement Learning Algorithms: An Overview and Classification
• Improving prediction of the terrestrial water storage anomalies during the GRACE and GRACE-FO gap with Bayesian convolutional neural networks
• HypoSVI: Hypocenter inversion with Stein variational inference and Physics Informed Neural Networks
• Evaluating aleatoric and epistemic uncertainties of time series deep learning models for soil moisture predictions
• Measuring Sample Quality with Kernels
• Learning to Draw Samples: With Application to Amortized MLE for Generative Adversarial Learning
• Stein Variational Policy Gradient
• Federated Unlearning: A Survey on Methods, Design Guidelines, and Evaluation Metrics
• The Internet of Federated Things (IoFT): A Vision for the Future and In-depth Survey of Data-driven Approaches for Federated Learning
• Interacting particle solutions of Fokker-Planck equations through gradient-log-density estimation
• Covariances, Robustness, and Variational Bayes
• Adaptive weighting of Bayesian physics informed neural networks for multitask and multiscale forward and inverse problems
• Multiagent Soft Q-Learning
• Bayesian Physics-Informed Neural Networks for the Subsurface Tomography based on the Eikonal Equation
• A Spectral Approach to Gradient Estimation for Implicit Distributions
• Stein Variational Gradient Descent With Matrix-Valued Kernels
• Deterministic Sampling of Expensive Posteriors Using Minimum Energy Designs
• A stochastic version of Stein Variational Gradient Descent for efficient sampling
• Adversarial Symmetric Variational Autoencoder
• The Information Autoencoding Family: A Lagrangian Perspective on Latent Variable Generative Models
• Fed-ensemble: Improving Generalization through Model Ensembling in Federated Learning
• Learning to Draw Samples with Amortized Stein Variational Gradient Descent
• Amortized Variational Inference: A Systematic Review
• Towards Overfitting Avoidance: Tuning-free Tensor-aided Multi-user Channel Estimation for 3D Massive MIMO Communications
• A Survey of Bayesian Statistical Approaches for Big Data
• Semi-Implicit Variational Inference
• Approximate Inference with Amortised MCMC
• HMCLab: a framework for solving diverse geophysical inverse problems using the Hamiltonian Monte Carlo method
• Transport Analysis of Infinitely Deep Neural Network
• Neural Parametric Fokker-Planck Equations
• All You Need is a Good Functional Prior for Bayesian Deep Learning
• Riemannian Stein Variational Gradient Descent for Bayesian Inference
• Computational Mean-field information dynamics associated with Reaction diffusion equations
• Projected Stein Variational Gradient Descent
• Stein variational gradient descent with local approximations
• Bayesian Active Meta-Learning for Few Pilot Demodulation and Equalization
• A Bayesian Deep Learning Approach to Near-Term Climate Prediction
• Learning Model Reparametrizations: Implicit Variational Inference by Fitting MCMC distributions
• Inference via low-dimensional couplings
• Birth-death dynamics for sampling: Global convergence, approximations and their asymptotics
• Distributionally Adversarial Attack
• Navigating the Trade-Off between Multi-Task Learning and Learning to Multitask in Deep Neural Networks
• Geometric Methods for Sampling, Optimisation, Inference and Adaptive Agents
• Particle Optimization in Stochastic Gradient MCMC
• Parametric Fokker-Planck equation
• Energy-Based Sequence GANs for Recommendation and Their Connection to Imitation Learning
• Kernel embedding of maps for sequential Bayesian inference: The variational mapping particle filter
• Stochastic Maximum Likelihood Optimization via Hypernetworks
• Stochastic Stein Discrepancies
• Path Integral Sampler: a stochastic control approach for sampling
• Kernel Implicit Variational Inference
• Bayesian Multi-Task Transfer Learning for Soft Prompt Tuning
• Modelling Bounded Rationality in Multi-Agent Interactions by Generalized Recursive Reasoning
• Sqrt(d) Dimension Dependence of Langevin Monte Carlo
• Learning interaction kernels in stochastic systems of interacting particles from multiple trajectories
• Nearly Consistent Finite Particle Estimates in Streaming Importance Sampling
• Trust-Region Variational Inference with Gaussian Mixture Models
• Stein particle filtering
• SVGD as a kernelized Wasserstein gradient flow of the chi-squared divergence
• Convergence of Stein Variational Gradient Descent under a Weaker Smoothness Condition
• Annealed Flow Transport Monte Carlo
• Constructing sampling schemes via coupling: Markov semigroups and optimal transport
• A General Framework for Empirical Bayes Estimation in Discrete Linear Exponential Family
• A Fourier representation of kernel Stein discrepancy with application to Goodness-of-Fit tests for measures on infinite dimensional Hilbert spaces
• A Particle-Based Algorithm for Distributional Optimization on \textit{Constrained Domains} via Variational Transport and Mirror Descent
• Measure Transport with Kernel Stein Discrepancy
• Deep PDF: Probabilistic Surface Optimization and Density Estimation
• Learn Quasi-stationary Distributions of Finite State Markov Chain
• Active Slices for Sliced Stein Discrepancy
• The Renyi Gaussian Process: Towards Improved Generalization
• Stein Variational Goal Generation for adaptive Exploration in Multi-Goal Reinforcement Learning
• Doubly Adaptive Importance Sampling
• Integrable Nonparametric Flows
• Variational Inference MPC using Tsallis Divergence
• An Iterative Closest Points Approach to Neural Generative Models
• On Unifying Randomized Methods For Inverse Problems
• Variational Inference over Non-differentiable Cardiac Simulators using Bayesian Optimization
• Variational Inference for Policy Gradient
• Dynamic Flows on Curved Space Generated by Labeled Data
• Learning Approximately Objective Priors
• Accelerated Information Gradient flow
• Transfer Bayesian Meta-learning via Weighted Free Energy Minimization