Scalable Approximate Inference and Some Applications

Abstract

Approximate inference in probability models is a fundamental task in machine learning. Approximate inference provides powerful tools to Bayesian reasoning, decision making, and Bayesian deep learning. The main goal is to estimate the expectation of interested functions w.r.t. a target distribution. When it comes to high dimensional probability models and large datasets, efficient approximate inference becomes critically important. In this thesis, we propose a new framework for approximate inference, which combines the advantages of these three frameworks and overcomes their limitations. Our proposed four algorithms are motivated by the recent computational progress of Stein's method. Our proposed algorithms are applied to continuous and discrete distributions under the setting when the gradient information of the target distribution is available or unavailable. Theoretical analysis is provided to prove the convergence of our proposed algorithms. Our adaptive IS algorithm iteratively improves the importance proposal by functionally decreasing the KL divergence between the updated proposal and the target. When the gradient of the target is unavailable, our proposed sampling algorithm leverages the gradient of a surrogate model and corrects induced bias with importance weights, which significantly outperforms other gradient-free sampling algorithms. In addition, our theoretical results enable us to perform the goodness-of-fit test on discrete distributions. At the end of the thesis, we propose an importance-weighted method to efficiently aggregate local models in distributed learning with one-shot communication. Results on simulated and real datasets indicate the statistical efficiency and wide applicability of our algorithm.

PhD thesis, Dartmouth College (2019)

References in corpus (32)

• Auto-Encoding Variational Bayes
• Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation
• Variational Inference with Normalizing Flows
• Binarized Neural Networks
• The pseudo-marginal approach for efficient Monte Carlo computations
• Optimal Distributed Online Prediction using Mini-Batches
• Communication Efficient Distributed Optimization using an Approximate Newton-type Method
• Bayesian Dark Knowledge
• Adversarial Variational Bayes: Unifying Variational Autoencoders and Generative Adversarial Networks
• Stein Variational Gradient Descent as Gradient Flow
• Measuring Sample Quality with Stein's Method
• A Kernel Test of Goodness of Fit
• Measuring Sample Quality with Kernels
• Learning to Draw Samples: With Application to Amortized MLE for Generative Adversarial Learning
• Stein Variational Policy Gradient
• Understanding Straight-Through Estimator in Training Activation Quantized Neural Nets
• Discontinuous Hamiltonian Monte Carlo for discrete parameters and discontinuous likelihoods
• Stein Variational Gradient Descent With Matrix-Valued Kernels
• Distributed Estimation, Information Loss and Exponential Families
• Probabilistic Path Hamiltonian Monte Carlo
• Gradient-free Hamiltonian Monte Carlo with Efficient Kernel Exponential Families
• Kernel Adaptive Metropolis-Hastings
• Stein Variational Gradient Descent Without Gradient
• Advances in Variational Inference
• Hamiltonian Annealed Importance Sampling for partition function estimation
• Stein Variational Adaptive Importance Sampling
• Adaptive Importance Sampling via Stochastic Convex Programming
• Variational Inference with Tail-adaptive f-Divergence
• Action-depedent Control Variates for Policy Optimization via Stein's Identity
• Binary Ensemble Neural Network: More Bits per Network or More Networks per Bit?
• High efficiently numerical simulation of the TDGL equation with reticular free energy in hydrogel
• Taxable Stock Trading with Deep Reinforcement Learning