Learning Model Reparametrizations: Implicit Variational Inference by Fitting MCMC distributions
arXiv:1708.01529
Abstract
We introduce a new algorithm for approximate inference that combines reparametrization, Markov chain Monte Carlo and variational methods. We construct a very flexible implicit variational distribution synthesized by an arbitrary Markov chain Monte Carlo operation and a deterministic transformation that can be optimized using the reparametrization trick. Unlike current methods for implicit variational inference, our method avoids the computation of log density ratios and therefore it is easily applicable to arbitrary continuous and differentiable models. We demonstrate the proposed algorithm for fitting banana-shaped distributions and for training variational autoencoders.
16 pages, 6 figures
References in corpus (9)
- Auto-Encoding Variational Bayes
- Stochastic Backpropagation and Approximate Inference in Deep Generative Models
- Markov Chain Monte Carlo and Variational Inference: Bridging the Gap
- Stein Variational Gradient Descent: A General Purpose Bayesian Inference Algorithm
- Automatic Differentiation Variational Inference
- Learning to Draw Samples: With Application to Amortized MLE for Generative Adversarial Learning
- Variational Inference using Implicit Distributions
- Approximate Inference with Amortised MCMC
- Adversarial Message Passing For Graphical Models
Cited by in corpus (12)
- Normalizing Flows for Probabilistic Modeling and Inference
- NeuTra-lizing Bad Geometry in Hamiltonian Monte Carlo Using Neural Transport
- ZhuSuan: A Library for Bayesian Deep Learning
- Advances in Variational Inference
- Subspace Inference for Bayesian Deep Learning
- Bayesian Neural Ordinary Differential Equations
- Adaptive Monte Carlo augmented with normalizing flows
- Conditional deep surrogate models for stochastic, high-dimensional, and multi-fidelity systems
- Cycle-Consistent Adversarial Learning as Approximate Bayesian Inference
- Efficient Bayesian Sampling Using Normalizing Flows to Assist Markov Chain Monte Carlo Methods
- MCMC-Interactive Variational Inference
- Reparameterized Sampling for Generative Adversarial Networks