Amortised Learning by Wake-Sleep

Abstract

Models that employ latent variables to capture structure in observed data lie at the heart of many current unsupervised learning algorithms, but exact maximum-likelihood learning for powerful and flexible latent-variable models is almost always intractable. Thus, state-of-the-art approaches either abandon the maximum-likelihood framework entirely, or else rely on a variety of variational approximations to the posterior distribution over the latents. Here, we propose an alternative approach that we call amortised learning. Rather than computing an approximation to the posterior over latents, we use a wake-sleep Monte-Carlo strategy to learn a function that directly estimates the maximum-likelihood parameter updates. Amortised learning is possible whenever samples of latents and observations can be simulated from the generative model, treating the model as a "black box". We demonstrate its effectiveness on a wide range of complex models, including those with latents that are discrete or supported on non-Euclidean spaces.

References in corpus (22)

• Auto-Encoding Variational Bayes
• Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
• GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
• Categorical Reparameterization with Gumbel-Softmax
• Variational Inference with Normalizing Flows
• Which Training Methods for GANs do actually Converge?
• Mining gold from implicit models to improve likelihood-free inference
• Implicit Reparameterization Gradients
• Variational Inference using Implicit Distributions
• On gradient regularizers for MMD GANs
• Gradient Estimators for Implicit Models
• A Wrapped Normal Distribution on Hyperbolic Space for Gradient-Based Learning
• Continuous Hierarchical Representations with Poincaré Variational Auto-Encoders
• The continuous Bernoulli: fixing a pervasive error in variational autoencoders
• DVAE++: Discrete Variational Autoencoders with Overlapping Transformations
• A Test of Relative Similarity For Model Selection in Generative Models
• Tighter Variational Bounds are Not Necessarily Better
• Variational Message Passing with Structured Inference Networks
• Flexible and accurate inference and learning for deep generative models
• Revisiting Reweighted Wake-Sleep for Models with Stochastic Control Flow
• Reweighted Expectation Maximization
• Kernel Recursive ABC: Point Estimation with Intractable Likelihood