Learning Energy-Based Models by Diffusion Recovery Likelihood

Abstract

While energy-based models (EBMs) exhibit a number of desirable properties, training and sampling on high-dimensional datasets remains challenging. Inspired by recent progress on diffusion probabilistic models, we present a diffusion recovery likelihood method to tractably learn and sample from a sequence of EBMs trained on increasingly noisy versions of a dataset. Each EBM is trained with recovery likelihood, which maximizes the conditional probability of the data at a certain noise level given their noisy versions at a higher noise level. Optimizing recovery likelihood is more tractable than marginal likelihood, as sampling from the conditional distributions is much easier than sampling from the marginal distributions. After training, synthesized images can be generated by the sampling process that initializes from Gaussian white noise distribution and progressively samples the conditional distributions at decreasingly lower noise levels. Our method generates high fidelity samples on various image datasets. On unconditional CIFAR-10 our method achieves FID 9.58 and inception score 8.30, superior to the majority of GANs. Moreover, we demonstrate that unlike previous work on EBMs, our long-run MCMC samples from the conditional distributions do not diverge and still represent realistic images, allowing us to accurately estimate the normalized density of data even for high-dimensional datasets. Our implementation is available at https://github.com/ruiqigao/recovery_likelihood.

References in corpus (20)

• Denoising Diffusion Probabilistic Models
• GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
• Conditional Image Generation with PixelCNN Decoders
• Wide Residual Networks
• Deep Unsupervised Learning using Nonequilibrium Thermodynamics
• LSUN: Construction of a Large-scale Image Dataset using Deep Learning with Humans in the Loop
• Energy-based Generative Adversarial Network
• Generating Long Sequences with Sparse Transformers
• A Connection between Generative Adversarial Networks, Inverse Reinforcement Learning, and Energy-Based Models
• Generalized Denoising Auto-Encoders as Generative Models
• A Theory of Generative ConvNet
• Improved Techniques for Training Score-Based Generative Models
• Implicit Generation and Generalization in Energy-Based Models
• Your Classifier is Secretly an Energy Based Model and You Should Treat it Like One
• Deep Directed Generative Models with Energy-Based Probability Estimation
• Maximum Entropy Generators for Energy-Based Models
• Deep Energy Estimator Networks
• Introspective Classification with Convolutional Nets
• Quality Aware Generative Adversarial Networks
• Measuring the reliability of MCMC inference with bidirectional Monte Carlo

Cited by in corpus (8)

• Knowledge Distillation in Iterative Generative Models for Improved Sampling Speed
• Diffusion Schrödinger Bridge with Applications to Score-Based Generative Modeling
• On Fast Sampling of Diffusion Probabilistic Models
• Noise Estimation for Generative Diffusion Models
• Non Gaussian Denoising Diffusion Models
• Path Integral Sampler: a stochastic control approach for sampling
• Denoising Diffusion Gamma Models
• Particle Dynamics for Learning EBMs