Generalized Energy Based Models

Abstract

We introduce the Generalized Energy Based Model (GEBM) for generative modelling. These models combine two trained components: a base distribution (generally an implicit model), which can learn the support of data with low intrinsic dimension in a high dimensional space; and an energy function, to refine the probability mass on the learned support. Both the energy function and base jointly constitute the final model, unlike GANs, which retain only the base distribution (the "generator"). GEBMs are trained by alternating between learning the energy and the base. We show that both training stages are well-defined: the energy is learned by maximising a generalized likelihood, and the resulting energy-based loss provides informative gradients for learning the base. Samples from the posterior on the latent space of the trained model can be obtained via MCMC, thus finding regions in this space that produce better quality samples. Empirically, the GEBM samples on image-generation tasks are of much better quality than those from the learned generator alone, indicating that all else being equal, the GEBM will outperform a GAN of the same complexity. When using normalizing flows as base measures, GEBMs succeed on density modelling tasks, returning comparable performance to direct maximum likelihood of the same networks.

References in corpus (16)

• GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
• MADE: Masked Autoencoder for Distribution Estimation
• Underdamped Langevin MCMC: A non-asymptotic analysis
• Your Classifier is Secretly an Energy Based Model and You Should Treat it Like One
• Gradient descent GAN optimization is locally stable
• Approximation and Convergence Properties of Generative Adversarial Learning
• Deep Compressed Sensing
• LOGAN: Latent Optimisation for Generative Adversarial Networks
• Your GAN is Secretly an Energy-based Model and You Should use Discriminator Driven Latent Sampling
• Flows for simultaneous manifold learning and density estimation
• Efficient Algorithms for Smooth Minimax Optimization
• Fractional Underdamped Langevin Dynamics: Retargeting SGD with Momentum under Heavy-Tailed Gradient Noise
• Subsampling Generative Adversarial Networks: Density Ratio Estimation in Feature Space with Softplus Loss
• Smoothness and Stability in GANs
• Discriminator optimal transport
• Training Deep Energy-Based Models with f-Divergence Minimization

Cited by in corpus (3)

• A case for new neural network smoothness constraints
• Learning Discrete Energy-based Models via Auxiliary-variable Local Exploration
• Particle Dynamics for Learning EBMs