Variational Walkback: Learning a Transition Operator as a Stochastic Recurrent Net
arXiv:1711.02282
Abstract
We propose a novel method to directly learn a stochastic transition operator whose repeated application provides generated samples. Traditional undirected graphical models approach this problem indirectly by learning a Markov chain model whose stationary distribution obeys detailed balance with respect to a parameterized energy function. The energy function is then modified so the model and data distributions match, with no guarantee on the number of steps required for the Markov chain to converge. Moreover, the detailed balance condition is highly restrictive: energy based models corresponding to neural networks must have symmetric weights, unlike biological neural circuits. In contrast, we develop a method for directly learning arbitrarily parameterized transition operators capable of expressing non-equilibrium stationary distributions that violate detailed balance, thereby enabling us to learn more biologically plausible asymmetric neural networks and more general non-energy based dynamical systems. The proposed training objective, which we derive via principled variational methods, encourages the transition operator to "walk back" in multi-step trajectories that start at data-points, as quickly as possible back to the original data points. We present a series of experimental results illustrating the soundness of the proposed approach, Variational Walkback (VW), on the MNIST, CIFAR-10, SVHN and CelebA datasets, demonstrating superior samples compared to earlier attempts to learn a transition operator. We also show that although each rapid training trajectory is limited to a finite but variable number of steps, our transition operator continues to generate good samples well past the length of such trajectories, thereby demonstrating the match of its non-equilibrium stationary distribution to the data distribution. Source Code: http://github.com/anirudh9119/walkback_nips17
To appear at NIPS 2017
References in corpus (11)
- Adam: A Method for Stochastic Optimization
- Stochastic Backpropagation and Approximate Inference in Deep Generative Models
- Deep Unsupervised Learning using Nonequilibrium Thermodynamics
- Improved Techniques for Training GANs
- DRAW: A Recurrent Neural Network For Image Generation
- Path ensembles averages in systems driven far-from-equilibrium
- Better Mixing via Deep Representations
- Random feedback weights support learning in deep neural networks
- Near-optimal protocols in complex nonequilibrium transformations
- Why are deep nets reversible: A simple theory, with implications for training
- Accurate and Conservative Estimates of MRF Log-likelihood using Reverse Annealing