Learning Deep Generative Models With Doubly Stochastic Gradient MCMC

Deep generative models (DGMs), which are often organized in a hierarchical manner, provide a principled framework of capturing the underlying causal factors of data. Recent work on DGMs focussed on the development of efficient and scalable variational inference methods that learn a single model under some mean-field or parameterization assumptions. However, little work has been done on extending Markov chain Monte Carlo (MCMC) methods to Bayesian DGMs, which enjoy many advantages compared with variational methods. We present doubly stochastic gradient MCMC, a simple and generic method for (approximate) Bayesian inference of DGMs in a collapsed continuous parameter space. At each MCMC sampling step, the algorithm randomly draws a mini-batch of data samples to estimate the gradient of log-posterior and further estimates the intractable expectation over hidden variables via a neural adaptive importance sampler, where the proposal distribution is parameterized by a deep neural network and learnt jointly along with the sampling process. We demonstrate the effectiveness of learning various DGMs on a wide range of tasks, including density estimation, data generation, and missing data imputation. Our method outperforms many state-of-the-art competitors.

[1]  Koray Kavukcuoglu,et al.  Pixel Recurrent Neural Networks , 2016, ICML.

[2]  Neural Adaptive Sequential Monte Carlo Supplementary Material , 2015 .

[3]  Tianqi Chen,et al.  Stochastic Gradient Hamiltonian Monte Carlo , 2014, ICML.

[4]  Brendan J. Frey,et al.  Graphical Models for Machine Learning and Digital Communication , 1998 .

[5]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[6]  Tara N. Sainath,et al.  FUNDAMENTAL TECHNOLOGIES IN MODERN SPEECH RECOGNITION Digital Object Identifier 10.1109/MSP.2012.2205597 , 2012 .

[7]  Yoshua Bengio,et al.  Gradient-based learning applied to document recognition , 1998, Proc. IEEE.

[8]  Dustin Tran,et al.  Variational Gaussian Process , 2015, ICLR.

[9]  Zhe Gan,et al.  Learning Deep Sigmoid Belief Networks with Data Augmentation , 2015, AISTATS.

[10]  Tapani Raiko,et al.  Enhanced Gradient for Training Restricted Boltzmann Machines , 2013, Neural Computation.

[11]  Jason Yosinski,et al.  Deep neural networks are easily fooled: High confidence predictions for unrecognizable images , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[12]  Jeffrey Pennington,et al.  Dynamic Pooling and Unfolding Recursive Autoencoders for Paraphrase Detection , 2011, NIPS.

[13]  Yee Whye Teh,et al.  A Fast Learning Algorithm for Deep Belief Nets , 2006, Neural Computation.

[14]  Geoffrey E. Hinton.  Training Products of Experts by Minimizing Contrastive Divergence , 2002, Neural Computation.

[15]  Yoshua Bengio,et al.  Understanding the difficulty of training deep feedforward neural networks , 2010, AISTATS.

[16]  Michael I. Jordan,et al.  An Introduction to Variational Methods for Graphical Models , 1999, Machine-mediated learning.

[17]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[18]  Julien Cornebise,et al.  Weight Uncertainty in Neural Network , 2015, ICML.

[19]  Tapani Raiko,et al.  Iterative Neural Autoregressive Distribution Estimator NADE-k , 2014, NIPS.

[20]  Yoshua Bengio,et al.  Deep Generative Stochastic Networks Trainable by Backprop , 2013, ICML.

[21]  Alexander Binder,et al.  Evaluating the Visualization of What a Deep Neural Network Has Learned , 2015, IEEE Transactions on Neural Networks and Learning Systems.

[22]  Julien Cornebise,et al.  Weight Uncertainty in Neural Networks , 2015, ArXiv.

[23]  Max Welling,et al.  Auto-Encoding Variational Bayes , 2013, ICLR.

[24]  Ruslan Salakhutdinov,et al.  Importance Weighted Autoencoders , 2015, ICLR.

[25]  Michael I. Jordan,et al.  Mean Field Theory for Sigmoid Belief Networks , 1996, J. Artif. Intell. Res..

[26]  Geoffrey E. Hinton,et al.  The "wake-sleep" algorithm for unsupervised neural networks. , 1995, Science.

[27]  Kai Fan,et al.  High-Order Stochastic Gradient Thermostats for Bayesian Learning of Deep Models , 2015, AAAI.

[28]  Nando de Freitas,et al.  Inductive Principles for Restricted Boltzmann Machine Learning , 2010, AISTATS.

[29]  Geoffrey E. Hinton,et al.  Bayesian Learning for Neural Networks , 1995 .

[30]  Yoshua Bengio,et al.  Reweighted Wake-Sleep , 2014, ICLR.

[31]  Shane Legg,et al.  Human-level control through deep reinforcement learning , 2015, Nature.

[32]  Yali Amit,et al.  Mixtures of Sparse Autoregressive Networks , 2015 .

[33]  Yoshua Bengio,et al.  Modeling Temporal Dependencies in High-Dimensional Sequences: Application to Polyphonic Music Generation and Transcription , 2012, ICML.

[34]  Hugo Larochelle,et al.  The Neural Autoregressive Distribution Estimator , 2011, AISTATS.

[35]  Daan Wierstra,et al.  Deep AutoRegressive Networks , 2013, ICML.

[36]  Karol Gregor,et al.  Neural Variational Inference and Learning in Belief Networks , 2014, ICML.

[37]  Alex Graves,et al.  DRAW: A Recurrent Neural Network For Image Generation , 2015, ICML.

[38]  Bo Zhang,et al.  Learning to Generate with Memory , 2016, ICML.

[39]  David J. C. MacKay,et al.  A Practical Bayesian Framework for Backpropagation Networks , 1992, Neural Computation.

[40]  Jen-Tzung Chien,et al.  Bayesian Recurrent Neural Network for Language Modeling , 2016, IEEE Transactions on Neural Networks and Learning Systems.

[41]  Sean Gerrish,et al.  Black Box Variational Inference , 2013, AISTATS.

[42]  Ole Winther,et al.  Ladder Variational Autoencoders , 2016, NIPS.

[43]  Ryan P. Adams,et al.  Learning the Structure of Deep Sparse Graphical Models , 2009, AISTATS.

[44]  Zhe Gan,et al.  Scalable Deep Poisson Factor Analysis for Topic Modeling , 2015, ICML.

[45]  H. Robbins.  A Stochastic Approximation Method , 1951 .

[46]  Pieter Abbeel,et al.  Max-margin Classification of Data with Absent Features , 2008, J. Mach. Learn. Res..

[47]  Ryan Babbush,et al.  Bayesian Sampling Using Stochastic Gradient Thermostats , 2014, NIPS.

[48]  Radford M. Neal.  Connectionist Learning of Belief Networks , 1992, Artif. Intell..

[49]  Daan Wierstra,et al.  Stochastic Backpropagation and Approximate Inference in Deep Generative Models , 2014, ICML.

[50]  Miguel Lázaro-Gredilla,et al.  Doubly Stochastic Variational Bayes for non-Conjugate Inference , 2014, ICML.

[51]  Hugo Larochelle,et al.  A Deep and Tractable Density Estimator , 2013, ICML.

[52]  Yee Whye Teh,et al.  Bayesian Learning via Stochastic Gradient Langevin Dynamics , 2011, ICML.

[53]  Geoffrey E. Hinton,et al.  Deep Boltzmann Machines , 2009, AISTATS.

[54]  Ruslan Salakhutdinov,et al.  Evaluating probabilities under high-dimensional latent variable models , 2008, NIPS.

[55]  Andriy Mnih,et al.  Variational Inference for Monte Carlo Objectives , 2016, ICML.

[56]  Ahn.  Bayesian Posterior Sampling via Stochastic Gradient Fisher Scoring , 2012 .

[57]  Samy Bengio,et al.  Modeling High-Dimensional Discrete Data with Multi-Layer Neural Networks , 1999, NIPS.

[58]  Ruslan Salakhutdinov,et al.  On the quantitative analysis of deep belief networks , 2008, ICML '08.

[59]  Pierre Priouret,et al.  Adaptive Algorithms and Stochastic Approximations , 1990, Applications of Mathematics.

[60]  Joshua B. Tenenbaum,et al.  One-shot learning by inverting a compositional causal process , 2013, NIPS.