Preconditioned stochastic gradient langevin dynamics for deep neural networks

Chunyuan Li, Changyou Chen, David Carlson, Lawrence Carin

Research output: Chapter in Book/Report/Conference proceedingConference contribution

68 Scopus citations

Abstract

Effective training of deep neural networks suffers from two main issues. The first is that the parameter spaces of these models exhibit pathological curvature. Recent methods address this problem by using adaptive preconditioning for Stochastic Gradient Descent (SGD). These methods improve convergence by adapting to the local geometry of parameter space. A second issue is overfitting, which is typically addressed by early stopping. However, recent work has demonstrated that Bayesian model averaging mitigates this problem. The posterior can be sampled by using Stochastic Gradient Langevin Dynamics (SGLD). However, the rapidly changing curvature renders default SGLD methods inefficient. Here, we propose combining adaptive preconditioners with SGLD. In support of this idea, we give theoretical properties on asymptotic convergence and predictive risk.We also provide empirical results for Logistic Regression, Feedforward Neural Nets, and Convolutional Neural Nets, demonstrating that our preconditioned SGLD method gives state-of-the-art performance on these models.
Original languageEnglish (US)
Title of host publication30th AAAI Conference on Artificial Intelligence, AAAI 2016
PublisherAAAI press
Pages1788-1794
Number of pages7
ISBN (Print)9781577357605
StatePublished - Jan 1 2016
Externally publishedYes

Fingerprint Dive into the research topics of 'Preconditioned stochastic gradient langevin dynamics for deep neural networks'. Together they form a unique fingerprint.

Cite this