Summary
The 2018 International Symposium on Information Theory and Its Applications　(ISITA2018)
2018
Session Number:Mo-AM-1-2
Session:
Number:Mo-AM-1-2.2
Fast yet Simple Natural-Gradient Descent for Variational Inference in Complex Models
Mohammad Emtiyaz Khan, Didrik Nielsen,
pp.31-35
Publication Date:2018/10/18
Online ISSN:2188-5079
DOI:10.34385/proc.55.Mo-AM-1-2.2
PDF download
Summary:
Bayesian inference plays an important role in advancing machine learning, but faces computational challenges when applied to complex models such as deep neural networks. Variational inference circumvents these challenges by formulating Bayesian inference as an optimization problem and solving it using gradient-based optimization. In this paper, we argue in favor of natural-gradient approaches which, unlike their gradientbased counterparts, can improve convergence by exploiting the information geometry of the solutions.We show how to derive fast yet simple natural-gradient updates by using a duality associated with exponential-family distributions. An attractive feature of these methods is that, by using natural-gradients, they are able to extract accurate local approximations for individual model components. We summarize recent results for Bayesian deep learning showing the superiority of natural-gradient approaches over their gradient counterparts.