Importance Sampled Stochastic Optimization for Variational Inference

Show full item record



Permalink

http://hdl.handle.net/10138/232532

Citation

Sakaya , J & Klami , A 2017 , Importance Sampled Stochastic Optimization for Variational Inference . in G Elidan & K Kersting (eds) , 33rd Conference on Uncertainty in Artificial Intelligence 2017 : Sydney, Australia, 11-15 August 2017 . AUAI Press , Conference on Uncertainty in Artificial Intelligence , Sydney , Australia , 11/08/2017 . < http://auai.org/uai2017/proceedings/papers/138.pdf >

Title: Importance Sampled Stochastic Optimization for Variational Inference
Author: Sakaya, Joseph; Klami, Arto
Editor: Elidan, Gal; Kersting, Kristian
Contributor: University of Helsinki, Helsinki Institute for Information Technology
University of Helsinki, Department of Computer Science
Publisher: AUAI Press
Date: 2017-08
Language: eng
Number of pages: 10
Belongs to series: 33rd Conference on Uncertainty in Artificial Intelligence 2017 Sydney, Australia, 11-15 August 2017
ISBN: 978-1-5108-4779-8
URI: http://hdl.handle.net/10138/232532
Abstract: Variational inference approximates the posterior distribution of a probabilistic model with a parameterized density by maximizing a lower bound for the model evidence. Modern solutions fit a flexible approximation with stochastic gradient descent, using Monte Carlo approximation for the gradients. This enables variational inference for arbitrary differentiable probabilistic models, and consequently makes variational inference feasible for probabilistic programming languages. In this work we develop more efficient inference algorithms for the task by considering importance sampling estimates for the gradients. We show how the gradient with respect to the approximation parameters can often be evaluated efficiently without needing to re-compute gradients of the model itself, and then proceed to derive practical algorithms that use importance sampled estimates to speed up computation. We present importance sampled stochastic gradient descent that outperforms standard stochastic gradient descent by a clear margin for a range of models, and provide a justifiable variant of stochastic average gradients for variational inference.
Subject: 113 Computer and information sciences
Rights:


Files in this item

Total number of downloads: Loading...

Files Size Format View
138.pdf 821.7Kb PDF View/Open

This item appears in the following Collection(s)

Show full item record