Low-variance black-box gradient estimates for the Plackett-Luce distribution

Gadetsky, Artyom, Struminsky, Kirill, Robinson, Christopher, Quadrianto, Novi and Vetrov, Dmitry (2020) Low-variance black-box gradient estimates for the Plackett-Luce distribution. Thirty-Fourth AAAI Conference on Artificial Intelligence, Hilton New York Midtown, New York, New York, USA, February 7-12 2020. Published in: Thirty-Fourth AAAI Conference on Artificial Intelligence. 34 (6) 10126-10135. AAAI ISSN 2159-5399 ISBN 9781577358350

[img] PDF - Other
Restricted to SRO admin only

Download (331kB)
[img] PDF (Main) - Accepted Version
Available under License All Rights Reserved.

Download (459kB)
[img] PDF (Supplementary) - Accepted Version
Available under License All Rights Reserved.

Download (279kB)

Abstract

Learning models with discrete latent variables using stochastic gradient descent remains a challenge due to the high variance of gradient estimates. Modern variance reduction techniques mostly consider categorical distributions and have limited applicability when the number of possible outcomes becomes large. In this work, we consider models with latent permutations and propose control variates for the Plackett-Luce distribution. In particular, the control variates allow us to optimize black-box functions over permutations using stochastic gradient descent. To illustrate the approach, we consider a variety of causal structure learning tasks for continuous and discrete data. We show that our method outperforms competitive relaxation-based optimization methods and is also applicable to non-differentiable score functions.

Item Type: Conference Proceedings
Schools and Departments: School of Engineering and Informatics > Informatics
Research Centres and Groups: Data Science Research Group
Related URLs:
Depositing User: Novi Quadrianto
Date Deposited: 25 Nov 2019 11:26
Last Modified: 19 Jan 2021 15:27
URI: http://sro.sussex.ac.uk/id/eprint/88244

View download statistics for this item

📧 Request an update
Project NameSussex Project NumberFunderFunder Ref
EthicalML: Injecting Ethical and Legal Constraints into Machine Learning ModelsG2034EPSRC-ENGINEERING & PHYSICAL SCIENCES RESEARCH COUNCILUnset