Saved in:
Bibliographic Details
Main Authors: Cundy, Chris, Desai, Rishi, Ermon, Stefano
Format: Preprint
Published: 2020
Subjects:
Online Access:https://arxiv.org/abs/2012.15019
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914755556409344
author Cundy, Chris
Desai, Rishi
Ermon, Stefano
author_facet Cundy, Chris
Desai, Rishi
Ermon, Stefano
contents As reinforcement learning techniques are increasingly applied to real-world decision problems, attention has turned to how these algorithms use potentially sensitive information. We consider the task of training a policy that maximizes reward while minimizing disclosure of certain sensitive state variables through the actions. We give examples of how this setting covers real-world problems in privacy for sequential decision-making. We solve this problem in the policy gradients framework by introducing a regularizer based on the mutual information (MI) between the sensitive state and the actions. We develop a model-based stochastic gradient estimator for optimization of privacy-constrained policies. We also discuss an alternative MI regularizer that serves as an upper bound to our main MI regularizer and can be optimized in a model-free setting, and a powerful direct estimator that can be used in an environment with differentiable dynamics. We contrast previous work in differentially-private RL to our mutual-information formulation of information disclosure. Experimental results show that our training method results in policies that hide the sensitive state, even in challenging high-dimensional tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2012_15019
institution arXiv
publishDate 2020
record_format arxiv
spellingShingle Privacy-Constrained Policies via Mutual Information Regularized Policy Gradients
Cundy, Chris
Desai, Rishi
Ermon, Stefano
Machine Learning
Cryptography and Security
As reinforcement learning techniques are increasingly applied to real-world decision problems, attention has turned to how these algorithms use potentially sensitive information. We consider the task of training a policy that maximizes reward while minimizing disclosure of certain sensitive state variables through the actions. We give examples of how this setting covers real-world problems in privacy for sequential decision-making. We solve this problem in the policy gradients framework by introducing a regularizer based on the mutual information (MI) between the sensitive state and the actions. We develop a model-based stochastic gradient estimator for optimization of privacy-constrained policies. We also discuss an alternative MI regularizer that serves as an upper bound to our main MI regularizer and can be optimized in a model-free setting, and a powerful direct estimator that can be used in an environment with differentiable dynamics. We contrast previous work in differentially-private RL to our mutual-information formulation of information disclosure. Experimental results show that our training method results in policies that hide the sensitive state, even in challenging high-dimensional tasks.
title Privacy-Constrained Policies via Mutual Information Regularized Policy Gradients
topic Machine Learning
Cryptography and Security
url https://arxiv.org/abs/2012.15019