Privacy-Constrained Policies via Mutual Information Regularized Policy Gradients
Abstract
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.
- Publication:
-
arXiv e-prints
- Pub Date:
- December 2020
- DOI:
- 10.48550/arXiv.2012.15019
- arXiv:
- arXiv:2012.15019
- Bibcode:
- 2020arXiv201215019C
- Keywords:
-
- Computer Science - Machine Learning;
- Computer Science - Cryptography and Security
- E-Print:
- Accepted to AISTATS 2024