Privacy-Constrained Policies via Mutual Information Regularized Policy Gradients

arXiv:2012.15019v3 Announce Type: replace-cross
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 …

actions algorithms arxiv attention cs.cr cs.lg decision disclosure examples information policies policy privacy problems real reward sensitive sensitive information state task techniques training world