1 paper · 1 filter
Alexandre Rio, Merwan Barlier, Igor Colin +1
We address private deep offline reinforcement learning (RL), where the goal is to train a policy on standard control tasks that is differentially private (DP) with respect to indiv…