Research map / Federated learning and privacy / Differential privacy
Differentially Private Regret Minimization in Episodic Markov Decision Processes
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use a central private count mechanism for JDP rather than local privatization | Private-UCB-PO and Private-UCB-VI · Finite-horizon tabular MDP | regret bound privacy cost: JDP: S^2AH^3/ε; LDP: S^2A√(H^5T)/ε | Helped |
| Use value-aware concentration for the optimal value function and a Bernstein-type transition bound | Private-UCB-VI · Finite-horizon tabular MDP | privacy-independent regret term: \sqrt{SAH^3T} vs. \sqrt{S^2AH^3T} | Helped |
| Apply the Private-UCB-VI framework instead of Private-UCB-PO | Private-UCB-PO · Finite-horizon tabular MDP, JDP and LDP | privacy-independent regret term: \sqrt{S^2AH^3T} -> \sqrt{SAH^3T} | Helped |
| Use a local privatizer instead of a central privatizer | Private-UCB-PO and Private-UCB-VI · Finite-horizon tabular MDP | privacy cost in regret bound: JDP: S^2AH^3/ε; LDP: S^2A√(H^5 T)/ε | Hurt |
Closest papers by meaning
- Local Differential Privacy for Regret Minimization in Reinforcement Learning (NeurIPS 2021 · similarity 0.96)
- Differentially Private Reinforcement Learning with Self-Play (NeurIPS 2024 · similarity 0.92)
- Towards Differentially Private Reinforcement Learning with General Function Approximation (NeurIPS 2026 · similarity 0.89)
- On the Sample Complexity of Differentially Private Policy Optimization (NeurIPS 2025 · similarity 0.89)
- Connecting Thompson Sampling and UCB: Towards More Efficient Trade-offs Between Privacy and Regret (ICML 2025 · similarity 0.88)
- Achieving Tractable Minimax Optimal Regret in Average Reward MDPs (NeurIPS 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.