Research map / Federated learning and privacy / Differential privacy
Differentially Private Policy Evaluation
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 LSW instead of LSL for non-private Monte Carlo policy evaluation | Monte Carlo policy evaluation · Synthetic chain MDP | convergence speed | Helped |
| Use DP-LSW or DP-LSL with sufficiently large batches | Differentially private policy evaluation · Synthetic chain MDP | RMSE | No clear effect |
| Use LSL rather than LSW with small batches | Policy evaluation · Synthetic chain MDP | RMSE | Helped |
| Use LSW rather than LSL with large batches | Policy evaluation · Synthetic chain MDP | RMSE | Helped |
| Set the DP-LSL regularization schedule proportional to the square root of batch size | DP-LSL · Synthetic chain MDP | RMSE | Helped |
| Use more aggressive function approximation to estimate fewer parameters | DP-LSW and DP-LSL · Synthetic chain MDP with state aggregation | convergence speed | Helped |
| Increase the batch size | DP-LSW and DP-LSL · Synthetic chain MDP | RMSE | Helped |
Closest papers by meaning
- On the Sample Complexity of Differentially Private Policy Optimization (NeurIPS 2025 · similarity 0.93)
- Towards Differentially Private Reinforcement Learning with General Function Approximation (NeurIPS 2026 · similarity 0.90)
- Differentially Private Reinforcement Learning with Self-Play (NeurIPS 2024 · similarity 0.90)
- Shifted Interpolation for Differential Privacy (ICML 2024 · similarity 0.85)
- Connecting Thompson Sampling and UCB: Towards More Efficient Trade-offs Between Privacy and Regret (ICML 2025 · similarity 0.85)
- Instance-Wise Laplace Mechanism via Deep Reinforcement Learning (Student Abstract) (AAAI 2024 · similarity 0.84)
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.