KL-regularization Itself is Differentially Private in Bandits and RLHF

Documenti analoghi

• Efficient Policy Optimization in Robust Constrained MDPs with Iteration Complexity Guarantees
  di: Ganguly, Sourav, et al.
  Pubblicazione: (2025)
• Hybrid Transfer Reinforcement Learning: Provable Sample Efficiency from Shifted-Dynamics Data
  di: Qu, Chengrui, et al.
  Pubblicazione: (2024)
• Differentially Private Data Release on Graphs: Inefficiencies and Unfairness
  di: Fioretto, Ferdinando, et al.
  Pubblicazione: (2024)
• Distributionally Robust Constrained Reinforcement Learning under Strong Duality
  di: Zhang, Zhengfei, et al.
  Pubblicazione: (2024)
• Model-Free Robust $ϕ$-Divergence Reinforcement Learning Using Both Offline and Online Data
  di: Panaganti, Kishan, et al.
  Pubblicazione: (2024)