Metraipolis › archive › dc_e97d37545df5

Vanilla Policy Optimization Is Both Optimal and Differentially Private for Stochastic Contextual Bandits — arXiv:2609.33888

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.33888
License (as archived)
CC0-1.0
Archived
2026-10-07T22:37:50.604Z
SHA-256 of the archived text
a4f108d05e3af3fb77e747ae36d39d47f280d4e80ff10eb58fdc5e719879bda8
Stable URL of this page
https://metraipolis.com/documento/dc_e97d37545df5

How to cite

Vanilla Policy Optimization Is Both Optimal and Differentially Private for Stochastic Contextual Bandits — arXiv:2609.33888 — arXiv (en), https://arxiv.org/abs/2609.33888 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e97d37545df5 (Metraipolis archive dc_e97d37545df5, sha256 a4f108d05e3af3fb77e747ae36d39d47f280d4e80ff10eb58fdc5e719879bda8)
BibTeX
@misc{dce97d37545df5,
  title = {Vanilla Policy Optimization Is Both Optimal and Differentially Private for Stochastic Contextual Bandits — arXiv:2609.33888},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.33888},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e97d37545df5; sha256 a4f108d05e3af3fb77e747ae36d39d47f280d4e80ff10eb58fdc5e719879bda8},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.33888. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Can vanilla policy optimization explore enough to achieve near-optimal regret in stochastic contextual bandits? We show that standard exponential policy updates driven by offline regression do so under realizability, without exploration bonuses or importance weighting. For $A$ actions, $T$ rounds, and a finite prediction class $F$, vanilla PO achieves $\widetilde O(\sqrt{AT\log(|F|)})$ regret with high probability. Our analysis reveals an implicit exploration mechanism of independent interest: gradual policy updates prevent actions from losing probability too quickly, allowing the regression oracle to learn their expected losses. We further develop a batched version using only $O(\log T)$ regression calls and policy switches, and show how private regression oracles yield differentially private contextual bandit algorithms without composition across batches. For a finite class, this gives pure $\varepsilon_{\rm priv}$-DP and regret $\widetilde O\left( \sqrt{AT \log(|F|/\delta)}(1+\varepsilon_{\rm priv}^{-1/2}) \right)$. Finally, experiments across oracle-based contextual bandit algorithms, with and without privacy, demonstrate the practical effectiveness of policy optimization and the value of explicit exploration under stronger privacy constraints.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.33888 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-07T22:14:28.052Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).