Robust approachability and regret minimization in games with partial monitoring

Abstract : Approachability has become a standard tool in analyzing earning algorithms in the adversarial online learning setup. We develop a variant of approachability for games where there is ambiguity in the obtained reward that belongs to a set, rather than being a single vector. Using this variant we tackle the problem of approachability in games with partial monitoring and develop simple and efficient algorithms (i.e., with constant per-step complexity) for this setup. We finally consider external regret and internal regret in repeated games with partial monitoring and derive regret-minimizing strategies based on approachability theory.
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-00595695
Contributor : Gilles Stoltz <>
Submitted on : Wednesday, February 15, 2012 - 3:19:50 PM
Last modification on : Tuesday, April 2, 2019 - 2:15:12 PM
Long-term archiving on : Wednesday, May 16, 2012 - 2:35:30 AM

Files

Mannor-Perchet-Stoltz--Robust-...
Files produced by the author(s)

Identifiers

  • HAL Id : hal-00595695, version 3
  • ARXIV : 1105.4995

Citation

Shie Mannor, Vianney Perchet, Gilles Stoltz. Robust approachability and regret minimization in games with partial monitoring. 2012. ⟨hal-00595695v3⟩

Share

Metrics

Record views

622

Files downloads

333