Optimistic Thompson Sampling for No-Regret Learning in Unknown Games
IEEE Transactions on Signal Processing · 2026 · 74 · 2434-2447
DOI: 10.1109/tsp.2026.3695892 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Signal Processing, 2026. Verfasst von Yingru Li und 4 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 35 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Low-Altitude Airspace Strategic Program Portfolio, Shenzhen Research Institute of Big Data.
Abstract (Originalsprache)
We study the problem of learning to play a repeated multi-player game with an unknown reward function and bandit feedback. The central challenge arises from the need to balance exploration and exploitation under bandit feedback while strategically responding to other players. To address this, we propose Thompson Sampling-based algorithms that leverage available information about opponents’ actions and reward structures, resulting in a significant reduction in regret bound. Building on these insights, we introduce the Optimism-then-NoRegret framework, which encompasses various game algorithms as special cases. Simulation evaluations on three distinct types of games show that our proposed algorithms consistently and substantially outperform standard baselines.
Förderung
- National Natural Science Foundation of China
- Low-Altitude Airspace Strategic Program Portfolio
- Shenzhen Research Institute of Big Data
Zitieren
APA
Li, Y., Liu, L., Pu, W., Liang, H., & Luo, Z. (2026). Optimistic Thompson Sampling for No-Regret Learning in Unknown Games. IEEE Transactions on Signal Processing, 74, 2434-2447. https://doi.org/10.1109/tsp.2026.3695892
Vancouver
Li Y, Liu L, Pu W, Liang H, Luo Z. Optimistic Thompson Sampling for No-Regret Learning in Unknown Games. IEEE Transactions on Signal Processing. 2026;74:2434-2447. doi:10.1109/tsp.2026.3695892
BibTeX
@article{Li2026,
author = {Li, Yingru and Liu, Liangqi and Pu, Wenqiang and Liang, Hao and Luo, Zhi-Quan},
title = {Optimistic Thompson Sampling for No-Regret Learning in Unknown Games},
journal = {IEEE Transactions on Signal Processing},
year = {2026},
volume = {74},
pages = {2434--2447},
doi = {10.1109/tsp.2026.3695892}
}
EUVIMED