MetaGameBO: Hierarchical Game-Theoretic Driven Robust Meta-Learning for Bayesian Optimization
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (43) · 36705-36713
DOI: 10.1609/aaai.v40i43.40995 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Hui Li und 7 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref).
Abstract (Originalsprache)
Meta-learning for Bayesian optimization accelerates optimization by leveraging knowledge from previous tasks, but existing methods optimize for average performance and fail on challenging outlier tasks critical in practice. These limitations become particularly severe when target tasks exhibit distribution shifts or when optimization budgets are limited in real-world applications. We introduce MetaGameBO, a hierarchical game-theoretic framework that formulates meta-learning as robust optimization through CVaR-based task selection and diversity-aware sample learning. Our approach incorporates uncertainty-aware adaptation via probabilistic embeddings and Thompson sampling for robust generalization to out-of-distribution targets. We establish theoretical guarantees including convergence to game-theoretic equilibria and improved sample complexity, and demonstrate substantial improvements with 95.7% reduction in average loss and 88.6% lower tail risk compared to state-of-the-art methods on challenging tasks and distribution shifts.
Zitieren
APA
Li, H., Liu, H., Fu, Y., Lin, S., Zhang, B., Ouyang, D., Jing, L., & Yu, J. (2026). MetaGameBO: Hierarchical Game-Theoretic Driven Robust Meta-Learning for Bayesian Optimization. Proceedings of the AAAI Conference on Artificial Intelligence, 40(43), 36705-36713. https://doi.org/10.1609/aaai.v40i43.40995
Vancouver
Li H, Liu H, Fu Y, Lin S, Zhang B, Ouyang D, et al. MetaGameBO: Hierarchical Game-Theoretic Driven Robust Meta-Learning for Bayesian Optimization. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(43):36705-36713. doi:10.1609/aaai.v40i43.40995
BibTeX
@article{Li2026,
author = {Li, Hui and Liu, Huafeng and Fu, Yiran and Lin, Shuyang and Zhang, Baoxin and Ouyang, Deqiang and Jing, Liping and Yu, Jian},
title = {MetaGameBO: Hierarchical Game-Theoretic Driven Robust Meta-Learning for Bayesian Optimization},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {43},
pages = {36705--36713},
doi = {10.1609/aaai.v40i43.40995}
}
EUVIMED