analiza aiqbench
This research addresses the practical challenge of leveraging offline datasets to improve early-stage performance in stochastic linear bandits without sacrificing long-term exploration. The proposed algorithm achieves sublinear regret by dynamically balancing reliance on historical data with active exploration as the interaction horizon expands. For practitioners, this offers a theoretically grounded method to optimize decision-making in environments where both prior data and real-time feedback are available.
Offline-to-Online Learning in Linear Bandits
arXiv:2606.04305v1 Announce Type: new Abstract: We study online learning with an additional offline dataset in the stochastic linear bandit setting. Although this problem arises frequently in practice, the offline-to-online tradeoff remains poorly understood in structured environments. We propose a linear bandit algorithm that balances this tradeoff: it relies on offline data during early rounds, and increasingly favors exploration as the horizon grows. We establish regret bounds showing that our method is simultaneously competitive with both purely online and purely offline solutions. In particular, it achieves sublinear regret relative to the optimal action in the number of online interactions, while its regret relative to an offline reference decreases as the number of offline samples grows. Empirical results further demonstrate its effectiveness across various problem parameters.
Abstract:We study online learning with an additional offline dataset in the stochastic linear bandit setting. Although this problem arises frequently in practice, the offline-to-online tradeoff remains poorly understood in structured environments. We propose a linear bandit algorithm that balances this tradeoff: it relies on offline data during early rounds, and increasingly favors exploration as the horizon grows. We establish regret bounds showing that our method is simultaneously competitive with both purely online and purely offline solutions. In particular, it achieves sublinear regret relative to the optimal action in the number of online interactions, while its regret relative to an offline reference decreases as the number of offline samples grows. Empirical results further demonstrate its effectiveness across various problem parameters.
Machine Learning (cs.LG); Machine Learning (stat.ML)
arXiv:2606.04305 [cs.LG]
(or arXiv:2606.04305v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2606.04305 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 216 cuvinte
Subiecte conexe
Sequential Batch Learning in Finite-Action Linear Contextual Bandits
arXiv · 19 aug. 2026
Efficient Adversarial Attacks on High-dimensional Offline Bandits
arXiv · 5 iun. 2026
Gaussian-Width Bounds for Offline-to-Online Linear Bandits
arXiv · 12 oct. 2026
Near-Optimal Regret Bounds for Monotone Stochastic Optimization under Semi-Bandit Feedback
arXiv · 14 oct. 2026