New paper studies offline-to-online learning in linear bandits

Arxiv CS.LG (machine learning)

The paper proposes a method combining offline data and online exploration in stochastic linear bandits. A key finding is a phase transition based on the offline dataset's coverage.

linear-bandits

Arxiv CS.LG (machine learning)·Kushagra ChandakToshinori KitamuraXiaoqi Tan

7 days ago