Towards Minimax Optimal Best Arm Identification in Linear Bandits.

Junwen Yang Vincent Y. F. Tan

Published in: CoRR (2021)

Keyphrases

worst case
optimal linear
closed form
optimal solution
data sets
regret bounds
piecewise linear
semi infinite programming
dynamic programming
special case
linear constraints
automatic identification
linear space
imperfect information
search algorithm
objective function
multi armed bandit
information systems