Optimal Best-arm Identification in Linear Bandits.

Yassir Jedra Alexandre Proutière

Published in: NeurIPS (2020)

Keyphrases

closed form
optimal linear
regret bounds
optimal solution
piecewise linear
dynamic programming
semi infinite programming
worst case
online learning
least squares
optimal design
learning algorithm
multiscale
closed form solutions
automatic identification
asymptotically optimal
case study
real world