Tight First- and Second-Order Regret Bounds for Adversarial Linear Bandits.

Shinji Ito Shuichi Hirahara Tasuku Soma Yuichi Yoshida

Published in: NeurIPS (2020)

Keyphrases

regret bounds
lower bound
upper bound
online learning
linear regression
multi armed bandit
worst case
higher order
bregman divergences
objective function
optimal solution
pairwise
knn
nearest neighbor
least squares
linear predictors