One Arrow, Two Kills: An Unified Framework for Achieving Optimal Regret Guarantees in Sleeping Bandits.

Pierre Gaillard Aadirupa Saha Soham Dan

Published in: CoRR (2022)

Keyphrases

unified framework
regret bounds
multi armed bandit
worst case
lower bound
online learning
optimal solution
multi armed bandits
optimal design
probabilistic model
approximation algorithms
asymptotically optimal
special case
feature selection
expert advice
data sets
real time