Problem Dependent View on Structured Thresholding Bandit Problems
James Cheshire 1 Pierre Menard 1 Alexandra Carpentier 1
Abstract of error - i.e. the probability that the learner mis-classifies
We investigate the problem dependent regime at least one arm - and consider therefore the problem de-
in the stochastic Thresholding Bandit problem pendent regime.
(TBP) under several shape constraints. In the The focus of this p ...


雷达卡




京公网安备 11010802022788号







