发帖
雷达卡
On Limited-Memory Subsampling Strategies for Bandits
Dorian Baudry * 1 Yoan Russac * 2 Olivier Cappé 2
Abstract Multi-armed bandits models have been used to address a
There has been a recent surge of interest in non- wide range of sequential optimization tasks under uncer-
parametric bandit algorithms based on subsam- tainty: online recommendation (Li et al., 2011; 2016), strate-
pling. One drawback however of these approaches ...
京公网安备 11010802022788号
