Taming the Monster: A Fast and Simple Algorithm for Contextual Bandits
8 950
25.5
Microsoft Research328 тыс
Следующее
08.07.16 – 1331:37:10Популярные
27.10.22 – 22850:10
27.10.22 – 36839:24Опубликовано 8 июля 2016, 0:11
IMS-Microsoft Research Workshop: Foundations of Data Science - Taming the Monster: A Fast and Simple Algorithm for Contextual Bandits
We present a new algorithm for the contextual bandit learning problem, where the learner repeatedly takes one of K actions in response to the observed context, and observes the reward only for that chosen action. Our method assumes access to an oracle for solving fully supervised cost-sensitive classification problems and achieves the statistically optimal regret guarantee with only ˜O(√KT) oracle calls across all T rounds. By doing so, we obtain the most practical contextual bandit learning algorithm amongst approaches that work for general policy classes. We further conduct a proof-of-concept experiment which demonstrates the excellent computational and prediction performance of (an online variant of) our algorithm relative to several baselines. [Joint work with Daniel Hsu, Satyen Kale, John Langford, Lihong Li and Rob Schapire]
We present a new algorithm for the contextual bandit learning problem, where the learner repeatedly takes one of K actions in response to the observed context, and observes the reward only for that chosen action. Our method assumes access to an oracle for solving fully supervised cost-sensitive classification problems and achieves the statistically optimal regret guarantee with only ˜O(√KT) oracle calls across all T rounds. By doing so, we obtain the most practical contextual bandit learning algorithm amongst approaches that work for general policy classes. We further conduct a proof-of-concept experiment which demonstrates the excellent computational and prediction performance of (an online variant of) our algorithm relative to several baselines. [Joint work with Daniel Hsu, Satyen Kale, John Langford, Lihong Li and Rob Schapire]
Свежие видео
2 дня – 22 8809:14
3 дня – 1 816 84714:15
3 дня – 47748:11
10 дней – 1 940 4129:41
11 дней – 3 353 9403:22Случайные видео
143 дня – 35 4680:41
185 дней – 4 714 7869:19
242 дня – 40 7700:34
272 дня – 5 624 3430:39
08.01.20 – 133 2936:09
01.09.11 – 59 56911:12
3 дня – 6090:57
4 дня – 76 8730:16
8 дней – 4 3770:17
9 дней – 203 3900:16
1 день – 3 2671:30