[Computer-go] FYI KL-UCB

Łukasz Lew lukasz.lew at gmail.com
Mon Jul 22 06:04:52 PDT 2013


KL-UCB algorithm
http://arxiv.org/pdf/1102.2490v4.pdf

"Thus, KL-UCB is optimal for Bernoulli distributions and strictly dominates
α-UCB for any
bounded reward distributions."
http://www.princeton.edu/~sbubeck/SurveyBCB12.pdf (page 18)

-- 
Łukasz
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://computer-go.org/pipermail/computer-go/attachments/20130722/b45a49de/attachment.html>


More information about the Computer-go mailing list