3 ms·
I don't believe that is correct, re: UCB1 There are probably a few ways to prove the result now, but one (nicely general) way to prove that online learning res
by idunning 13y ago
I don't believe that is correct, re: UCB1
There are probably a few ways to prove the result now, but one (nicely general) way to prove that online learning result would be the method of multiplicative weights: https://www.cs.princeton.edu/~arora/pubs/MWsurvey.pdf https://www.cs.princeton.edu/~arora/pubs/MWsurvey.pdf