Slacker News Slacker News logo featuring a lazy sloth with a folded newspaper hat
  • top
  • new
  • show
  • ask
  • jobs
Library
← Back to context

Comment by conductrics

13 years ago

You might want to look at botlzman/softmax if you want to weight the prob of selection as a function of the current estimated value. One tricky bit is figuring out a good setting for the temperature parameter. Another poster alluded to softmax. In my experience it dosn't really perform better than a simple e-greedy approach, but maybe it has worked well for others?

0 comments

conductrics

Reply

No comments yet

Contribute on Hacker News ↗

Slacker News

Product

  • API Reference
  • Hacker News RSS
  • Source on GitHub

Community

  • Support Ukraine
  • Equal Justice Initiative
  • GiveWell Charities