Multi-Armed Bandits, Conjugate Models and Bayesian Reinforcement Learning | Eigenfoo

[utterances-bot]

Multi-Armed Bandits, Conjugate Models and Bayesian Reinforcement Learning | Eigenfoo

In this blog post I hope to show that there is more to Bayesianism than just MCMC sampling and suffering, by demonstrating a Bayesian approach to a classic reinforcement learning problem: the multi-armed bandit.

https://eigenfoo.xyz/bayesian-bandits/

[eigenfoo]

Just thought I'd share some amazing updates and feedback I've gotten:

• This article made the #2 spot on the Hacker News front page!
• Two data scientists share lessons learnt from two years of designing, building and maintaining Bayesian bandit algorithms: check it out here and here.
• Very unrelated, but a kind stranger on Mastodon explained how manufacturers actually tune the payback percentages of slot machines!

NOTE (01-08-19): I've migrated from Disqus to utterances to provide comments, so I've lost almost all of the comments on my blog posts. I just thought this comment was worth saving, so I've copied and pasted it in.