Thompson sampling, one armed bandits, and the Beta distribution

Thompson sampling is a strategy to explore a space while exploiting the wins. In this video we see an application to winning at a game of one-armed bandits.
Beta distributions video: • The Beta distribution ...
Tom Denton blog: inventingsituations.net/
Icons made by Freepik from www.flaticon.com
Announcement: Book by Luis Serrano! Grokking Machine Learning. bit.ly/grokkingML
40% discount code: serranoyt

[@marcin.sobocinski]

That's the clearest and easiest to understand explanation of Thompson sampling I've ever seen! Thank you!

[@arturoaltamirano1376]

A real talent to explain complex math in simple and clear terms. Excellent use of graphics and animation.

[@aritramandal5855]

Such an informative video. All these complex concepts are cleared so nicely and simply.

[@srinivasanbalan2469]

You are the best teacher always, Dr. Serrano. Thanks for the upload.

[@andrewt929]

“Best” (most intuitive, concise, and practical) explanation of both Thompson sampling and OAB I’ve found. Rivtik should be jealous.

[@lenkapenka6976]

EXCELLENT Video! And one of the few that explain the Bandit problem clearly and succinctly with Thompson sampling.

[@rng1160]

another great video, really appreciate the simplicity and knowledge in it

[@JackLiu-xn4jz]

Because there's no Chinese subtitle in the video so I would like use English to reply the comment。 the video is cool and easy to understand and I understand it and benefit from it lot. Thank you, thank you very much

[@etienneboutet7193]

Thank you for this video. Very informative as always !

[@priyasearcher]

Thank u so much sir... Want more videos on machine learning and deep learning topics

[@abdealiarsiwala5485]

Amazing explaination!

[@cernejr]

Nice video, I learned something today.

[@robmarks6800]

Very well done!

[@cssambit]

Great video ☺️

[@sgodse]

Great video.

[@samvrittiwari351]

Happy Teachers day sir Love from India. 🙂

[@busn311]

8:23 an error in the third graph when adding the third trial. The graph should move skew to the right instead of left.

[@SerranoAcademy]

You’re right, thank you for the correction!

[@ullibowyer]

9:17 you say "picking a random point from the distribution" which makes it seem like the x coordinate is randomly chosen. I think it's much clearer to say draw a random sample from the distribution.

[@yashasvibhatt1951]

Hi Luis, I have a doubt, at 8:46 how did M2, M3 and M4 achieved a Right Skewed, Left Skewed and Right Skewed curves respectively

[@user-wr4yl7tx3w]

How doe Thompson sampling ensure we choose the unexplored machine? Did you make that clear in the video explanation?

[@matheussales4861]

Awesome

[@kasraamanat5453]

❤️

[@tvvt005]

5:52 kindly distinguish between likelihood and probability,i am a little lost here