🔒 Locked

Q-Learning (Machine Learning)

Started by firefly2442 Jun 5, 2006 at 5:19 PM 2 replies 1.3k views

Original Post

Jun 5, 2006 at 5:19 PM

Hello. I'm trying to learn the Q-Learning Machine Learning Algorithm and I'm having a hard time. I have various papers on the topic but all of them start out OK but then delve into math and equations which doesn't really help me. Does anyone know of any explanations or tutorials online that will explain the topic without using so much math and symbols? I've looked on google but haven't found much. Thanks in advance. :)

Timkin

149

Timkin

Jun 5, 2006 at 9:56 PM

Go directly to the source (and still one of the best introductions to this material) and get a hold of Rich Sutton and Andrew Barto's book, "Reinforcement Learning: An Introduction".

Cheers,

Timkin

XXX_Andrew_XXX

Jun 5, 2006 at 10:17 PM

I second the sutton and barto book. You can even look at a copy online:

http://www.cs.ualberta.ca/%7Esutton/book/ebook/the-book.html

firefly2442

Author

firefly2442

Jun 12, 2006 at 4:55 PM

Thanks for the help. I think I understand the algorithm for the most part although I have one question:

Given the Q-learning equation:

Qnow = Qnow + alpha(reward + gamma * max(Qnext) - Qnow)

I understand that alpha is a variable that determines how effective that learning is given that point in time, but I'm still a little lost on the gamma variable. Thanks. :)

Topic Locked

This topic has been locked by a moderator. New replies are not allowed.