search

magician-david / aforge

Automatically exported from code.google.com/p/aforge
Other
0 stars 0 forks source link

Add Deep Reinforcement Learning to MachineLearning namespace #403

Open GoogleCodeExporter opened 8 years ago

GoogleCodeExporter commented 8 years ago 
A new set of algorithms are being developed that combine reinforcement learning 
such as Q learning and SARSA with neural networks to provide much more 
sophisticated and adaptable AI.

Here are some papers and implementations:

JavaScript implementation:
http://cs.stanford.edu/people/karpathy/reinforcejs/waterworld.html
https://github.com/karpathy/reinforcejs
http://cs.stanford.edu/people/karpathy/convnetjs/demo/rldemo.html

Google research into Atari game playing AI
http://googleresearch.blogspot.co.uk/2015/02/from-pixels-to-actions-human-level.
html
http://deepmind.com/publications.html

Many thanks.

Original issue reported on code.google.com by mchallis...@googlemail.com on 19 May 2015 at 10:56