Deep Pink

[[toc]]
**[[Home]] * [[Engines]] * Deep Pink**

**<span style="color: #ff1493;">Deep Pink</span>**,  <ref>color: #ff1493; [[https://en.wikipedia.org/wiki/Shades_of_pink#Deep_pink|Deep Pink - Shades of pink from Wikipedia]]</ref>
an experimental [[Open Source Engines|open source chess engine]] by [[Erik Bernhardsson]] written in [[Python]] as an attempt to [[Deep Learning|learn]] and play chess. It applies [[Negamax|negamax]] [[Alpha-Beta|alpha-beta]] and a [[Neural Networks#Deep|deep neural network]] as [[Evaluation|evaluation function]], using [[https://en.wikipedia.org/wiki/Theano_%28software%29|Theano]] <ref>[[http://deeplearning.net/software/theano/|Welcome — Theano 0.7 documentation]]</ref>, [[python-chess]], and [[Sunfish]] <ref>[[https://github.com/erikbern/deep-pink|erikbern/deep-pink · GitHub]]</ref> . The input representation seems similar to [[Octavius]] with 12x64 nodes, not feeding in [[Side to move|side to move]], [[Castling rights|castling rights]], and [[En passant|en passant]] target square. The first hidden layer has 2048 neurons as well. Bernhardsson used a [[GPU]] instance to train the net with 100M games for about four days using [[https://en.wikipedia.org/wiki/Stochastic_gradient_descent|stochastic gradient descent]] with [[Mathematician#YNesterov|Nesterov]] <ref>[[https://en.wikipedia.org/wiki/Yurii_Nesterov|Yurii Nesterov from Wikipedia]]</ref> momentum <ref>[[http://erikbern.com/2014/11/29/deep-learning-for-chess/|Deep learning for… chess]] by [[Erik Bernhardsson]], November 29, 2014</ref> <ref>[[https://blogs.princeton.edu/imabandit/2013/04/01/acceleratedgradientdescent/|ORF523: Nesterov’s Accelerated Gradient Descent]] by [[Sébastien Bubeck]], [[https://blogs.princeton.edu/imabandit/|I’m a bandit]], April 1, 2013</ref> <ref>[[https://blogs.princeton.edu/imabandit/2014/03/06/nesterovs-accelerated-gradient-descent-for-smooth-and-strongly-convex-optimization/|Nesterov’s Accelerated Gradient Descent for Smooth and Strongly Convex Optimization]] by [[Sébastien Bubeck]], [[https://blogs.princeton.edu/imabandit/|I’m a bandit]], March 6, 2014</ref>. 

=See also= 
* [[Neural Networks#engines|Chess Engines with Neural Networks]]
* [[Deep Learning]]
* [[Neural MoveMap Heuristic]]
* [[Octavius]]

=Postings= 
* [[http://erikbern.com/2014/11/29/deep-learning-for-chess/|Deep learning for… chess]] by [[Erik Bernhardsson]], November 29, 2014 » [[Deep Learning]]
* [[https://news.ycombinator.com/item?id=8685840|Deep learning for chess Comments]] [[https://en.wikipedia.org/wiki/Hacker_News|Hacker News]]
> [[https://news.ycombinator.com/item?id=8686995|As the author of sunfish....]] by [[Thomas Dybdahl Ahle|Thomas Ahle]], December 2, 2014
> [[https://news.ycombinator.com/item?id=8687273|This has been tried many times before, with better-but-still-lackluster results....]] by halfcat, [[https://en.wikipedia.org/wiki/Hacker_News|Hacker News]], December 2, 2014
* [[http://erikbern.com/2014/12/08/deep-learning-for-chess-addendum/|Deep learning for… chess (addendum)]] by [[Erik Bernhardsson]], December 8, 2014
* [[http://blog.yhat.com/posts/deep-learning-chess.html|ŷhat | Deep Learning for ... Chess]] by [[Erik Bernhardsson]], February 02, 2017
* [[http://www.talkchess.com/forum/viewtopic.php?t=63063|Deep Pink: a chess engine using deep learning]] by [[Chao Ma]], [[CCC]], February 05, 2017

=External Links= 
* [[https://github.com/erikbern/deep-pink|GitHub - erikbern/deep-pink: Deep Pink is a chess AI that learns to play chess using deep learning]]

=References= 
<references />
=What links here?= 
[[include page="Deep Pink" component="backlinks" limit="20"]]
**[[Engines|Up one level]]**