建议按序阅读
1.Deep Learning 概述
Deep Learning(深度学习)学习笔记整理系列 : http://blog.csdn.net/zouxy09/article/details/8775360/
2.Neural Networks:
Pattern Recognition And Machine Learning - Chapter 5
PRML勘误表 :
http://research.microsoft.com/en-us/um/people/cmbishop/PRML/pdf/prml-errata-1st-20110921.pdf
3.重要预备知识:
受限玻尔兹曼机RBM:http://blog.csdn.net/itplus/article/details/19168937
参考资料:M-H采样 百度文库 <MetropolisHastings自适应算法及其应用>
4.基于梦境与现实概念的生成和识别权重修正
wake-sleep 算法:The wake-sleep algorithm for unsupervised neural networks
参考资料:
[深度学习] Wake-Sleep算法 :http://www.tuicool.com/articles/3myyau
[深度神经网络](deep neural network)具体是怎样工作的 : http://www.zhihu.com/question/19833708?group_id=15019075
5.从PoE概念引入到CD-k的算法的推导与分析
论文:Training Products of Experts by Minimizing Contrastive Divergence
6.基于DBNs的分类框架
A fast learning algorithm for deep belief nets
参考资料:explaining away & Complementary Prior : http://www.douban.com/note/344374111/
7.基于DBNs的降维框架
Reducing the dimensionality of data with neural networks.
8.一个很好的总结Hinton工作的PPT(也可以在学习DL最开始的时候先过一遍)
http://www.cs.purdue.edu/homes/alanqi/Courses/ML-08/Deep_Belief_nets.pptx
9.其他:
deeplearning code [RBM+BP] :http://www.cs.toronto.edu/~hinton/MatlabForSciencePaper.html
另外一个整理材料 :http://ycool.com/post/g79bn66
机器学习——深度学习(Deep Learning) http://blog.csdn.net/abcjennifer/article/details/7826917
百度文库 <深度学习研究综述>