@ 马辰The_answer：主页君能否推荐一些用深度学习做推荐的文章资料，之前看过ICML2007的Restricted Boltzmann Machines for Collaborative Filtering，感觉dl在推荐上的应用不是太多，主页能不能再推荐一些资料，谢过~

[haoawesome]

私信:

@马辰The_answer：主页君能否推荐一些用深度学习做推荐的文章资料，之前看过ICML2007的Restricted Boltzmann Machines for Collaborative Filtering，感觉dl在推荐上的应用不是太多，主页能不能再推荐一些资料，谢过~ ~

http://www.weibo.com/5220650532/BiQUcxKod?ref=

[haoawesome]

keywords: recommender systems deep learning Restricted Boltzmann Machines (RBMs) collaborative filtering

readings

http://benanne.github.io/2014/08/05/spotify-cnns.html Recommending music on Spotify with deep learning

http://www.businessinsider.com/netflix-using-ai-to-suggest-better-films-2014-2 Netflix Is 'Training' Its Recommendation System By Using Amazon's Cloud To Mimic The Human Brain (2014)

Read more: http://www.businessinsider.com/netflix-using-ai-to-suggest-better-films-2014-2#ixzz3Aixvps00

http://nips.cc/Conferences/2013/Program/event.php?ID=4028 Deep content-based music recommendation (nips 2013 poster)

http://classes.soe.ucsc.edu/cmps290c/Spring13/proj/kailu_report.pdf The application of Deep Learning in Collaborative Filtering it is a student project, but reasonably good survey.

related

http://www.slideshare.net/kerveros99/learning-to-rank-for-recommender-system-tutorial-acm-recsys-2013 recommender system tutorial, mentioned deep learning at the end (2013)

http://ijcai13.org/files/tutorial_slides/td3.pdf Tutorial: Recommender Systems (IJCAI 2013)

[haoawesome]

问: @马辰The_answer：主页君能否推荐一些用深度学习做推荐的文章资料 答: 问答进展 http://memect.co/iMHKPqR 找到一篇今年8月新鲜出炉, Spotify的实习生博文: Recommending music on Spotify with deep learning .今年2月有人讲Netflix的电影推荐. 欢迎补充指正

http://www.weibo.com/5220650532/BiQUcxKod?ref=

[haoawesome]

@赵家平USC : Deep Learning除了在IT中有广泛的应用， 还能用在哪些领域呢？ 一年前Sebastian Seung 用CNN重建了老鼠视网膜里的plexiform layer；昨天UCI的学者在nature上撰文说DL可用于发现 希格斯玻色子 等exotic particles, 比传统的方法更精确http://t.cn/Rv3g1EL 看来DL在高能物理学，神经学，医学中都有应用 http://www.weibo.com/2288385870/BbPMYrazJ?mod=weibotime

@赵家平USC : Jeff Hinton组把deep CNN(CovNets)在ImageNet上train好的模型放到网上了，试了下classification, retrieval, image2text的在线demo, amazing! http://t.cn/Rvs0Pvj 最重要的是他们的source code以及installation & documentation 也一并公布，超过Rob Fergus学生的Clarifai http://t.cn/8kL993u http://www.weibo.com/2288385870/Bdcr2jitr?mod=weibotime

@赵家平USC : Fei-Fei组的 video classification with CNN, 结果并没达到Cordelia组的最好结果，从文中可知temporal-spatial CNN没有很好的学到Motion pattern. 不过看看Zisserman组还没publish文章：Two-stream convolutional networks for action recognition in videos，精度与crafted-feature recognition接近 http://www.weibo.com/2288385870/BfkIslHXi?mod=weibotime