深度学习Deep Learning资料大全

Posted 2020-09-19 Aaron_L

转载：http://www.cnblogs.com/charlotte77/p/5485438.html

 

　　最*在学深度学*相关的东西，在网上搜集到了一些不错的资料，现在汇总一下：

Free Online Books

1. Deep Learning66 by Yoshua Bengio, Ian Goodfellow and Aaron Courville
2. Neural Networks and Deep Learning42 by Michael Nielsen
3. Deep Learning27 by Microsoft Research
4. Deep Learning Tutorial23 by LISA lab, University of Montreal
5. Deep Learning:An MIT Press Book

Courses

1. Machine Learning10 by Andrew Ng in Coursera
2. Neural Networks for Machine Learning12 by Geoffrey Hinton in Coursera
3. Neural networks class2 by Hugo Larochelle from Université de Sherbrooke
4. Deep Learning Course14 by CILVR lab @ NYU

Video and Lectures

1. How To Create A Mind3 By Ray Kurzweil - Is a inspiring talk
2. Deep Learning, Self-Taught Learning and Unsupervised Feature Learning2 By Andrew Ng
3. Recent Developments in Deep Learning2 By Geoff Hinton
4. The Unreasonable Effectiveness of Deep Learning by Yann LeCun
5. Deep Learning of Representations by Yoshua bengio
6. Principles of Hierarchical Temporal Memory by Jeff Hawkins
7. Machine Learning Discussion Group - Deep Learning w/ Stanford AI Lab1 by Adam Coates
8. Making Sense of the World with Deep Learning1 By Adam Coates
9. Demystifying Unsupervised Feature Learning By Adam Coates
10. Visual Perception with Deep Learning3 By Yann LeCun
11. Oxford Deep Learning -Nando de Freitas ：在Oxford开设的深度学*课程，有全套视频

Papers

1. ImageNet Classification with Deep Convolutional Neural Networks5
2. Using Very Deep Autoencoders for Content Based Image Retrieval2
3. Learning Deep Architectures for AI2
4. CMU’s list of papers7
5. The Learning Machines - 一个导论性质的文章，让你大致了解深度学*是什么，用来干什么的。
6. Deep Learning - (Review Article in Nature, May 2015) 三大神 Yann LeCun, Yoshua Bengio, and Geoffrey Hinton的文章，不解释。
7. Growing Pains in Deep Learning
8. Deep Learning in Neural Networks - This technical report provides an overview of deep learning and related techniques with a special focus on developments in recent years. 主要看点是深度学**两年（2012-2014）的进展情况。

Tutorials

1. UFLDL Tutorial 120
2. Deep Learning Tutorial from Stanford:斯坦福的官方Tutorial
3. Deep Learning for NLP (without Magic)8
4. A Deep Learning Tutorial: From Perceptrons to Deep Networks5

WebSites

1. deeplearning.net7
2. deeplearning.stanford.edu6
3. Forum2

Datasets

1. MNIST1 Handwritten digits
2. Google House Numbers from street view
3. CIFAR-10 and CIFAR-10034.  IMAGENET1
4. Tiny Images1 80 Million tiny images6.  Flickr Data 100 Million Yahoo dataset
5. Berkeley Segmentation Dataset 500

Frameworks

1. Caffe92.  Torch73
2. Theano3
3. cuda-convnet25.  Ccv1
4. NuPIC3
5. DeepLearning4J：Java和Scala写的，能在Hadoop和Spark上应用，功能非常强大

Miscellaneous

1. Google Plus - Deep Learning Community
2. Caffe Webinar4
3. 100 Best Github Resources in Github for DL5
4. Word2Vec3
5. Caffe DockerFile2
6. TorontoDeepLEarning convnet
7. Vision data sets1
8. Fantastic Torch Tutorial4 My personal favourite. Also check out gfx.js1

Github

1. DeepLearn Toolbox
2. Caffe Webinar4
3. 100 Best Github Resources in Github for DL5
4. Word2Vec3
5. GitHub - Eniac-Xie/PyConvNet: Convolutional Neural Network for python users  ：一个简单的CNN实现（Python）

几个常见应用领域

• Video Recognition - finding and/or identifying specific items in videos or images
• Self-Driving Cars - just like it says, cars that drive without humans
• Natural Language Processing - getting computers to understand human vocal languages
• And others - A free book chapter on many applications of deep learning

几个常用的深度学*代码库

• H2O - 一个开源的可扩展的库，支持Java, Python, Scala, and R

• Deeplearning4j - Java库，整合了Hadoop和Spark

• Caffe - Yangqing Jia读研究生的时候开发的，现在还是由Berkeley维护。

• Theano - 最流行的Python库

News

• Deep Learning News - 紧跟深度学*的新闻、研究进展和相关的创业项目。

 

CV和NLP方面的应用（左边的链接是论文，右边的是代码）

• Page on Toronto, Home Page of Geoffrey Hinton
• Page on Toronto, Home Page of Ruslan R Salakhutdinov
• Page on Wustl, ynd/cae.py · GitHub
• Page on Icml, https://github.com/lisa-lab/pyle...
• Page on Jmlr, pylearn2)
• On the difficulty of training recurrent neural networks, trainingRNNs
• ImageNet Classification with Deep Convolutional Neural Networks, cuda-convnet - High-performance C++/CUDA implementation of convolutional neural networks - Google Project Hosting
• Linguistic Regularities in Continuous Space Word Representations, word2vec - Tool for computing continuous distributed representations of words. - Google Project Hosting

 

最后一定得推荐这个Github：

 

机器学*(Machine Learning)&深度学*(Deep Learning)资料(Chapter 2)（篇目一是机器学*的资料汇总，篇目二是深度学*的汇总，并且在不断更新中）

 

• 《Image Scaling using Deep Convolutional Neural Networks》

介绍:使用卷积神经网络的图像缩放.

• 《Proceedings of The 32nd International Conference on Machine Learning》

介绍:ICML2015 论文集,优化4个+稀疏优化1个；强化学*4个，深度学*3个+深度学*计算1个；贝叶斯非参、高斯过程和学*理论3个；还有计算广告和社会选择.ICML2015 Sessions.

• 《Image Scaling using Deep Convolutional Neural Networks》

介绍:使用卷积神经网络的图像缩放.

• 《Microsoft researchers accelerate computer vision accuracy and improve 3D scanning models》

介绍:，第28届IEEE计算机视觉与模式识别(CVPR)大会在美国波士顿举行。微软研究员们在大会上展示了比以往更快更准的计算机视觉图像分类新模型，并介绍了如何使用Kinect等传感器实现在动态或低光环境的快速大规模3D扫描技术.

• 《Machine Learning for Humans》

介绍:(文本)机器学*可视化分析工具.

• 《A Plethora of Tools for Machine Learning》

介绍:机器学*工具包/库的综述/比较.

• 《The art of visualizing visualizations: a best practice guide》

介绍:数据可视化最佳实践指南.

• 《MIT Machine Learning for Big Data and Text Processing Class Notes - Day 1》

介绍:Day 1、Day 2、Day 3、Day 4、Day 5.

• 《Getting “deep” about “deep learning”》

介绍:深度学*之“深”——DNN的隐喻分析.

• 《Mixture Density Networks》

介绍:混合密度网络.

• 《Interview Questions for Data Scientist Positions》

介绍:数据科学家职位面试题.

• 《Accurately Measuring Model Prediction Error》

介绍:准确评估模型预测误差.

• 《Continually updated Data Science Python Notebooks》

介绍:Continually updated Data Science Python Notebooks.

• 《How to share data with a statistician》

介绍:How to share data with a statistician.

• 《The Eyescream Project NeuralNets dreaming natural images》

介绍:来自Facebook的图像自动生成.

• 《How to share data with a statistician》

介绍:How to share data with a statistician.

• 《A Neural Conversational Model》

介绍:(Google)神经(感知)会话模型.

• 《The 50 Best Masters in Data Science》

介绍:The 50 Best Masters in Data Science.

• 《NLP常用信息资源》

介绍:NLP常用信息资源.

• 《Conditional Random Fields as Recurrent Neural Networks》

介绍:语义图像分割的实况演示,通过深度学*技术和概率图模型的语义图像分割.

• 《Fully Convolutional Networks for Semantic Segmentation》

介绍:Caffe模型/代码:面向图像语义分割的全卷积网络,模型代码.

• 《Growing Pains for Deep Learning》

介绍:深度学*——成长的烦恼.

• 《Clustering Text Data Streams – A Tree based Approach with Ternary Function and Ternary Feature Vector 》

介绍:基于三元树方法的文本流聚类.

• 《Foundations and Advances in Data Mining》

介绍:Free Ebook:数据挖掘基础及最新进展.

• 《The Deep Learning Revolution: Rethinking Machine Learning Pipelines》

介绍:深度学*革命.

• 《The Definitive Guide to Do Data Science for Good》

介绍:数据科学(实践)权威指南.

• 《Microsoft Academic Graph》

介绍:37G的微软学术图谱数据集.

• 《Challenges and Opportunities Of Machine Learning In Production》

介绍:生产环境(产品级)机器学*的机遇与挑战.

• 《Neural Nets for Newbies》

介绍:神经网络入门.

• 《A Nearly-Linear Time Framework for Graph-Structured Sparsity》

介绍:来自麻省理工的结构化稀疏论文.

• 《Optimal and Adaptive Algorithms for Online Boosting》

介绍:来自雅虎的机器学*小组关于在线Boosting的论文 .

• 《Top 20 Python Machine Learning Open Source Projects》

介绍:20个最热门的开源(Python)机器学*项目.

• 《The Parallel C++ Statistical Library for Bayesian Inference: QUESO》

介绍:C++并行贝叶斯推理统计库QUESO,github code.

• 《《Deep learning》Yann LeCun, Yoshua Bengio, Geoffrey Hinton (2015) 》

介绍:Nature:LeCun/Bengio/Hinton的最新文章《深度学*》,Jürgen Schmidhuber的最新评论文章《Critique of Paper by "Deep Learning Conspiracy" (Nature 521 p 436)》.

• 《Palladium》

介绍:基于Scikit-Learn的预测分析服务框架Palladium.

• 《Advances in Structured Prediction》

介绍:John Langford和Hal Daume III在ICML2015上关于Learning to Search的教学讲座幻灯片.

• 《100 open source Big Data architecture papers for data professionals》

介绍:读完这100篇论文 就能成大数据高手,国内翻译.

• 《Social Media & Text Analytics》

介绍:NLP课程《社交媒体与文本分析》精选阅读列表.

• 《Machine Learning for Developers》

介绍:写给开发者的机器学*指南.

• 《Hot news detection using Wikipedia》

介绍:基于维基百科的热点新闻发现.

• 《Harvard Intelligent Probabilistic Systems Group》

介绍:(Harvard)HIPS将发布可扩展/自动调参贝叶斯推理神经网络.

• 《An Empirical Exploration of Recurrent Network Architectures》

介绍:面向上下文感知查询建议的层次递归编解码器.

• 《Efficient Training of LDA on a GPU by Mean-for-Mode Estimation》

介绍:GPU上基于Mean-for-Mode估计的高效LDA训练.

• 《From the Lab to the Factory: Building a Production Machine Learning Infrastructure》

介绍:从实验室到工厂——构建机器学*生产架构.

• 《6 Useful Databases to Dig for Data (and 100 more)》

介绍:适合做数据挖掘的6个经典数据集(及另外100个列表).

• 《Deep Networks for Computer Vision at Google – ILSVRC2014》

介绍:Google面向机器视觉的深度学*.

• 《How to choose a machine learning API to build predictive apps》

介绍:构建预测类应用时如何选择机器学*API.

• 《Exploring the shapes of stories using Python and sentiment APIs》

介绍:Python+情感分析API实现故事情节(曲线)分析.

• 《Movie selection using R》

介绍:(R)基于Twitter/情感分析的口碑电影推荐,此外推荐分类算法的实证比较分析.

• 《A Tutorial on Graph-based Semi-Supervised Learning Algorithms for NLP》

介绍:CMU(ACL 2012)(500+页)面向NLP基于图的半监督学*算法.

• 《Arbitrariness of peer review: A Bayesian analysis of the NIPS experiment》

介绍:从贝叶斯分析NIPS，看同行评审的意义.

• 《Basics of Computational Reinforcement Learning》

介绍:(RLDM 2015)计算强化学*入

以上是关于深度学习Deep Learning资料大全的主要内容，如果未能解决你的问题，请参考以下文章

机器学习(Machine Learning)&深度学习(Deep Learning)资料

机器学习(Machine Learning)与深度学习(Deep Learning)资料汇总

机器学习(Machine Learning)与深度学习(Deep Learning)资料汇总

动手学深度学习 DIVE INTO DEEP LEARNING -- Day 01 -- 学习资料环境配置

深度学习FPGA实现基础知识6(Deep Learning（深度学习）学习资料大全及CSDN大牛博客推荐)

Deep Learning（深度学习）学习笔记整理系列之