1. 入门
Neural Network(别名):Deep Learning
Deep Learning比较流行的一个原因,主要是因为它能够自主地从数据上学到有用的feature。特别是对于一些不知道如何设计feature的场合,比如说图像和speech。deep learning可以学习到比以往比如说sift或者MFCC这样手工设计的feature更好的方法,而且这些feature有很强的semantic的含义。所以很多时候在用到其他的一些task的时候会很有效,这也是为什么我们可以用一个feature来实现很多比如说识别,检测,物体分割这样的不同task的缘故。
1) Machine Learning Mastery(初学者、从业人员、高级人员)
4) Neural Networks and Deep Learning
5) Neural Networks for Machine Learning (Geoffrey Hinton)
6) Machine Learning: 2014-2015(Nando de Freitas Oxford)
7) CS224d: Deep Learning for Natural Language Processing
9) Deep Learning Summer School, Montreal 2015(Deep Learning,Unsupervised Learning,Reinforcement Learning)
10) Yann LeCun 2015-2016 Deep Learning
11) Deep Learning for Computer Vision Barcelona
12) Computer Vision (CS 543 / ECE 549) – Spring 2015
13) CS3710: Visual Recognition, Spring 2015
13) 16-385 Computer Vision 2015
2. 进阶
- Neural Networks (General)
- Understanding LSTM Networks
- Tensorflow (by Google) tutorials
- Torch (by Facebook)
- Theano (by MILA)
3. Models
4. Computer Vision
5. 强化学习
6. Caffe实战
7. RNN
- recurrent neural nets augmented with differentiable memory modules (e.g. Memory Networks:
- Memory Networks (FAIR): https://scholar.google.co
m/citat... - Stack-Augmented RNN (FAIR): Google Scholar Citations
- Neural Turing Machine (DeepMind): [1410.5401] Neural Turing Machines
- End-toEnd MemNN (FAIR/NYU): https://scholar.google.co
m/citat...
- Memory Networks (FAIR): https://scholar.google.co
8. Hadoop和Spark DL工具
9. 他人博客
12. 图片检测
13. 对抗机器学习 (Adversarial Machine Learning)
1) Ian Goodfellow开源对抗机器学习库cleverhans v0.1
14. Human Pose Estimation
1) 2D : MPII Human Pose Models
参考:
1) http://ofir.io/How-to-Start-Learning-Deep-Learning/
2)