Deep learning new Journey (1)

Deep learning new Journey (1) [Email protected]http://www.cnblogs.com/swje/ Zhouw 2015-11-26

Statement:

1) The Deep Learning Learning Series is a collection of information from the online very big Daniel and the machine learning experts selfless dedication. Please refer to the references for specific information. Specific version statements are also referenced in the original literature.

2) This article is for academic exchange only, non-commercial. So each part of the specific reference does not correspond in detail. If a division accidentally violated the interests of everyone, but also look haihan, and contact bloggers deleted.

3) I Caishuxueqian, finishing summary of the time is inevitable error, but also hope that the predecessors, thank you.

Please contact: [Email protected]

The recent period of time in the deep learning, just contact, is still in the learning stage. Want to record their own learning process, learning to use the URL and notes, etc., so that in the future to review the time can be found at any time, but also hope to be able to share to everyone, the same in the deep learning learning the initial stage of the Friends of a case of a primer (that's Me,right~o (∩_ ∩) o~)

Because the undergraduate major is not the deep learning, even the computer major is not (my undergraduate major is the electronic Institute of Information Engineering, graduate stage of the direction of the computer), just want to graduate design to do a computer-related content, for the postgraduate Stage Lay foundation, early contact with some, So I followed the tutor who did the CV. Also because is the graduation design needs, the postgraduate stage direction also is not the DL, therefore did not have so much time to learn from the machine learning, only wants to get started as soon as possible, can study some interesting things. Let's talk about my study process.

At first I saw a lot of messy information, the instructor recommended a lot of cutting-edge learning materials, but I opened the Web page, do not know where to start, see all English pages, immediately do not know where to start learning. Post your tutor's recommended learning site.

CSC321 Winter 2015: (University of Toronto this year's deep learning courseware material):http://www.cs.toronto.edu/~rgrosse/csc321/calendar.html

UFLDL Tutorial-ufldl (Stanford's deeplearning Introductory tutorial): Http://deeplearning.stanford.edu/wiki/index.php/UFLDL_Tutorial

Deep Learning Concise Tutorial: http://openclassroom.stanford.edu/MainFolder/CoursePage.php?course=DeepLearning

Welcome to Lasagne:http://lasagne.readthedocs.org/en/latest/index.html

github:https://github.com/a very good open source website, unfortunately I still not enough to make full use of, is studying diligently!

--How can a novice use github? http://www.zhihu.com/question/21669554

--git version control software combines GitHub from getting started to mastering common Command learning Manuals: http://www.ihref.com/read-16369.html#1

Later, the tutor let me read a Minerva paper, and asked me to translate, I read two times the thesis, the language basically overcome, but there are many unfamiliar algorithms and professional terminology I do not understand, so I would like to understand the deep learning, familiar with some nouns. Inadvertently, I saw a blog on the introduction of DL, very detailed, but also relatively easy to understand, which is the depth that the beginner can accept. There are many chapters, in layman's sense. Http://blog.csdn.net/zouxy09/article/details/8775360. Intermittent look for a few days, feel good, than directly read some books or video effects better, than to read more vivid, compared to watching video has more time to think.

In understanding some of the concepts, I also refer to some other resources, the following Web site is the interpretation of some nouns and understanding, I think it is better information:

Boosting:

BP (back propagation) Reverse propagation algorithm: http://galaxy.agh.edu.pl/~vlsi/AI/backp_t_en/backprop.html

Gradient diffusion (gradient diffusion):

Come here first, and you'll continue to update later ...

2015-12-01

After two or three days of time, I translated Minerva This article, the following is the link to this article, but also hope that you give a change of opinion. Http://www.cnblogs.com/swje/p/5023773.html. (Summary: Minerva-a scalable and highly efficient Training Platform for deep learning.) This article is a translation of Minerva's introduction. I just contact the depth of learning direction, the professional terminology of very little understanding, ventured to translate this article, there are many languages are not fluent and inaccurate. Hope and the same is the introduction of friends to share this article, but also want to learn from all walks of life the great God feel free! The first time to send a public blog, look forward to your valuable suggestions, hope to work together with June progress! ）

After a few days out to play, back to continue fighting, hey.

2015-12-07

After coming back, there are a few important things to do.

1. Learning mxnet:http://mxnet.readthedocs.org/en/latest/index.html

Https://github.com/dmlc/mxnet

Focus will be on the Python package document (inside the tutorials and Python API documents) and developer documents as well as programming models for the deep Learning the three-piece focus.

2. Learn Python:

The first is a combination of "Python actual combat fourth Edition" that book study;

The second is the combination of Theano science: Https://github.com/fchollet/keras;

Third, through the official website documents: https://www.python.org/doc/

3, Learning neural Network: See CNN, RNN paper

Nn-lectures: University of Toronto courseware CSC321 Winter 2015:introduction to neural Networks

Link: http://pan.baidu.com/s/1gdpuo1t Password: Ueib

4, install Ubuntu dual system (Ubuntu 14.04.1 LTS), and familiar with command operation, learn the use of vim command

5, LSTM:

Lstm Neural network in layman's: http://www.csdn.net/article/2015-06-05/2824880

RNN and lstm Introduction and Formula combing: http://blog.csdn.net/Dark_Scope/article/details/47056361

　　

Well, set goals here, and here's a one to conquer, and I'll update my learning experience, learning notes and learning materials. With June Mutual encouragement!

Deep learning new Journey (1)