Deep Learning (depth learning) Chinese translation

Https://github.com/exacity/deeplearningbook-chinese

In the help of many netizens and proofreading, the draft slowly became a first draft. Although there are still many problems, at least 90% of the content is readable and accurate. We kept the meaning of the original book Deep learning as much as possible and kept the original book's statement.

However, we have limited levels and we cannot eliminate the variance of many readers. We still need your advice and help to reduce the translation bias.

All you have to do is read, then summarize your suggestions and mention issue (preferably not one by one). If you are sure that you do not need to discuss the proposal, you can directly initiate PR.

Corresponding translators: 1th, 4, 7, 10, 14, 20 and 12.4, 12.5 are responsible for the 2nd, 5, 8, 11, 15, 18 chapters by the @liber145 responsible for the 3rd, 6, 9 chapters by the @KevinLee1110 responsible for 13th, 16, 17, Chapters 19 and 12.1 to 12.3 are @futianfan-oriented readers

Please download PDF reading directly.

This version of the accuracy has been improved, readers can be based on the Chinese version of the main, English version supplemented to read the study, but we still suggest that researchers read the original. Publishing and open source reasons

The book will be published by the press, but we are not sure of the exact date. So we can first look at the PDF electronic version, after all, technology changes rapidly.

If you feel that the Chinese version of the PDF is helpful to you, I hope you will be able to support the publication of the paper in the future. If you think the Chinese version is not good, I hope you can make more suggestions. Thank you very much.

Here are the specific reasons for open source: We are not a literary worker, not a full-time translator. Relying on us alone, can not give today's translation, many netizens have given us a valuable advice, so open source helped a lot of busy. Publishers will give us royalties (we do not know how much, maybe 20,000 or so), we are embarrassed to use their own, the discussion after the donation is the most appropriate, in the name of all the users who have contributed. PDF electronic version for the technical books is very important, at any time need to inquire, take the paper version everywhere is obviously not suitable. Many foreign technical books have corresponding electronic version (although not necessarily genuine), and the domestic almost no. Personally think that this is a publishing house or the author thinks that the national quality is not high enough to actively pay for the realm of knowledge, so do not want to "leak" electronic version. Times are progressing and we need to change. In particular, the general quality of translation works is not high in the case, to dare to the world first. Deep learning develops too fast and changes rapidly, so we want you to learn the relevant knowledge earlier. I think the original author of the Open PDF version also has a similar consideration, that is, pay after reading. We think that the quality of Chinese population is high enough to pay for knowledge. Of course, this is not paid to us, it is paid to the publishers, the publishers to pay the original author. We do not want the sales of the Chinese version to fall due to the presence of the PDF version. Publishers only value back to the copyright in the future to introduce more excellent books. Our open source translation precedent will not be a negative case, there will be more PDF version later. Open source also involves copyright issues, for copyright reasons, we no longer update this first edition PDF file, please the final paper version of the subject. Acknowledgements

We have 3 categories of proofreading staff. The person in charge is the corresponding translator. Simple reading, the statement is not fluent or difficult to understand where the proposed changes. In contrast, Chinese and English reading should be carried out to eliminate the situation of double-flipping.

All proofing suggestions are saved in the Annotations.txt file for each chapter.

Chapters	principal	Simple Reading	comparison between China and Britain
Preface to Chapter I.	@swordyork	LC, @SiriusXDJ, @corenel, @NeutronT	@linzhp
Chapter II Linear algebra	@liber145	@SiriusXDJ, @angrymidiao	@badpoem
Chapter III Probability and information theory	@KevinLee1110	@SiriusXDJ	@kkpoker, @Peiyan
Fourth Chapter numerical calculation	@swordyork	@zhangyafeikimi	@hengqujushi
The fifth chapter on machine learning Basics	@liber145	@wheaio, @huangpingchun	@fairmiracle, @linzhp
Sixth chapter Depth Feedforward network	@KevinLee1110	David_chow, @linzhp, @sailordiary
Seventh. Regularization in deep learning	@swordyork
Eighth. Optimization in depth model	@liber145	@happynoom, @codeVerySlow	@huangpingchun
Nineth Convolution Network	@KevinLee1110	@zhaoyu611, @corenel	@zhiding
Tenth series Modeling: loops and recursive networks	@swordyork	Lc	@zhaoyu611, @yinruiqing
The 11th chapter of practical Methodology	@liber145
12th Chapter Application	@swordyork, @futianfan		@corenel
13th Chapter Linear Factor model	@futianfan	@cloudygoose	@ZhiweiYang
14th Chapter Self-encoder	@swordyork		@Seaball, @huangpingchun
The 15th chapter shows that learning	@liber145	@cnscottzheng
16th, structural probabilistic models in deep learning	@futianfan
The 17th Chapter Monte Carlo method	@futianfan		@sailordiary
The 18th chapter faces the distribution function	@liber145
The 19th chapter approximate inference	@futianfan		@sailordiary, @hengqujushi
20th chapter Depth Generation model	@swordyork

We will be in the paper version of the official publication, in the Book of thanks, formally thank you for your contributions to the students.

Many students have put forward a lot of suggestions, we are listed here.

@tttwwy @tankeco @fairmiracle @GageGao @huangpingchun @MaHongP @acgtyrant @yanhuibin315 @Buttonwood @RuiZhang1993 @zymiboxpay @xingkongliang @oisc @tielei @yuduowu @Qingmu -2016 @HC @xiaomingabc of weijy026a @bengordai n @JoyFYan @minoriwww @khty2000 @gump88 @zdx3578 @PassStory @imwebson @wlbksy the @roachsinai @Elvinczp @endymecy Name:yue-daj Iong @9578577 @linzhp @cnscottzheng @germany-zhu @zhangyafeikimi @showgood163 @gump88 @kangqf @NeutronT, @badpoem @kkpoker @Seaball @wheaio @angrymidiao @ZhiweiYang @corenel @zhaoyu611 @SiriusXDJ @dfcv24 emisxxy flyingfire vsooda @friskit-china

If there is any omission, please notify us and email to echo c3dvcmquew9ya0bnbwfpbc5jb20k | Base64-d. This is what we have to thank, so don't be embarrassed. TODO Typesetting note A variety of questions or suggestions can be issue, recommended to use Chinese. Due to copyright issues, we can not upload pictures and bib, please forgive me. Due to copyright issues, we would not upload figures and the bib file. may be used for study purposes and may not be used for any commercial conduct. Thank you. markdown Format

This format is indeed more important, easy to access, and easy to index. After the initial conversion, generate Web pages, see Deeplearningbook-chinese. Note that this conversion does not put the graph in, nor does it put a graph. Currently using a single script, latex-based file conversion may later change but the principle is not to directly modify the MD file. Students who need it can modify their scripts themselves.

Updating .....