Author: He Congqing

In recent days, I often see my friends in the wechat group asking me “How to get started with machine learning and what materials to see?” , so I want to introduce how to get started with machine learning and what materials should I read according to the author’s learning experience of more than two years. Here I will organize the resources for getting started with machine learning from the following aspects:

(1) Language: commonly used language in machine learning.

(2) Books: There is a golden room in books. For many mathematical theories involved in machine learning, it is difficult to obtain a complete knowledge framework only by watching videos or blogs.

(3) Video: The derivation of some formulas in the book is difficult to understand, so you can watch the simple and profound courses of Danniu.

(4) Blog: often read some of the great share, for the expansion of knowledge has a certain help.

(5) Competition: Practice is an important criterion for testing learning results. Participating in some algorithm competitions is helpful for understanding algorithms.

(6) Thesis: For some masters, innovation is an important embodiment of testing learning ability.

language

“Life is short, I use Python.” Python has become the dominant language in machine learning due to its rich library of algorithms.

Numpy: one of the most basic Python libraries \

Address: www.numpy.org/

2, Pandas: **** a library commonly used for data processing

Address: pandas.pydata.org/pandas-docs…

3. Scipy: Scipy is an open source Python algorithm library and math toolkit.

Address: docs.scipy.org/doc/scipy/r…

4. Scikit-learn: SkLearn includes many algorithmic interfaces, from supervised learning to semi-supervised learning to unsupervised learning. There are also evaluation indicators, feature selection, etc.

Address: scikit-learn.org/

5, SciKit-multilearn: multi-label algorithm library.

Address: scikit. Ml /

There are also libraries of deep learning algorithms such as:

6. Keras: The best algorithm library for beginners of deep learning.

Address: keras. IO/useful /

There are also libraries of more difficult deep-learning algorithms, such as TensorFlow and PyTorch.

books

1, “Statistical learning methods” : Teacher Li Hang’s “statistical learning methods” this book is a classic, many students rely on this book to find the ideal job, strongly recommended! For many people who want to get started with machine learning, it is recommended to read this book several times to understand every detail of the algorithm.

2. Machine Learning: The book Machine Learning by Zhou Zhihua, also known as watermelon book by many people, is also very helpful. It basically covers all branches of machine learning, such as supervised learning, unsupervised learning, semi-supervised learning, reinforcement learning, feature selection, etc.

3. “Recommendation System Combat” : The book “Recommendation System Combat” by Dr. Xiang Liang is very suitable for those who want to know about the recommendation system.

4. Probability theory and Mathematical Statistics: Many machine learning algorithms are developed from statistical probability theory. For those who lack knowledge of probability and statistics, it is recommended to read this book.

5. Pattern Recognition and Machine Learning: If you are good at English, you can also read THE classic book PRML.

Reinforcement Learning: An Introduction. If you want to study Reinforcement Learning, this book is a great Introduction to Reinforcement Learning.

The PDF version of the above information has been uploaded to the web disk. If you are interested, please pay attention to “The Heart of AI Algorithm” and reply to “Introduction to Machine Learning” in the background.

video

If the above books seem difficult for friends, it is difficult to understand the ins and outs of the algorithm, it is suggested to combine the book (recommended for beginners: statistical Learning Methods) with the video to promote each other.

1. Mr. Ng’s open class: There are his lectures on netease Cloud and Coursera, which are very basic versions. I suggest you watch this video more when you are getting started. Personally, COURsera courses are easier.

Address on netease Cloud: open.163.com/special/ope…

Address on Coursera:

www.coursera.org/learn/machi…

2, Teacher Li Hongyi’s courses: Teacher Li Hongyi’s courses are also relatively good, worth learning.

Here are cleared up version: blog.csdn.net/soulmeetlia…

blog

Domestic:

1, flame flickering: Tencent technology masters blog

Address: www.flickering.cn/

2, Meituan technical team’s blog: there are also a lot of dry goods:

Address: tech.meituan.com/

Su Jianlin’s blog is also full of dry goods

Address: Spaces. Ac. Cn /

4. There are also some large blog websites, such as Blog Garden, Jianshu, CSDN, Zhihu and so on.

Abroad:

1. Netflix: Netflix tech blog, lots of dry stuff.

Address: medium.com/netflix-tec…

2. Towards Data Science: Sharing concepts, ideas and code.

Address: towardsdatascience.com/

3, Github: All code is here.

The game

In the process of learning machine learning, how do you test your learning results? The competition is a good direction, in fact, the competition may be in order to achieve results, the gap between the thousandths, hundredths, but in fact, in the competition thinking is the most important. How to apply these classical algorithms to industry, the advantages and disadvantages of these algorithms in industry? Slowly experience!

Large domestic algorithm platforms include:

Tianchi Big Data:

tianchi.aliyun.com/home/

Datacastle:

www.pkbigdata.com/

Datafountain:

www.datafountain.cn/

Biendata:

biendata.com/

Kesci:

www.kesci.com/

Jdata:

jdata.jd.com/

Large foreign algorithm platforms include:

Kaggle:

www.kaggle.com/

There are many platforms, and these are some of the more famous ones. You can go to the official website to have a look, there are a lot of games going on. In addition, there are many other platforms, which I will not introduce here. In recent days, my friends and I are also thinking about this question. Is it possible to build a website that integrates these contest websites and academic evaluation contests of famous foreign conferences? Welcome to discuss in the comments section!! Give me an opinion by the way!

The paper

A lot of seniors are going to graduate and enter the graduate life, or are they still worried about graduation? Small papers become the old problem of Master’s graduation in China! In fact, writing a relatively simple CCF C class paper is not very difficult, perhaps CCF B CCF A class paper is really difficult! How do you get started? See the top conferences and journal papers on machine learning and artificial intelligence in recent years (conference papers are faster). I just want to sort out the conference papers.

Conference articles worth reading:

1. Data mining:

SIGKDD: Top data Mining paper.

2019: Under review

Accepted Paper 2018

www.kdd.org/kdd2018/acc…

In 2017 accepted paper:

www.kdd.org/kdd2017/acc…

In 2016 accepted paer:

www.kdd.org/kdd2016/pro…

SIGIR: Top recommendation Systems paper

2019 Accepted Paper: Under review

In 2018 accepted paper:

Sigir.org/sigir2018/a…

In 2017 accepted paper:

Sigir.org/chiir2017/a…

In 2016 accepted paper:

Sigir.org/sigir2016/f…

Sigir.org/sigir2016/s…

There are also some top-level meetings: CIKM/ECML-PKDD/ICDM/SDM/WSDM

2. Machine Learning

AAAI: Comprehensive Conference on Top-level Artificial Intelligence

In 2019 accepted paper:

Aaai.org/Conferences…

In 2018 accepted paper:

Aaai.org/Conferences…

In 2017 accepted paper:

www.aaai.org/Conferences…

IJCAI: Comprehensive Conference on Top-level Artificial Intelligence

2019 Accepted Paper: Under review

In 2018 accepted paper:

www.ijcai-18.org/accepted-pa…

In 2017 accepted paper:

Ijcai-17.org/accepted-pa…

ICML: Top machine learning conference \

2019 Accepted Paper: Under review

In 2018 accepted paper:

Icml. Cc/Conferences…

In 2017 accepted paper:

Icml. Cc/Conferences…

NIPS: The Top Comprehensive Artificial Intelligence Conference

2019 AccpeTED Paper: Papers called for

In 2018 accepted paper:

Nips. Cc/Conferences…

In 2017 accepted paper:

Nips. Cc/Conferences…

There are other specialized AI conferences: ACL/EMNLP/NAACL/COLING in natural language processing. Statistics-biased AI conference: AISTATS.

Conference on Artificial Intelligence in Images: CVPR/ICCV/ECCV. Partners can see some of the conference papers related to their own, in view of the shortcomings of the paper method, thinking about the method of improvement!

Author He Congqing’s official account:

Please follow and share ↓↓↓\

Machine learning beginners \

QQ group: 727137612

(Note: there are 7 QQ groups on this site, those who have joined any of them do not need to add more)

ID of knowledge planet of this site: 92416895

Past wonderful review \

  • Conscience Recommendation: Introduction to machine learning and learning recommendations (2018 edition) \

  • Github Image download by Dr. Hoi Kwong (Machine learning and Deep Learning resources)

  • Printable version of Machine learning and Deep learning course notes \

  • Machine Learning Cheat Sheet – understand Machine Learning like reciting TOEFL Vocabulary

  • Introduction to Deep Learning – Python Deep Learning, annotated version of the original code in Chinese and ebook

  • Zotero paper Management tool

  • The mathematical foundations of machine learning

  • Machine learning essential treasure book – “statistical learning methods” python code implementation, ebook and courseware

  • Blood vomiting recommended collection of dissertation typesetting tutorial (complete version)

  • The encyclopaedia of Machine learning introduction – A collection of articles from the “Beginner machine Learning” public account in 2018

  • Installation of Python environment (Anaconda+Jupyter Notebook +Pycharm) \