In recent years, artificial intelligence and machine learning technologies have repeatedly refreshed people’s cognition of “what computers can do”, and made breakthroughs in application fields such as man-machine game, computer vision, biometric recognition, unmanned driving and medical diagnosis. Machine learning is so popular that as a programmer or general practitioner, is it necessary to transition to machine learning?

It is necessary. But I’m not advocating blindly quitting your job to do deep learning, or artificial intelligence. With the popularity of big data, machine learning, especially deep learning, has become the advanced direction of many engineers. But any technology should be based on actual business needs. For example, we’re not going to be the front-end developer in machine learning, we’re going to be the front-end developer in machine learning, and that’s your advantage.


Machine learning started to catch fire, which pleased me as a professional, but also worried the people outside machine learning, because they were following the path from my 10-year undergraduate to doctor’s degree, and they learned by themselves in their spare time, which was difficult to imagine. I thought about it and gave him the following two tips:

1. Don’t let hundreds of gigabytes of data get in the way

When we are getting started, we usually collect massive learning materials, such as “100 PDF books required for machine learning”, “internal data of XXX College”, “from the beginning to the actual combat * 100 G data”, and then put them into the network disk step by step. But 90 percent of people look at the data and say it’s too much and they don’t know where to start. So a really good primer is more important than collecting thousands of gigabytes of information.

2. Give up 0 basic entry, curve breakthrough

Machine learning is a complex technology that combines probability theory, linear algebra, convex optimization, computer and neuroscience. But regardless of the academic needs, most people will not end up working in algorithms, but in front-line applications. Mathematical theory learning is not so necessary at the beginning of the entry, it is best to have a systematic understanding from the top framework, and then from practice to theory, targeted inspection of machine learning knowledge points. From macro to micro, from the whole to the details, more conducive to machine learning fast start! And in terms of learning enthusiasm, it also plays a “positive feedback” role.


My name is Beck Wang, (PhD) graduated from Tsinghua University, majoring in data and artificial intelligence. Now I am a lecturer in a key university in Beijing. I have published many papers in TKDE/ TKDD and IJCAI/AAAI. He once worked in Microsoft Research Asia and a Wall Street fund company. He has rich experience in machine learning in the field of artificial intelligence.


This course a total of 10 section 41, systematically shows the machine learning to master the theory knowledge, in order to everyone can understand the point of view of combining specific example python, enables learners to use machine learning algorithms to solve specific practical problems, and ultimately to master basic theory of machine learning in practice, a simple, reverse thinking breakthrough in machine learning 0 to 1.

1\ Simple, high school graduates will be able to use the mathematical foundation

As a basic course, it is probably the easiest machine learning course to get started in the whole web. There is no complex formula derivation and theoretical analysis in the course. Of course, you’d better have some knowledge of “matrix-vector” multiplication in common with programmers.

2. Learn python, the leading machine learning development language

As an interpreted language, Python is inherently superior to other languages in terms of human interaction, and it has a number of open source frameworks. This course will help you improve your Python programming skills, using algorithms and tuning models flexibly.

3\ Theory combined with practice, 14 application cases, 23 programming examples

In order to make the technology more closely with the practical application, this course uses a case-driven approach to explain the basic, practical and advanced methods and skills of machine learning, covering the theory of machine learning algorithm, model tuning and solution.

If you want to get the video and courseware of this course for free, you can contact the teaching assistant weixin: BT474849