Big data digest editing group

Google has just brought a big gift to machine learners around the world.

As the first step to the field of artificial intelligence education, Google's artificial intelligence learning website learn with Google AI is online today, and has launched a machine learning crash course (MLCC), which provides interactive teaching videos and exercises and teaches machine learning concepts free of charge.

Course website:

https://developers.google.cn/machine-learning/crash-course/

In this release, Google provides five versions of English, Spanish, French, Korean and Mandarin to break the language barrier and benefit the global AI learners.

As a machine learning course, this textbook also has a lot of AI elements. The whole set of Chinese audio and Chinese ppt are generated by machine learning technology, and all of them are embedded in the course video, which is quite painstaking.

The course is based on a two-day internal training course designed to help Google engineers learn machine learning concepts. According to reports, more than 10000 Google engineers have studied the same course before.

Google said that through the training of the course, Google engineers can have more innovations in the field of AI: for example, strengthening camera calibration of daydream devices, creating virtual reality for Google Earth, and improving the quality of YouTube's streaming media. "The success of MLCC at Google has inspired us to open these training courses to all."

MLCC is also the first of many courses and resources that Google intends to provide through the new ai.google website. More courses and resources are coming.

Google hopes that the website will become a repository for machine learning and artificial intelligence, and a "base camp" for all machine learners, attracting all levels of AI enthusiasts from senior researchers to beginners.

Machine learning crash course is about 15 hours long, including interactive courses, lectures by Google researchers and more than 40 exercises.

Abstract bacteria also made a simple evaluation of this course. Generally speaking, this course is a very basic introduction to the concept of artificial intelligence. It is suitable for new students who have a certain Python foundation but have no machine learning experience.

So for beginners, how to choose the latest courses?

Compared with the deep learning.ai series courses newly launched by Wu Enda, Google is obviously more inclined to the introduction of machine learning, and learners with programming foundation can quickly understand the relevant concepts of machine learning and get started through this course, which is undoubtedly good news for most programmers.

For in-depth learning in computer vision, natural language processing and other directions, this course is not involved. At the same time, you can only learn from Google's own TensorFlow framework. But if the goal is to get started, it's enough to learn.

Last year at the I / O Developers Conference, Google announced new educational programs on its website. In recent years, Google has repeatedly said that its goal is to democratize artificial intelligence and provide tools for everyone. This course is clearly an important step.

Let's see how Google officials describe what the course aims to solve:

• How is machine learning different from traditional programming?

How is machine learning different from traditional programming?

• What is loss and how to measure it?

What is loss and how to measure it?

• How does gradient descent work?

How does gradient descent work?

• How can I determine if my model is valid?

How can I determine if my model is valid?

• How to provide my data for machine learning?

How to provide my data for machine learning?

• How to build deep neural network?

How to build deep neural network?

Google recommends that learners master introductory algebra, basic programming knowledge and python.

The official website has the following requirements for course learners:

• Master basic algebra knowledge. You should be familiar with variables and coefficients, linear equations, function graphs and histograms (familiarity with higher-level mathematical concepts such as logarithm and derivative will help, but is not necessary);

Master basic algebra knowledge. You should be familiar with variables and coefficients, linear equations, function graphs and histograms (familiarity with higher-level mathematical concepts such as logarithm and derivative will help, but is not necessary);

• Master the basic knowledge of programming, and have some experience of programming in Python.

Master the basic knowledge of programming, and have some experience of programming in Python.

• The programming exercises in the machine learning crash course are coded using tensorflow and python. You don't need to have experience with tensorflow, but you should be able to read and write Python code that includes basic programming structures such as function definitions / calls, lists and dictionaries, loops, and conditional expressions.

The programming exercises in the machine learning crash course are coded using tensorflow and python. You don't need to have experience with tensorflow, but you should be able to read and write Python code that includes basic programming structures such as function definitions / calls, lists and dictionaries, loops, and conditional expressions.

Take a look at the course catalog:

Machine learning concept:

Introduction to machine learning (3 minutes)

Frame handling (15 minutes)

Learn more about machine learning (20 minutes)

Reduce losses (60 minutes)

Basic steps to use TF (60 minutes)

Generalization (15 minutes)

Training and test sets (25 minutes)

Validation (40 minutes)

Representation (65 minutes)

Feature combination (70 minutes)

Regularization: simplicity (40 minutes)

Logistic regression (20 minutes)

Classification (90 minutes)

Regularization: sparsity (40 minutes)

Introduction to neural network (55 minutes)

Training neural network (40 minutes)

Multi class neural network (50 minutes)

Embedded (80 minutes)

Machine learning engineering:

Production environment machine learning system (3 minutes)

Static training and dynamic training (7 minutes)

Static reasoning and dynamic reasoning (7 minutes)

Relationship since data (14 minutes)

Examples of machine learning applications in the real world:

Cancer prediction (5 minutes)

18th century literature (5 minutes)

Real world application guidelines (2 minutes)

From the catalog point of view, this is a machine learning, especially the introduction to tensorflow. Just like the English name of this course, crash course. But the highlight of this course is that it provides a large number of supporting practice and practice environment (IPython notebook), as well as machine learning glossary of Chinese voice, Chinese PPT and Chinese English comparison.

Course video page

At the beginning of opening the web page, the abstract fungus did not see Chinese. At this time, just select Chinese as the language in the lower left corner of the page.

Perhaps the best thing Google can do is to use machine learning technology to generate a full set of Chinese audio and Chinese PPT for the course, and all of them are embedded in the course video. Magical! Abstract bacteria have never seen this operation before.

In other words, the Chinese that uncle said in the video above is voiced by AI!

Let's take a look at the matching exercises of the course.

The following figure shows the IPython notebook cloud platform used for the exercises. Learners can run the code directly in the browser according to the instructions in the tutorial, which is very convenient. Similarly, the page can be freely switched between Chinese and English. Compared with English and Chinese interface, there is no trace of machine translation.

English interface of supporting exercises

Matching practice Chinese interface

Google also provides a Chinese English machine learning glossary, which is also very thoughtful.

What are you waiting for? Open the link to learn! Remember to share your experience with abstract fungus! A kind of

https://developers.google.cn/machine-learning/crash-course/

Today's machine learning concepts

Have a Great Definition

Excellent course recommendation

Data science training camp phase 5

Recommended by excellent ta

Are you dazzled and disorganized by the current data science training market? Don't panic. The old potato driver will hold your hand and point out clearly for you: what is the quality of data science education and training?

The course is full of dry goods and interesting. The lecturer is energetic and loves to share. The assistant teacher carefully corrects and gives enthusiastic feedback.

You're right! Data science training camp is such a star course! From basic Python Programming and scrapy crawler to skillful use of numpy / pandas / Matplotlib / Seaborn / scikit learn and other Python libraries, get through the two channels of machine learning, and complete the gorgeous transformation from data science to ashes players under the polishing of real data science competition cases and data mining projects!

As the assistant teacher of the 4th / 5th training camp, I send a message to the students: keep up with the progress of the course, finish all the assignments on time, and make learning notes carefully. Finally, you will be able to easily start data science!