The relationship between data analysis, machine learning, deep learning, and artificial intelligence I drew this picture
Let me explain this picture.
The emergence of all technologies is to solve real problems, and real problems are divided into simple problems and complex problems. Simple problems require simple analysis. We use data analysis . Complex problems require complex analysis. We use machine learning .
1. What is a simple question?
For example, company leaders want to know the sales situation every week. This is a simple problem. Simple problems can be handled by data analysis, and useful information can be analyzed by analyzing the data.
The simplest is that you use excel to analyze the sales data of a Grocery store. The company will send you a weekly report and find out that the sales volume has decreased in recent months, and then analyze the reasons for the decline in sales. Develop corresponding strategies to increase sales.
Let's look at a real case. Airbnb, the world's largest travel home rental community, struggled with the slow growth of new users in 2011. One day, their data analysis team found that the exquisiteness of the photos of the listings was very positively related to the number of people booking the listings.
So they put forward a hypothesis that "homes with professional photography photos are more sought after, so homeowners will definitely be willing to apply for this service provided by Airbnb."
They quickly launched a version that provided professional photography photo services, and then did an A / B Test with the original version. They found that the same property, using professional photography services, had 2-3 times more orders than did not use it.
In late 2011, Airbnb hired 20 professional photographers to help homeowners on the platform take photos of their houses. At almost the same time, Airbnb's order volume curve had a steep increase.
2. What is a complex problem?
For example, Taobao we use every day, it will recommend products you may be interested in based on your historical shopping habits (data). How does Taobao do it? For such complex problems, machine learning is used behind Taobao .
Let me give another example, how today's headline relied on machine learning to become the boss of news clients.
Around 2010, the three giants Netease, Sohu, and Tencent, which emerged in the portal era, transformed into mobile terminals, which almost monopolized the news client market at the time. And just 2 years later, today's headline, using the "machine learning" Dragon Sword to personally recommend news that users are interested in, breaking the monopoly of giants and becoming the boss of news clients. Although Tencent and NetEase later launched similar products, Daily Express and NetEase, to counter headlines, they failed due to late start and immature algorithms.
An analysis like "people you might be interested in" is a complex analysis that needs to be found through a machine learning algorithm, similar to recommending movies you are interested in on Netflix, and recommending products you are interested in on Amazon.
3. What is deep learning?
There are many methods (algorithms) for machine learning, and different methods solve different problems. Deep learning is a branch method in machine learning .
Deep learning has achieved very good results in the classification and recognition of rich media such as images and speech, so major research institutions and companies have invested a lot of manpower in related research and development. Let me give an example, you must have heard of it. That is, in 2016, Google ’s DeepMind company developed AlphaGo to defeat human top Go players. The main working principle of Alpha Go is "deep learning".
4. What is artificial intelligence?
Artificial intelligence, its scope is very broad. Artificial intelligence in a broad sense refers to the realization of human mind thinking through computers (machines), making machines to make decisions like humans.
Machine learning is a technology that enables artificial intelligence . So I put artificial intelligence, machine learning, and deep learning in different circles. The three of them are inclusive relationships:
Now you have a clear understanding of the relationship between data analysis> machine learning> deep learning> machine learning. When we look at it from the perspective of solving real problems, many concepts will become clear. Deal with different problems and use different methods.
5. What is the relationship between data analysis and artificial intelligence?
You may ask: "I didn't see any relationship between data analysis and artificial intelligence in the above picture. Is it not useful to learn data analysis? Then I learned machine learning from the beginning so that I can directly enter the era of artificial intelligence , Enjoy the dividend of the times? "
It is wrong to think so.
Machine learning is the fusion of knowledge in many disciplines, and data analysis is the foundation of machine learning. Only by learning how to analyze and process data can you understand machine learning. This is like, if you want to go to junior high school (machine learning), you must finish elementary school (data analysis) before you can.
Therefore, I have drawn two yellow lines in the picture below to indicate the two directions of data analysis. If you like to in-depth technology and learn data analysis, you can lay the foundation and learn machine learning. If you like business content, you can go in the direction of artificial intelligence business.
Professional social networking site LinkedIn said in the "2018 Emerging Jobs Report" that in 2018, 6 of the 15 new positions were related to artificial intelligence, which shows that artificial intelligence-related skills have begun to penetrate various industries, not just It's the technology industry.
LinkedIn defines artificial intelligence skills as skills that develop and effectively use artificial intelligence tools and technologies . This is the fastest growing skill on LinkedIn. Globally, this skill increased by 190% from 2015 to 2017.
Many people used to be zero-based before, but bought a bunch of machine learning courses and books to learn, and finally they looked dizzy and felt unsuitable.
Actually, this is the wrong way. If you are zero-based and want to enter the related profession of artificial intelligence, you must start with data analysis.
6.Summary
1) Artificial intelligence instructs machines to make decisions like humans
2) Machine learning is a technology to realize artificial intelligence
3) There are many methods (algorithms) for machine learning, and different methods solve different problems. Deep learning is a branch method in machine learning.
4) Data analysis can help you move from zero to the age of artificial intelligence. If you like in-depth technology and learned data analysis, you can lay the foundation and learn machine learning. If you like business content, you can go in the direction of artificial intelligence business.
5) The following picture is the relationship between them
0 Comments