Huawei scientists decipher: Big data is commercial hype? What is the future technology trend?

Author: Yangqiang, academician of Chinese Jieshou International Advanced Artificial Intelligence Association (AAAI), director of Huawei Noah's Ark Laboratory, professor of Computer Department of Hong Kong University of Science and Technology, outstanding Scientist of ACM. Here are answers to three hot questions for the big data industry Yangqiang.

The first time I heard the word "Big data" was a large data seminar for the American Academy of Sciences, held in Singapore in 2011. As a result of the sharp drop in data acquisition costs, resulting in a large number of figures, this phenomenon has become the focus of attention for the first time. There was no unified view of what was meant by the sharp increase in data at the time, and the only sanction was to name the phenomenon "Big data".

Today, all walks of life have different interpretations of big data. What can the telecoms industry do with "big data"? "So far there is no conclusion. After the publication of the article on the large data, the readers put forward many questions, because the space is limited, choose three questions of high concern to answer, give a discussion for everybody to discuss further.

The industry defines large data as 4 "V", that is, large volume (Volume), multiple species (produced), fast (velocity), and high authenticity (veracity). If you just look at these dimensions, then the big data is hype. Because it does not explain the nature of large data. The essence of large data is the ability to think like a person by adding a computer to the data. This ability is embodied in business as a new and better mode of business operation.

Noah's Ark laboratory in Huawei Terminal "wisdom sinks Cloud" application, is a typical large data application case. Huawei is a cloud store in Huawei's terminal, with tens of millions of users and hundreds of thousands of of apps, each user's habits, interests, choices and uninstall applications are different. The sum of the data for these behaviors creates a unique app recommendation question for each user: When the user comes to Huawei again, how to accurately recommend to the user the apps he wants most, rather than simply recommending the most popular apps to the user?

This business model can be expressed very clearly: "How to improve the users ' acceptance of app recommendation by using the big data analysis of Huawei mobile phone app market to enhance the user experience of Huawei Terminal?" ”

Finally, we use tens of thousands of user characteristics of a variety of dimensions, the establishment of a very sophisticated user model. The recommended effect is more than 70% higher than the previous wisdom cloud.

The key to the success of large data applications is also to see if we have a clear commercial (or scientific) purpose. The definition of this business model is necessary.

Pipeline big Data and Internet big data, in the end who is the dominant? What's the difference? Where are their values? This may be the most tangled issue of big data for operators and device providers.

The relationship between the Internet and the operator can be understood by the relationship between the car and the road. Vehicles on the road can be viewed as the Internet, and the goods and passengers and transportation systems on the vehicles can be seen as data and applications of the Internet, and the highways on which the vehicles take are similar to those provided by the operators. For the internet, they are more concerned with the passengers and goods, and the people and freight to the destination. But from the operator's point of view, they are more concerned about the smooth road. From this point of view, the Internet data are related to people and goods, and operators of data is the flow of traffic and road congestion level. Therefore, the data of the Internet is the end user's data, and the operator's data is about the data.

Data on data is of great importance in the telecommunications industry. There is, of course, a prerequisite: resources are limited at any time.

It's a metaphor for cars and roads. How to open a fast track for some important regulars? You need to know which are important regulars first. Which major vehicle companies are attracted by rival highway companies and are considering changing routes? You need to analyze the pain points of these companies. Which areas need new highways? You need to analyze the location and the flow of traffic (open up new business operators). Which areas can be directly built? You need to understand the status of the region and the stage (for mature operators can be directly up to 5G).

The need for data analysis has also improved with the advance of operational technology. In the 5G scenario, we need to provide more intensive, faster and more personalized telecommunications services to the public. Then, we need to know the user's use of the law, Pain points, service weaknesses in where. A high-end service to your shadow is not a myriad of waiters waiting in all the places you might be, but a smart waiter in time when you need it.

Future network technologies, such as software definition networks (SDN), require large data support: SDN's brain, which can generate, modify, and accurately predict the future end-to-end communication needs of a large network of data, continues to learn from the data. In this way, the entire network, like our human brain, becomes more and more intelligent.

The changes brought about by big data are only one step in the transformation of computer technology. The process of change, like other important changes in human history, must be regulated by the original accumulation of resources (i.e. large data), the differentiation of business and social services, and the normalization of human's industry and society to the virtual world to solve the fair and reasonable allocation of data resources. When the primitive accumulation of large data and its technology is stabilized, people will enter the stable state after the digital application.

As a corollary, the next-generation technology, triggered by large data, is likely to be a larger, digital-oriented shift that will transform and integrate many traditional industries into the digital world. This shift has led many areas that require many experts to appear in another form, and have led to changes in the overall "food chain" upstream and downstream of many industries. For example, will the "tall" industries such as doctors and scientists become "workers" who are only responsible for data collection and interpretation of analytical results? Or become a partner in a large data-driven robot?