Focusing on the core issue of 'data', Intel held the 2018 Chinese media 'Enrichment Meeting' in Beijing on April 18. At the roundtable forum of this event, Intel teamed up with a number of industry experts on the future of the Nuggets data. "Innovation and change" this topic, talk about the trend of data flow, talk data innovation and application.
Intel China Media's "Enjoy Club" Holds Roundtable Forum: Nuggets Data Future, Detonates Innovation
The torrent of data surpasses imagination and is the foundation for future innovation
As researchers and explorers at the forefront of the industry, the experts first shared their insights and insights on the data flow.
'For example, a region that wants to focus on optimizing traffic management can all be digitally mapped to the virtual world. In the real world, people, cars, things are mixed together, but in the virtual world it is easy to pull out and perform simulation experiments', Intel Song Jiqiang, president of the China Research Institute, believes that this is not only shallow processing of data, but in-depth data mining, using internal models and laws to make predictions, 'This is a very large value brought by the data, the data flow may be far beyond us. The previous imagination'.
According to Tang Libo, director of the Industrial Internet Research Department of the China Institute of Information and Communications, data is the basis for innovation in all walks of life. All industries will build their future competitiveness in the ocean of data. 'Macroscopically, data promotes us like waves. Going forward, microscopically, we are increasingly conscious of the fact that data has become an indispensable part of social life. He also stated that digitization in various fields is still in its infancy and is moving towards a comprehensive digital society. There is still a long way to go. go.
Tsinghua University professor of computer science and doctoral tutor, artificial intelligence expert Deng Zhidong took autopilot as an example to illustrate the continuous emergence of massive data, we need to approach, reach or exceed the human level of intelligent methods to deal with, analyze data, and explore data Value. In his opinion, we must face specific application scenarios and specific goals, and provide two supports for data-centricity: First, stronger computing power, and second, innovative algorithms.
Hu Guoping, senior vice president of HKUST and president of the Institute, said: 'Data is the foundation for many applications and many 'nutrients' of the algorithm. ' The data can continue to grow, one reason is the strong demand, especially artificial intelligence algorithms. The ability to process data becomes stronger, and because we have to do the corresponding algorithm, for example to obtain more data with more sensors in order to achieve autopilot, this is the trend of positive or even exponential growth.
At present, the data is infiltrating into all walks of life and has become an important production factor. The medical imaging industry is experiencing an explosion of data. Huang Xiongdong, president of Xi'an Yinggu Technology Co., Ltd. believes that in the past, many data were island-style and one device was used. Generates an island of data, a hospital produces an island of data, when these 'islands' are all connected together, is the real big data, 'medical data because of the 'island' is broken, single, frequency and total All have huge data flow, and this data flow has great value for the future medical health.
Release data value, which application scenarios?
Data flooding is not only an explosion of data volume, but also a change in the form of data. If data processing and analysis cannot be implemented effectively, data that cannot generate value will become a 'data corpse'. How can we avoid this trap?
Intel joins hands with industry guests to discuss data flow
Song Jiqiang believes that there are two key points: First, continuous innovation of algorithms and calculations, and secondly, to make data worthwhile in a suitable scenario. He illustrated that 'the deep learning algorithm is used in speech recognition, face recognition, It is to use it well, and there is computational support; applied in security, consumption, automatic driving, this is the scene where the application falls. Both of these are satisfied, and the data is really playing its value.
'Data at the application level, using algorithms to turn it into useful things, is generally partial structure, this is the first layer. The second layer is to knowledge or even decision-making.' Hu Guoping said, 'The core is Find the scene, and then find the data that is helpful to the decision in the scene. In the scene, 'find the end to find the data'.
From the transition from 4G to 5G era, autopilot, car networking is considered as the typical scene of data analysis and processing. So, what exactly does the relationship between 5G and autopilot have? What Tang Libo thinks, 5G will drive automatically. Car networking is taking a new height. It is the general direction of future development. According to Tang Libo, many features of 5G, such as high bandwidth, low latency, high reliability, and compatibility with massive devices, are very suitable for automatic use. Driving and other industrial applications. He further said: 'Car networking is not simply a concept of network connectivity, is the construction of the future ecological system, 5G applications in the future is very important.'
Deng Zhidong added that 5G is more in need of automatic driving than automatic driving, which requires 5G. Automatic driving may be one of the main areas that can really provide demand for 5G and promote 5G commercial landing. Many areas did not put forward such demand. It seems that 5G will bring about disruptive changes, and even require subversive innovations in algorithms.
Massive data also brought unprecedented opportunities for the outbreak of artificial intelligence. In 2017, visual intelligence found a large security scene in China. What area will the artificial intelligence shine in next?
In Song Jiqiang's view, smart retail is a promising application scenario. In order to seize this prestige, the entire retail sales industry, including distribution and supply chain management, has become a system. It is very easy to add portals directly to the front end. Companies that have gained more traffic have an opportunity. They must also have the ability to deploy cloud computing capabilities and can fully mine the value of the data.
Hu Guoping believes that medical imaging and factories will achieve industrial upgrading because of their combination with artificial intelligence. 'There is also a relatively large opportunity in the education industry. Education is a typical slow fever, combined with everyone's investment and willingness to pay, artificial intelligence. The value of technology in this field will be greater and greater.'