Shulou(Shulou.com)11/24 Report--

At the beginning of November, the Chengdu Station was successfully held by Kunpeng developers. Ren Fuji, chief scientist of the University of Electronic Science and Technology, attended the main forum and made an opening speech. Under the background that everyone is talking about what the next "popular style" application will be, as a scientific and educational worker who has been engaged in advanced intelligence research all over the world for many years, the academician begins with a clear meaning at the beginning. It is pointed out that only by facing squarely the building of computing platform and the training of international computing talents, and doing these basic work in a down-to-earth manner, can the independent innovation of the computing industry be stable.

Ren Fuji, academician, chief scientist of the University of Electronic Science and Technology

As a scientist in the field of advanced intelligence and emotional computing, Academician Ren has a profound understanding of the current scientific research difficulties and future development in the field of computing. He gives three points:

In the future, there will be more and more topics based on interdisciplinary research, and even become the mainstream of future scientific research.

In order to make a breakthrough in the field of emotional computing, Ren Academician and his team need to do a lot of "interdisciplinary research", not only collecting big data, but also deeply studying brain science and psychology. For example, people's psychological status can be divided into several independent states, through a large number of population survey statistics, big data analysis, get a transfer map, that is, an emotional computing engineering model. When people's emotions are stimulated by the outside world, based on this engineering model, any academician can predict the present or even the future state through calculation. Based on these research results, Academician Ren first put forward the theory of "mental state migration network" in the world, which laid the foundation for advanced human-computer interaction and emotional robot.

Second, before the emergence of revolutionary computing platforms, supercomputing will always be an indispensable infrastructure for basic scientific research and the application of artificial intelligence, and intelligent applications such as deep learning will lead to a revolutionary technological integration of intelligence and supercomputing. The development of AI for Science will accelerate this process.

Academician Ren pointed out that at present, a large number of data, models, and new scientific hypotheses continue to emerge, and the collection of big data, analytical data, and machine learning pose a great challenge to the performance indicators of the computing platform. As for the data-driven artificial intelligence, the theoretical dividend has been exhausted, so we must explore a new research paradigm and a new model of computing platform. Before 2010, the computing power required for training data basically complied with Moore's Law, doubling every 20 months; since 2010, machine learning has doubled every six months; by 2022, machine training requires 10 billion times as much computing power as it did in 2010. Industry calls for a new generation of transformational computing platform. On the other hand, Kunpeng Computing adopts the route of ARM technology and develops the hardware and software infrastructure of full-stack general computing. Based on the hardware root technology, create an open source operating system and database, and have a full set of tools such as application enabling. Ren Academician believes that Kunpeng Computing is a very comprehensive and open industrial ecology. At the same time, Huawei also has Teng AI Computing, and Kunpeng Computing has the ability to create a future-oriented "supercomputing + intelligent" computing platform that can meet the mixed accuracy.

Academician Ren called on more Chinese companies like Huawei to have the courage to explore, research and provide such a computing platform as soon as possible.

Third, the biggest difference between computing and other industries is that it takes a vibrant ecology to release the energy of technology. Technology is only one of the development of independent innovation in the computing industry, and talents with industrial perspective and innovative technology are also the key. Both of them are indispensable in the process of promoting independent innovation in the computing industry, such as "two wings of birds" and "two wheels of cars". Computing developers are the foundation, and they are the key force to go deep into basic research, make up for the shortcomings of the industry, and realize industrial innovation. Facing the future, it is necessary to build a new talent training mechanism and train developers with interdisciplinary background and taking into account both theory and practice, so as to adapt to the future research and innovation and transform the value of technology.

Finally, the industry that simply focuses on hardware performance is different. The computing industry is an industry that needs to be based on a high-performance and open hardware platform to create a lot of "soft technology" and "soft power". In order to truly form a prosperous ecology. Only with such an ecology, can we finally realize the independent innovation of the computing industry under the background of technological innovation changing with each passing day.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.