As artificial intelligence has become the core sector of the new infrastructure, in order to export computing power, data, algorithms and even talents to the society in an inclusive manner, and promote the more efficient development of the industry, the head players have taken the step of establishing an artificial intelligence enabling platform. Following the establishment of Artificial Intelligence Computing Center AIDC by Shangtang technology at the end of January this year, the first artificial intelligence Open Innovation Center (hereinafter referred to as "Deyang science and innovation center") of artificial intelligence internet of things company tesilian officially landed and operated in tesilian Deyang aipark.
\u3000\u3000 "In fact, intelligent development faces many challenges. At the technical level, high technology threshold, large talent gap, high investment cost and long cycle have become the difficulties faced by enterprises. In terms of industry, we can see that the majority of small, medium-sized and micro enterprises in second and third tier cities have very scattered needs, but it is difficult to find appropriate solutions to adapt them. At the enterprise level, the enterprise itself is subject to the traditional concept Thinking and path dependence have serious restrictions on talents and technology, and it is difficult to break through the threshold. " Liu Bin, senior vice president of Tesla technology group, said in an interview with the reporter of Securities Daily.
"Therefore, we have built an AI open innovation center in Deyang. AI open innovation center is an important carrier that focuses on the key segments of AI, continuously outputs the core R & D and service capabilities of AI through the integration of technical resources and industrial chain resources, and realizes the AI empowerment at the city level from the three directions of industry, education and training and scientific research," he said
three directions help AI enable
The concept of "smart city" has been put forward for more than ten years. Since the construction of smart city was raised to the national strategic planning in 2014, it has become people's beautiful vision of urban life. In recent years, driven by the rapid development of 5g, big data and artificial intelligence, a large number of smart park projects have been implemented across the country. But at the same time, people also found that at present, the repeated construction of China's smart parks is serious, "one side of a thousand cities", which only achieves the effect of space leasing, and the experience of artificial intelligence is not significant.
On the other hand, AI belongs to an industry that costs both money and manpower. Therefore, although small and medium-sized enterprises have huge fragmented AI needs, it is difficult for traditional AI enterprises to provide targeted solutions for them due to cost considerations, which has also become a major obstacle to the implementation and application of AI.
"So where should the wisdom of the park be reflected? Artificial intelligence can bring strong added value to traditional smart parks or smart cities. Therefore, we build a science and innovation center of artificial intelligence by introducing artificial intelligence technology, so that we can connect the local academic ecology and industrial ecology, and make our aicity (smart City created by Tesla) through cost sharing It can provide the surrounding small, medium-sized and micro enterprises with the core elements such as computing power, data and algorithms they need in AI, so as to promote the more efficient development of the whole industry. " Liu Bin said, "what we expect is to develop the smart park or smart city from traditional space leasing and service leasing to technology leasing through Deyang science and innovation center."
The mission of Deyang science and technology innovation center is very clear. The first is the incubation of AI Artificial intelligence algorithms, the second is the sharing of scientific and technological achievements of industry, University and research, and the third is to promote the training of local AI talents.
According to Liu Bin, if the platform wants to play such a role, it needs to meet high technical requirements. "This platform must be simple, easy to learn, easy to use and easy to understand. It must be easily configured through a variety of low code and drag-and-drop methods. It must be able to give full play to the advantages of rich student resources of Higher Vocational Colleges in various cities, and then make use of these personnel. Only in this way can small and medium-sized enterprises in second and third tier cities be able to use it in a low-cost, low threshold and inclusive way Use the Tesla algorithm platform to solve various problems they encounter in their business. "
open up the industry, study and research back feeding industry
It is understood that Deyang science and technology innovation center has played an AI enabling role in many key directions such as double carbon, energy and medical treatment after its landing.
\u3000\u3000 "For example, there has always been room for improvement in the charging safety of urban charging piles. Some companies found us through Deyang science and technology innovation center and helped them establish a perfect model curve of normal battery charging. Through this cooperation, on the one hand, the academic state will continue to improve the model, on the other hand, the enterprise will continue to improve the system, so as to achieve the self upgrading mode of the whole system in technology, business and mode. All these processes are realized through low-cost shared leasing, and enterprises have established a new security system through small investment. " Liu Bin said.
In fact, it is closely related to the industry to solve the practical application of artificial intelligence in various scenarios, which is also regarded as the biggest advantage of teslian Deyang science and innovation center compared with other AI enabling platforms.
Liu Bin stressed that on the one hand, Tesla can deeply understand the problems of the industry. This does not mean that all computing power can be gathered together to establish a super large computing power platform, but by breaking up computing power, the core technology can be put into cloud computing and federal learning technology. Deyang science and technology innovation center sinks the computing power to the industrial data, which can more effectively revitalize the silent data of enterprises. In this way, the revitalized data are handed over to scientific research institutions. Scientific research institutions learn through the Federation. According to local regulations, they can only see some of these data conditionally. Based on Federated learning, he abstracted a pre training model and then fed back to the industry to make corresponding algorithms.
"We use the pre training model or this core technology to realize the connection between industry, University and research, rather than incubating through the horizontal topic release platform or the ecological industry system as in the past. What we do is to realize the connection between Industrial Education and scientific research through the core technology, so that professionals can do professional things, which is a highlight of us." He said.
For the role that Deyang science and technology innovation center can play and the market space in the future, Liu Bin is very optimistic: "In the past two or three years, although the epidemic has pressed the pause button for our life to some extent, on the other hand, it has also accelerated the development of intelligence. The current AI technology is far from just a slogan of industrialization, but has begun to penetrate into our life. In cities, our scenes are very rich and our needs are very diverse. Both urban governance and urban operation are important for the implementation of our AI technology Provides a broad space to play. "
