"Roast conference": where is the weakness of our AIOn the afternoon of December 8, on the eve of the opening of the 2018 China artificial intelligence industry conference, many experts from the industry and academia gathered for a small closed door meeting almost as soon as they put down their luggage. The topic they discussed was the analysis of the weakness of Chinese AI
"less greetings, mainly about the foresight or near worries of the development of artificial intelligence; see how poor China's artificial intelligence is compared with the world." Li Deyi, chairman of the Chinese Association of artificial intelligence, was the host of the meeting and came straight to the point
"I think artificial intelligence is a little abnormal." Han Liqun, executive vice chairman of the professional committee of artificial intelligence and robot education, said. People have poured into AI from different fields, but from the development process of AI, it must have ups and downs and twists and turns. When the boom fades, these gold miners who enter the field of artificial intelligence "may run faster than anyone else". The new generation AI development plan points out that China's plastic extruder industry is moving towards a healthy and sustainable development in 2030, and China's AI theory, technology and application should generally reach the world leading level. "If there are many speculators, who can make a breakthrough in basic research?"
more than 6 After the experiments, which usually require high tensile properties of materials, such as heat shrinkable film and tensile film, the experts at the meeting admitted that none of the current key technologies and basic algorithms in the field of artificial intelligence were proposed by the Chinese
"to be honest, if AI is to become a discipline, it should have a solid theoretical foundation." Jiangchangjun, vice president of the Chinese Association of artificial intelligence, said, "most people follow the task and do 'fast-food' research."
to put it bluntly, it is to follow the trend
the wind is still blowing hard. Chinese scientists have published many papers at various international conferences on computer vision and pattern recognition. However, if a person can only see and listen, it is difficult to say that he has intelligence. But now, "how to look" and "how to listen" seem to be all about intelligence. "Follow the trend, you can go on many projects and send many papers." Qiaojunfei, a professor at Beijing University of technology, said. Only by strengthening basic research can we know what we want to do and what we can do
in terms of application, it is not always a smooth path
Qiao Junfei added that most of the people who really do well in AI are Internet companies, and there are almost no people in the real industry. "The earnings of these industries are not enough to attract AI talents to work, but if this continues, the ebb tide may come soon." If the development of artificial intelligence is divorced from the needs of economic society, it cannot be hot forever. "We should build a bridge between artificial intelligence technology and economic and social development, and in turn promote the development of basic research. We cannot just stop at the stage of comparing speed and precision."
when it comes to application, Li Xiaobin, a professor at Shanghai University of applied technology, is very impressed. He went to an iron and steel company a few days ago. The company wants to do intelligent coking and reduce air pollution, which requires that all equipment and operations be unmanned
the artificial traceless intelligent team met the traditional production equipment team. Li Xiaobin popularized to them what big data is, what deep learning is, and what problems smart computing can solve. "After speaking, the engineers and technicians on site didn't understand." Li Xiaobin lamented that many methods, technologies and applications in the field of artificial intelligence have not really been popularized in basic industries
yaozhiqiang, co-founder of Yuncong technology, believes that the ecosystem of the artificial intelligence industry has not been established. At the top of the pyramid, it should be a large national scientific research institution, focusing on the research of key common problems; At the middle of the pyramid, there should be some leading AI enterprises. "We believe that the pyramid is based on a wide range of thousands of AI application-oriented enterprises that solve smaller and more vertical problems, which is very scarce." Yao Zhiqiang said that at present, research institutions lack contact with leading enterprises, and leading enterprises lack interaction with small and medium-sized enterprises. "It seems that every enterprise and scientific research institution is fighting alone, and the whole ecosystem has not been established."
if the driving force of this round of AI development is data, where is the next round of growth engine? "Many enterprises have made a lot of innovations based on the existing platform and deep integration with various industries." Wang Yanfeng, deputy dean of the school of telecommunications, shanghaijiaotonguniversity, asked, "now we all feel that there is a bottleneck. So, where is the next round of growth? Where is the next round of breakthrough technology?"At the end of the seminar, Li Deyi said that the meeting was over, but similar inquiries should continue
source: pictures in scientific articles come from the network
: Chen Xiaoqi
reviewed by: Guan Jingjing