Nowadays, artificial intelligence has become a hot potato. In the face of the artificial intelligence specialty offered by colleges and universities and the overwhelming variety of artificial intelligence products, Han Liqun thinks that many of them blindly follow the trend. "This is a very dangerous momentum."
This is not the feeling of Han Liqun alone. At the 8th Wu Wenjun Artificial Intelligence Science and Technology Award Ceremony and the 20th18th China Artificial Intelligence Industry Annual Meeting held recently, many experts expressed the same concern.
Why do you want to follow suit? Lu Ruqian, a researcher and academician of the Institute of Mathematics and Systems Science of the Chinese Academy of Sciences who won the highest achievement award of artificial intelligence in Wu Wenjun, bluntly said that our researchers lack self-confidence.
"Our country has done a good job in the application of artificial intelligence, but there is a big gap between basic research and foreign countries." Wang Haifeng, senior vice president of Beijing Baidu Netcom Technology Co., Ltd. pointed out. Wu Wenjun, winner of the Outstanding Contribution Award of Artificial Intelligence, said with regret that the basic theory and technology of artificial intelligence we are using at present are in the hands of others.
In 20 17, the State Council issued a new generation of artificial intelligence development plan, and proposed that by 2030, China's artificial intelligence theory, technology and application will reach the world's leading level and become the world's major artificial intelligence innovation center. According to experts, to achieve this goal, China's artificial intelligence needs to do basic research, follow the trend less and be more practical.
Investment or speculation
From tangible mobile phones and home appliances to unfathomable Alpha Go and driverless driving, artificial intelligence has penetrated into every aspect of life, attracting the attention and capital of all walks of life. According to the data of foreign research institutions, in 20 17 years, the financing amount of global artificial intelligence start-ups reached152 billion US dollars, with China accounting for 48%, ranking first.
Artificial intelligence is like a magic pen, and all walks of life want to add color to it. However, in Han Liqun's view, many investors pouring into the artificial intelligence industry from various fields are speculators, which makes this artificial intelligence craze full of gimmicks and tricks.
Why do you see it? For example, in the field of home appliances, artificial intelligence has become a new selling point, and voice control is realized with the help of technologies such as Internet of Things and voice recognition. However, Han Liqun believes that this is not artificial intelligence, but "information appliances". "Whether you are smart or not depends on whether you have learning ability, adaptability, or some human-like or brain-like functions."
According to Qiao Junfei, a professor at Beijing University of Technology, many artificial intelligence products only stay at the level of "seeing" and "listening". If a person can only see and listen, can he be said to be wise? Obviously not enough. "But we have replaced all intelligence with' seeing' and' listening'." He said that many so-called artificial intelligence products are actually just doing speech recognition and image processing, and the threshold is very low.
Don't do "fast food" research
Experts pointed out that another dangerous trend in the development of artificial intelligence in China is that it is keen to develop the application of artificial intelligence, but it does not pay enough attention to basic research.
"Which direction should we go now?" Qiao Junfei believes that between applied research and basic research, more attention should be paid to the latter. In order to make a fundamental breakthrough in the basic research of artificial intelligence, it is necessary to improve the ability of data reorganization and calculation.
Algorithms, computing power, chips, open source platforms, etc. It is considered as the foundation of the development of artificial intelligence, but this foundation is mostly controlled by European and American countries at present. Han Liqun admitted that some important algorithms were invented abroad, and our technology was too weak. In addition, we are also subject to people for a long time in terms of chips, and this restriction will be more obvious in the era of artificial intelligence. Another example is open source software and architecture. "There are some in our country, but compared with the United States, the gap is obvious." Wang Haifeng said.
As an important supporting platform for basic research in China, the National Natural Science Foundation of China has specially set up artificial intelligence application codes, and concentrated on accepting basic theories and key technology research projects of artificial intelligence. According to Zhang, deputy director of the Information Science Department of the National Natural Science Foundation of China, more than 2,800 applications have been received, but there is still a long way to go. "For example, some algorithms may be successful in some fields, but switching to another field is not so effective." He said that from weak artificial intelligence to strong artificial intelligence, we need to make breakthroughs in basic theories such as general algorithms, and our current research is not enough.
Internal work is not good, artificial intelligence is just a virtual fire. To practice internal skills well, you need to be down-to-earth, not do "fast food" research. Han Liqun warned that if there are too many followers and speculators, who will make a breakthrough in basic research?
"baton" calls for orderly development
It is necessary to strengthen basic research and standardize the development of the industry. Behind the popularity of artificial intelligence is the disorder of development.
At present, artificial intelligence is planned and supported as a key industry in many places. "Even more than a dozen prefecture-level cities in a province are developing artificial intelligence, and the planning is very similar, which may eventually lead to redundant construction and low-level development." Xin Renzhou, inspector of the Industrial Policy Department of the Ministry of Industry and Information Technology, reminded that how to make all parts of the country form a joint force and maintain their own characteristics is an urgent issue to be considered. In addition to regional imbalances, there are also imbalances within the industry. Qiao Junfei emphasized that most of the artificial intelligence talents in China are concentrated in Internet companies, and there are few physical industries. The treatment of these industries is not enough to attract artificial intelligence talents, but if this continues, the ebb tide may come soon.
To avoid low tide, it is necessary to standardize the artificial intelligence products themselves. Han Liqun pointed out that there are more and more products called artificial intelligence, but how to judge whether it is flickering? There is no standard. "Is there intelligence, intelligence is what level, intelligence or retarded? It is necessary to establish an index system suitable for evaluating the IQ level of artificial intelligence. " She said.
Han Liqun believes that the standard is a baton, which plays a guiding role and can promote the continuous improvement of the intelligence level of artificial intelligence products. In fact, the whole artificial intelligence industry needs a "baton" to guide everyone to push the artificial intelligence craze to a higher and more lasting height, rather than coming and going quickly.
Guangming daily (20 19 0 12 10/0 edition)