Guangdong to be new hub of AI industry

The Canton Fair, one of the world's largest trade shows that kicked off on Thursday in Guangzhou, South China's Guangdong Province, is located in the Pazhou area, a growing industrial cluster of information technology, artificial intelligence, industrial Internet and other innovative technologies, where over 30,000 companies have settled, including Alibaba, Tencent, Xiaomi, TCL and iFlyTek.

Apart from enterprises, laboratories are also moving into the region to enhance research and development (R&D). Pazhou Lab, or AI & DE Guangdong Province Lab (GZ), is an example.

The Pazhou Lab is located in the Pazhou industrial park of AI and digital economy for the Guangdong-Hong Kong-Macao Greater Bay Area, and vows to be a hub of science and technology innovation, Yuan Ziwei, a publicity agent of the Pazhou Lab, told the Global Times on Thursday. 

Yuan said that the laboratory serves as a source enterprise, which can solve the core problems of intelligent systems, in a bid to empower production enterprises. The Pazhou Lab just began to move into the industrial park in June.

Tian Yun, vice director of the Beijing Economic Operation Association, told the Global Times that the construction of laboratories and industrial parks will help Guangdong, and Shenzhen in particular, address its shortcomings in education and research, thus supporting Guangdong's goal of becoming a hub of scientific and technological innovation.

To increase the R&D ability in Guangdong, Han Jiuqiang, a professor of Xi'an Jiaotong University, was one of the experts, scholars and scientists that came to support Guangdong. Han just became the director of the Intelligent Production Line Technology Center in Pazhou Lab.

"Existing machines in industrial production lines can improve the efficiency of mass production. But now the situation is different, as people's consumption level is higher, resulting in more pursuit and demand for niche, personalized products, which is a challenge to the current production line. If the output is too small, the cost will increase. It would also be too costly to have different machines for each different product. So, we have to have machines and robots that can do different things at the same time, and that requires a lot more from the chips or software systems that are at the heart of how the machines work. With that out of the way, how the machines on the same line can work together is the next step."

According to Han, all smart devices today are not really smart, because these machines and robots are only capable of learning, and can only have one function.

"AlphaGo, for example, is very good at playing chess, but it can only play chess. If it is asked to play mahjong, it is not capable. Human beings, on the other hand, are versatile and can have countless skills at the same time. A machine that can learn and have multiple skills at the same time is truly intelligent, and that's what our lab is working on. We have made some breakthroughs, but it still needs time to realize in industrial production," Han added.

The power of AI industry construction has made some achievements in Guangdong Province. One of the examples is the Shenzhen-based Skz Intelligent Equipment Co, of which independently developed and produced intelligent steel processing equipment have been widely used in civil engineering fields such as highways, high-speed rails and tunnels. 

Products have been exported to Europe, America, the Middle East, Southeast Asia, South Africa and other international markets, a marketer named Hu told the Global Times on Friday. The company is also an exhibitor of the Canton Fair, and overseas orders are roughly flat from last year.

"The intelligent system, which is the core of the smart machinery and equipment, is developed by our own company. Our products and system have received a number of invention patents, utility model patents and software copyright in the field of intelligent construction," Hu explained.

But analysts warned that the investment in R&D and long-term return on investment are great burdens for enterprises, which will frighten some to invest large sums in research.

"The existing production enterprises cannot tolerate the status quo of long-term input without output. Therefore, the original innovation of high-tech in the smart field, such as the third generation of the semiconductor, AI and mechanical automation, is quite difficult, which cannot be undertaken by an enterprise alone. Instead, resources in many fields of industry, education and research need to be planned and developed as a whole. It needs the power of the nation to make overall planning, support theoretical research, and finally, complete the process from theory to industrial practice," Zhang Xiaorong, director of the Beijing-based Cutting-Edge Technology Research Institute, told the Global Times.

Han also said that overall planning of the country is necessary. "If the government can integrate and allocate resources, and make the country concentrate on research, China can surpass the US in AI technology within 5 to 10 years. At present, what China needs to strengthen is its theoretical research."