China's AI industry needs more fundamental research in real economy

By Zhang Hongpei in Chongqing Source:Global Times Published: 2018/8/27 17:33:40

Industry needs more fundamental research in real economy, stronger talent cultivation: experts

Visitors watch robotic arms play instruments during Smart China Expo in Southwest China's Chongqing on Thursday. Photo: Zhang Hongpei/GT

As China rapidly builds momentum in becoming a global innovation center for artificial intelligence (AI), the country's integral players who focus on advanced technology should "search deeper into fundamental research and applications in the real economy," industry experts told the Global Times during an inaugural technology forum held from Thursday to Saturday.

From voice to image recognition, Chinese tech companies are beefing up efforts in research and application, scrambling for a leading position as the country aims for the goal of becoming a world leader in AI by 2025.

A Beijing-based white-collar worker surnamed Zhao told the Global Times on Sunday that he has been using an online voice transcription app developed by Chinese AI-powered speech recognition leader iFlytek since last year.

"You can put your voice recording into the app and several minutes later, a word version will come out. Although it could understand most of the words, I still needed to go back to the recording to double check when I began to use it last year," Zhao noted.

However, the app's accuracy has improved a lot recently.

"If the recording has a clear voice in Putonghua, the app can recognize everything, even some technical terms. Now I don't need to listen to the original recording," noted Zhao.

Since the app is becoming smarter, it indicates that the machine behind it has been trained better thanks to deeper learning based on richer data collection, enhanced computing power and stronger algorithms, three key elements in the AI sector.

Lou Chao, general manager for Southwest China business at iFlytek, told the Global Times in a recent exclusive interview at the first Smart China Expo held in Southwest China's Chongqing that voice recognition is just the beginning of the phenomenon of machine to human being interaction, while mathematical bases and algorithm models are the technologies supporting from behind.

As far as Lou is concerned, basic research and application can promote each other when the two aspects jointly empower different industries such as financial services or healthcare.

"If technology is applied based on more data in the wider scope, it will in turn promote the advancement of core technology," Lou noted. This can produce a "ripple effect", as described by iFlytek Chairman Liu Qingfeng during the opening session of the Smart China Expo.

Founded in 1999, Shenzhen-listed iFlytek was named along with Baidu Inc, Alibaba Group Holding and Tencent Holdings as part of the government's first batch of national champions driving the country's AI mission.

Hunt for research, specialists

In July 2017, the State Council, China's cabinet, issued a plan for new-generation AI technology, pledging to make the sector a major new growth engine that would improve people's lives by 2020 and make the country a global center and leader of AI innovation by 2030.

Still, China lags behind developed countries in AI theory, algorithms, core components and even talent, experts noted.

Piero Scaruffi, founder of the Silicon Valley Artificial Intelligence Research Institute, told the Global Times during the expo that China really needs to put more effort into fundamental research and inventing the "next thing" when compared not only with the US, but also with other key global AI players like Germany, Switzerland and Israel.

He also said that China could become the best country in the world in terms of AI application.

"Even as early as the 1990s, people thought voice synthesizers could speak and that the technology could be applied in real life, but it still took 19 years for iFlytek to develop its voice recognition technology. That process was like a marathon. We needed to concentrate on and bury ourselves in basic research while withstanding solitude - factors of which determine how far you can go," Lou stressed.

As a significant factor supporting the realization of basic research in the industry, recruiting specialists has become a priority for AI powerhouses like the US and China.

By the end of 2017, China had 18,232 AI specialists, accounting for 8.9 percent of the world's total, while the US held a share of 13.9 percent, according to a report issued by Tsinghua University on July 13.

Scaruffi said China could take inspiration from US universities and tech behemoths to form a long-term plan for talent cultivation and to encourage specialists to be more creative and therefore achieve something that has never been achieved before.

Referring to Chinese collaboration with the US, Scaruffi believes that, "If you remove all these barriers [trade wars and tariffs], there are many ways [in which that could happen]."

"China has a cutting-edge advantage in data and smart city building. For us, the data is like a goldmine to conduct research. On the other hand, Chinese scientists could probably see what research is for us, and that is one way of 1,000 different ways [to collaborate]," Scaruffi noted.

Newspaper headline: China’s AI push should ‘go deeper’


blog comments powered by Disqus