The new AI front-runner

By Chen Qingqing Source:Global Times Published: 2017/9/12 17:38:39

With many Chinese talents returning home, China gets set to overtake the US in technology breakthroughs


Visitors experience artificial intelligence (AI) on the Baidu AI Bus during an AI exhibition held by Baidu in Nanjing, capital of East China's Jiangsu Province. Photo: IC





A recent report by Goldman Sachs indicated that China has the resources and ambitious, high-level government plans to support artificial intelligence (AI) development and machine learning over the next few years. In the report, the investment bank identified four key factors for the growth of the AI industry - talent, data, infrastructure and computing power. By now, China already has the first three factors needed to fully embrace AI. Some talents who studied and worked overseas shared their stories with the Global Times to explain why they came back to China.



Wang Jianzong, AI Senior Director of Ping An Group, said he was not hesitant when he made the decision two years ago to come back to China after spending many years in the US.

Wang graduated with a PhD in computer science from Huazhong University of Science and Technology (HUST), Central China's Hubei Province, and then went on to Rice University in Texas as a visiting scholar in 2009. Afterward, he spent the next several years in the US majoring in cloud computing research as well as artificial intelligence (AI) for his postdoctoral.

"I went to study in America because that's the birthplace of cloud computing. However, I came back as I hold optimistic views about China's AI prospective," Wang told the Global Times over the weekend. 

While working as a postdoctoral researcher at the University of Florida, the 36-year-old man also accumulated academic and research experience in deep-learning study and medical analysis.

"I had faith in AI's future at that time," he said.

Wang is among many young Chinese talents in the information technology sector who choose to return to China after studying abroad for years.

In the first quarter of 2017, there were in total more than 1.9 million AI talents worldwide. Over 50,000 of them were created by China, according to a report released by the professional social network LinkedIn in July.

Compared with AI development in the US, China is now outsmarting in the quality and quantity of data, Wang said during the Artificial Intelligence Conference (AICC) held in Beijing on Thursday.

It was the first AI conference in China that attracted hundreds of participants including Chinese Internet giants ''BAT'' - Baidu Inc, Alibaba Group Holding and Tencent Holdings.

The industry is now embracing significant changes thanks to the central government's efforts to push forward AI development in China.

A giant market

Wang is now focusing on deep-learning study and applying technologies in different financial scenarios, such as anti-fraud efforts within insurance services, voice recognition and user profiling.

"Thanks to my experiences overseas, I have built a deep-learning team within Ping An," he said, noting that China now enjoys abundant AI data as well as powerful algorithm platforms.

Also, compared to the US, China has more open access to data, he opined.

AI's development over recent years has been driven by three fundamentals - high chip processing performance, data and algorithms.

A server's processing speed is now 60 times faster than it was two decades ago, and its high performance, in addition to abundant data amid Internet of Things, is supporting the AI boom in China.

Both the US and Chinese governments have moved the AI industry to a strategic position recently. For example, the Obama administration rolled out several national plans to support the growing AI industry in 2016. And the US government has become fully aware of AI's impact on society, including on employment, education, public safety, and national security.

The State Council, China's cabinet, unveiled new guidelines for the development of the AI industry in July, which it hopes will motivate the sector to reach the same level as other advanced countries in terms of technology and applications by 2020.

Also, the national AI technology market scale is expected to surpass 150 billion yuan ($23 billion) by then, and the related industry scale will be over 1 trillion yuan.

"The government support is a boon for enterprises in both R&D and application," said Hu Leijun, vice president of Inspur, an information technology company based in Ji'nan, capital of East China's Shandong Province.

The company provided about 20.4 percent of the servers in the Chinese market in 2016, which have now become a major partner with BAT and other Internet start-ups working on AI development.

US chipmakers have taken a leading position in the basic theories of some core technologies of AI, such as central processing units and graphics processing units.

But Chinese tech firms have been catching up rapidly as the booming market attracts more investment and resources, Hu told the Global Times.

Application-driven growth

Some Chinese AI talents graduate with their first and second degrees in China but pursue PhDs in the US. Zhou Xi, president of CloudWalk Technology Co, a company specializing in facial recognition technology based in Southwest China's Chongqing, is one of them.

"I used to do research work at NEC labs, where many important figures in AI such as Yann LeCun and Leo Bouttou had tapped into the deep-learning sector," he told the Global Times.

The scientist was referring to US-based NEC Laboratories America Inc, which has locations in Princeton, New Jersey and Cupertino, California. Its areas of research span across various sectors including data science, machine learning as well as optical networking and sensing.

"Image recognition is meaningful to the universe. One day, I saw a news report about how a camera installed in a swimming pool could automatically recognize if someone was drowning by analyzing their behavior," Zhou said, noting that the technology can help people in many other ways too.

Between 2007 and 2011, Zhou and his team won several visual competitions, for example, the ImageNet Large Scale Visual Recognition Challenge, run by US-based large visual database ImageNet.

"Those honors made me happy. However, they [the technologies] have to be applied in daily life to help others, like the camera in the swimming pool," he said, noting that how to actually apply the technology was the reason why he returned to China and became an entrepreneur.

Today, the application of AI technologies such as image and voice recognition has extended into many sectors in China. A bus system can now use data algorithms to analyze daily passenger flow in order to manage bus schedules.

Facial recognition systems can help police identify criminals in public areas and are now being widely used in railway stations, metros, airports and so on.

Voice interaction technology is enabling home appliances to become smarter, and Chinese tech giants such as Xiaomi and Alibaba have all unveiled features like smart speakers for such products.

AI will reshape five major sectors - public safety, Internet and e-commerce, consumer electronics, automotive and medical services - in the coming years, according to a report released by China International Capital Corp (CCIC) in June.

And application by governments, other authorities and large-sized enterprises will soon account for over 60 percent of the market share in China, CCIC forecast.

 "If we always stay within the academic field without actually using it [technology] in real life, we can't make any real changes," Zhou said.



Posted in: INSIGHT

blog comments powered by Disqus