AI achievements
With artificial intelligence (AI) widely perceived to be a technology that will shape the future of how people live and work, companies in China have been beefing up investment in the sector in recent years. Scientists believe that the odds favor Chinese companies to be able to match their U.S. counterparts in the AI technology race.
While U.S. tech giants like Google and Microsoft are racing to infuse artificial intelligence into their core products, tech companies in China are keen to ready their artificial neutral network to solve complicated problems of city governance and healthcare.
Traffic congestion, a global problem that troubles urban residents, is an area that has long been a focus of the artificial intelligence (AI) team of Aliyun, Alibaba Group Holding's cloud computing unit.
The AI team of Aliyun told the Global Times on Monday that it has made progress in fixing the problem over the past year as its self-developed AI system, dubbed ET, is breaking new ground in areas like video recognition and traffic prediction.
In Guangzhou, capital of South China's Guangdong Province, at the intersection of Nanhua Middle Road and Baogang Avenue, one of the busiest crossings in the city, the average traffic jam index monitored by local authorities between 9 am and 1 pm fell 25.75 percent after using ET in September in 2016, the team disclosed in a statement.
How does this work?
Hua Xiansheng, a visual computing scientist with Aliyun, explained to the press in a meeting in October 2016 that after consistently learning and analyzing large amounts of the city's transportation data, ET built a virtual model of the city that is able to predict how many vehicles will be on the roads at any given time.
Based on those predictions, the computer will then come up "smart" measures such as intelligently controlling traffic lights to help traffic flow better, Hua said.
Reshaping city governance is just a part of Aliyun's AI empire.
Led by Zhou Jingren, who was part of the research group on Microsoft's Bing search infrastructure, the AI team vowed to train ET into the successor of Jack Ma Yun after 20 years.
Betting on AI
Alibaba has already shown its ambition to develop AI.
Another tech giant, Tencent -Holdings, led a group that made a $10 million investment in February 2016 to fund the AI start-up Diffbot in the U.S.
Baidu Inc, which some consider the Chinese equivalent of Google, has also been devoted to AI for a long time, with Andrew Ng, the former leader of Google's Brain project, leading its research. Ng joined Baidu in 2014, the same year the company reportedly invested nearly 7 billion yuan ($1.08 billion) in cutting-edge technologies such as AI.
Unlike Alibaba, Baidu is keen to use AI to improve the healthcare industry.
China's healthcare has massive potential that can be fulfilled with AI, Baidu founder Robin Li Yanhong said in a meeting on February 8.
In October 2016, the company has released an app known as Melody, designed to act like a medical assistant. The app has an AI-powered chat program that can help collect a patient's symptoms and then provide advice to a doctor. It is not a replacement for doctors yet.
Li predicted a future in which AI is used to diagnose illnesses and recommend treatment based on a gene database.
Baidu's advances were underscored in January when Lu Qi, a veteran AI expert, left Microsoft to join the Chinese company as its COO.
The Chinese search engine powerhouse is now considered to be the leader of the pack among Chinese Internet giants in the race to develop AI technology, Forbes reported on Monday.
What's interesting about AI at this point in its development is that leading technology companies are not the only players.
"As long as your innovative capabilities are strong enough to stay on the cutting edge, you don't have to worry about competition or copycatting from big rivals," Xu Li, CEO of Beijing-based AI start-up SenseTime Group, told the Global Times.
With more than 50 researchers and more than 80 engineers who used to work for Microsoft and Google, Xu's team have made notable achievements in image recognition that have been widely applied in surveillance by governments and companies.
He said that SenseTime's supercomputer DeepLink, which employs 200 graphic processing units (GPU), can recognize 4 million faces in six hours, calculating faster than AlphaGo's 170 GPU-powered computer.
Odds in China's favor
Scientists believe that Chinese companies will have a chance to match or surpass their U.S. competitors in the AI era.
"China [with the world's largest online population] can provide larger amounts of behavioral data than the U.S., which will put Chinese firms in an advantageous position in the AI battle," Chu Min, a scientist with the Aliyun AI team, told the Global Times on Monday.
As everyone knows, AI needs data to learn.
As of December 2016, China has 731 million Internet users, equal to the population of the European Union, with an Internet penetration rate of 53.2 percent, according to a report issued by the China Internet Network Center in late January.
In 2016, 496 million Internet users made payments with their smartphones, 168 million hailed rides and 239 million received government services via mobile apps.
In addition, the Chinese government has moved toward a quick adoption of new technologies, which will bring about new opportunities, Xu said, referring to China's "Made In China 2025" plan.
"Another thing in China's favor is that top AI talent, at least those who are Chinese, will return to China, as Trump administration's immigration policy has already sparked fear among immigrants," Yang Jing, an industry expert and the founder of Ai Era, which focuses on AI technology, told the Global Times on Monday.
In the U.S., AI professors expressed concern over the future of the technology's development, with some worrying that the Trump administration will "dramatically" cut funding for research in the area, according to an article published by the U.S. IT trade publication TechRepublic in November in 2016.
China's AI market is expected to grow to 38 billion yuan in 2018, up from 23.9 billion yuan in 2016, according to a report issued by Beijing-based CCID Consulting in late January.
Experts like Xiang Yang, an AI analyst with CCID, believes there are still some obstacles ahead.
"We have indeed achieved some world-class cutting-edge technologies in the areas like speech and image recognition. But when it comes to core hardware, such as the GPUs we use to run those AI technologies, they are all originally designed by U.S. companies," Xiang told the Global Times on Monday.
Xiang called on the central government to issue more detailed policy support and guidance for the AI industry to attract sufficient private funding to speed up the development of the technology.
