Almost zero chance of Ke winning a game as AlphaGo upgrades
Chinese Go grandmaster Ke Jie, the world's top-ranked player, lost the first game on Tuesday of a three-game match against Google's AlphaGo AI as part of a five-day Go event in Wuzhen, East China's Zhejiang Province, adding another footnote to the computer program's virtual invincibility in the ancient Chinese board game.
"AlphaGo has done a splendid job," 19-year-old Ke, a native of Lishui, Zhejiang, told a postgame press conference. Some of the moves AlphaGo made were never-before-seen and unexpected, he said, adding that humans would not play that way.
"I sincerely accept the defeat," stated the reigning world champion of the complex board game. The match will be his last face-off against AI-powered computer Go programs, as the chance of humans defeating computer Go programs will keep shrinking, Ke told reporters.
Ke faces AlphaGo again on Thursday, with the last match on Saturday.
Tuesday's result makes the case for the formidable power of the Go-playing machine. AlphaGo, playing against itself and human experts, has arguably established its reign over the game.
"There's almost zero chance of Ke winning a match against AlphaGo. With the enhanced processing power today, AI systems are capable of beating humans, especially in chess games," Xiao Xiao, partner and executive director of Changzhou High-Tech Zone Airwheel International AI Incubator, told the Global Times on Tuesday.
"Even if Ke wins one game, there's scant room for Ke to further improve, but there is significant room for AlphaGo to upgrade its computing power," Xiang Yang, an industry expert at Beijing-based CCID Consulting, told the Global Times on Tuesday.
In Ke's own words, "AlphaGo still played in a human-like fashion one year ago, but now it's become the God of Go." The AI won 4-1 against star South Korean player Lee Sedol in Seoul in March 2016.
"We want to use AlphaGo as a tool for the Go community to improve their knowledge of the game," Demis Hassabis, co-founder and CEO of Google DeepMind, the developer behind the AI, told reporters on Tuesday. He revealed that some announcements are likely to be made later in the week, on "details about the architecture we're using, maybe also some of the games AlphaGo play against itself."
Room for improvement
While AI is making inroads in a broad range of areas including medicine, education and entertainment, there are bottlenecks in its development. As CCID Consulting's Xiang put it, in terms of hardware, AI is still weak in interaction with humans and motion control, and in terms of software, there is much room to improve the quality of voice and images.
"These bottlenecks are affecting the user experience either for individual customers or for businesses," the analyst said.
Still, it is an arena that no one wants to miss out on. It's not only Google that is investing heavily in AI, but an increasing number of technology firms across the globe are betting on burgeoning AI technologies and applications.
Domestic tech behemoths, including Baidu, Alibaba and Tencent, have largely committed substantial resources into AI-related fields, hoping to tap the next big thing. Other than that, a multitude of homegrown startups have also joined the AI trend. Xiao's AI incubator, only launched last week, is a conspicuous suggestion of growing enthusiasm in the country toward investing in the AI space.
Industry insiders say that China and the U.S. are leaders in AI technology and China even beats the U.S. in terms of facial recognition technology and algorithm of images,
But China's weakness in the field lies in its much lower level of R&D investment in AI, compared to the U.S., according to CCID Consulting's Xiang.
"China's AI infrastructure development also lags behind, which means the economic benefit the technology can bring to China are lower than that in developed countries," he noted.
