Aja Huang, who holds a doctorate in computer science and information engineering from Taipei City-based National Taiwan Normal University, put his education and experience as a top amateur Go player to good use in co-masterminding Alphago’s 4-1 defeat of Lee Se-dol March 9-15 in Seoul.
A member of program developer London-headquartered Google DeepMind since 2012, Huang was responsible for key program adjustments before the series of headline-grabbing matches.
Huang burst on the scene in 2010 when self-developed Go program Erica beat Zen, widely considered the best of its kind at the time, during the Computer Olympiad in Japan.
Professor Lin Shun-shii—Huang’s thesis advisor at NTNU—said the AI expert spent 10 to 16 hours per day for nearly a decade in the lab developing Erica. Huang deserves full recognition for Alphago on the strength of his dedication, proven track record in related research and extensive knowledge of this ancient Chinese board game, he added.
According to a Go-neural networks research paper co-authored by Huang and published Jan. 28 in prestigious journal Nature, Alphago boasts a 99.8 percent winning rate against other programs. It also made history earlier this year after vanquishing a European champion professional Go player in a best-of-five series. (SFC-JSM)
Write to Taiwan Today at ttonline@mofa.gov.tw