Homegrown computer scientist powers Taiwan’s AI development
2018/03/26
While some are skeptical about the prospect of Taiwan in the latest global artificial intelligence race, the issue is best analyzed by Ethan Tu, co-founder of PTT, a text-based communication platform that has grown into the largest bulletin board system in the ethnic Chinese world. He is also founder of Taiwan AI Labs.
Tu, 41, began to show his talent at a very young age. During his sophomore year studying information engineering at National Taiwan University in Taipei City, he used a 486 personal computer and set up PTT. The following year, Tu joined the research team at Yam—one of the earliest web portals in Taiwan—to build the country’s first search engine.
In June 2000, the Human Genome Project public consortium announced it had assembled a working draft of the sequence of the human genome: the genetic blueprint for a human being. Inspired by this development, Tu joined a genome research project at the U.S. National Institute of Health. Three years later, he began working at the headquarters of Microsoft Corp. developing Bing, the software heavyweight’s search engine. It was also during this time Tu started taking part in AI research.
In the following decade, Tu and his team at Microsoft successfully created Cortana, the intelligent personal assistant for the Windows operating system. It was this achievement that helped Tu become the head of R&D for Microsoft in the Asia-Pacific. Under his leadership, Cortana expanded into other platforms, with more than 1,000 businesses using it to offer services.
According to Tu, he left Microsoft to set up Taiwan AI Labs largely due to the deep pool of homegrown software talent waiting to be tapped. “I know there are plenty of software talents in Taiwan,” he said. “When there is yet to be a new order, one can achieve one’s goal in any place.”
Tu’s determination for a homecoming was made more resolute after receiving strong support from Minister without Portfolio Audrey Tang and Minister of Science and Technology Chen Liang-gee.
With several of his partners contributing the necessary capital, Tu quickly assembled a team of software developers to take on his AI dream without having to solicit funding from venture capital firms or institutional investors.
For the first stage of his venture, Tu is focusing on medical applications, smart city development and human-machine interface. The choice of medical applications is a natural move given the wealth of medical data collected over the past two decades under the country’s National Health Insurance program and his genomic research experience.
Chen’s partners in this endeavor include Chen Chien-yu, a professor in the Department of Bio-Industrial Mechatronics Engineering at NTU, Dr. Chen Pay-long of the Department of Medical Genetics at NTU Hospital and Chang Tien-hao, a professor in the Department of Electrical Engineering at National Cheng Kung University in Tainan City, southern Taiwan.
According to Chen Pay-long, the team is applying AI to identify the correlation between specific genetic variations and existing known diseases. As genetic variations do not necessarily lead to a disease, the team is trying to discover what genetic variations are likely to cause harm to the human body. “Genomic medicine is full of explosive potential,” he said.
As to smart city development, Tu said the decision has everything to do with “Beyond Beauty: Taiwan from Above,” a 2013 documentary directed by the late filmmaker and photographer Chi Po-lin. After Chi died in a June 2017 helicopter crash while shooting footage for the short’s sequel, Tu said on his personal Facebook page he would carry on Chi’s legacy.
Tu believes high-risk filming can be carried out by unmanned aerial vehicles. After Tu proposed the idea on his Facebook page, the administration at Southern Taiwan Science Park invited him to join a Tainan smart development project using AI and UAV applications. In addition to filming and mapping, the project will employ AI to conduct environmental monitoring, historical site maintenance and traffic control.
According to Lai Wei-hsiang, chair of NCKU’s Department of Aeronautics and Astronautics—the body tasked with overseeing the project’s UAV development—AI technology is not yet advanced enough, but there has been some progress. The team is also working on enhancing the image quality to 4K resolution.
With the various projects moving ahead on schedule, Tu said he plans to expand the headcount of his lab to 60 later this year. But not just any talented techie will do.
“One must understand where one’s real interest lies,” he said. “Life experiences will make your work so much more complete.”
Tu also highlighted the fact that while many AI talents in Taiwan do not have an undergraduate background in the discipline, they continue to enrich their expertise through self-education and taking part in Kaggle—the biggest data analysis competition in the world. His team comprises people with such diverse backgrounds as civil and medical engineering, as well as medicine.
Whereas the global technological development was driven by hardware manufacturers in the past, the situation is now in the hands of software heavyweights such as Amazon.com Inc., Apple Inc., Google Inc. and Microsoft.
But the hardware manufacturing-oriented business environment of Taiwan leaves very little room for software professionals, and many were forced to look for opportunities outside the country, he said, describing those who remain in Taiwan as “software vagabonds.”
Tu recalled that when he returned to Taiwan a year ago, the skeptical said AI was nothing more than a bubble. But Tu remains determined, as he believes such doubt is natural during the rise of a new technology.
The key is to find the proper application to define the new order. “When nobody believes it is possible, it takes someone reckless like me to make the first move,” he said.
Source: Taiwan Today (https://taiwantoday.tw/news.php?unit=7&post=131517)
