Exclusive: Singapore’s plan to get ahead on AI

By Nurfilzah Rohaidi

Interview with Laurence Liew, Director of AI Industry Innovation, AI Singapore.

Twenty years before AlphaGo became permanently imprinted in our collective consciousness, there was Deep Blue. This supercomputer bested a reigning chess champion, all the way back in 1997.

Artificial intelligence has been part of our everyday lives for decades, often going unseen – smart washing machines and the Google search engine come to mind, says Laurence Liew, Director of AI Industry Innovation at AI Singapore (AISG). The difference then and now is the higher visibility of AI projects and applications, such as autonomous vehicles, facial recognition and medical diagnostics.

Singapore's national AI programme was launched a year ago to build knowledge, tools and talent to make Singapore an AI leader. Now, through various initiatives, the agency is keenly lowering the barriers to AI adoption for companies in Singapore, regardless of size, he tells GovInsider.

Robotic process automation for startups


AISG is driven out of a partnership between six government entities, and hopes to help Singapore seize the opportunities that AI brings. A 2017 report shows that AI could nearly double Singapore’s annual economic growth rates from 3.2% to 5.4% by 2035. Furthermore, it could increase the country’s labour productivity by 41% by that same year, more than the US (35%) and Japan (34%).

The ideal is to ensure that even small startups can harness AI in impactful ways. In an effort to help SMEs, AISG recently launched an open-source robotic process automation (RPA) tool, TagUI, free for anyone to use.

RPA allows companies to automate certain rote tasks, greatly boosting productivity and efficiency – but is typically a very expensive service for startups to buy or build on their own, Liew notes. “We host that tool now, and have tremendous uptake.”

AISG is also running a 100 Experiments programme which hopes to “infuse” AI into the wider ecosystem. Industry players come up with problem statements unique to their business, and AISG matches these companies with local researchers who are keen to solve those problems, Liew says. “At the end of 18 months, it’s a solution that they can deploy nearly immediately back in the organisation.”

These experiments are wide-ranging. One is exploring how AI can analyse social media and other non-finance data to create a credit score for “people who, for whatever reason, cannot get a credit line with the bank”, Liew shares. Another project, by infrastructure consultancy Surbana Jurong, is looking into predictive maintenance of lifts.

A third is delving into voice-to-text recognition, which could be “very impactful” for government services. Officers will be able to give more attention to a citizen complaint if they did not need to focus on transcribing citizens’ calls, Liew points out. “Multiple agencies have a need for that kind of technology,” he notes. “They have tested all the various commercially available solutions, and they’re just not accurate enough.”

Six Singapore agencies have expressed interest in this solution once complete, he shares.

100 ways to use AI


There are two main objectives for each 100 Experiments (100E) project: developing a minimum viable product; and training the company’s engineers in AI capabilities along the way. “Before we take the project, we want to make sure that you have an engineering team that can co-create a solution with the professor and the team.”

AISG will also provide up to S$250,000 worth of co-funding to work on the project, he adds. The company is expected to match up to this amount - 30% in cash, and 70% in kind to fund research manpower. Ten industry projects have been approved so far, according to him.

He goes on to stress that all of these projects will develop solutions that are currently not available within the industry, so as not to inadvertently take business away from local companies. “There are cases where companies say they want to do this and we say, this is facial recognition – why don’t you just talk to this other company, they’re already doing what you want to do,” he explains.

Startups can register their area of expertise on the AISG website, he adds, so that the AISG may “matchmake” interested companies with the startups that can provide the solution they need.

Key areas

 

Besides increasing industry adoption, AISG has three clear objectives: boosting AI research; increasing the pool of AI talent; and solving national challenges with AI, says Liew.

After half a year in the works, AISG will soon announce a national Grand Challenge for healthcare, which will provide a grant for delivering impactful solutions, he continues. Healthcare is one area seeing great benefit from AI tools and assistants – algorithms can now analyse chest x-rays, detect skin cancers, or screen for diabetes from a patient’s retina scan, he explains.

“Anything that is image-based, you can use AI to learn what the image means and therefore surface a possible diagnosis,” he remarks, adding that Grand Challenges will also be announced for solving issues in two other crucial areas: Finance and Urban Solutions.
 

“Anything that is image-based, you can use AI to learn what the image means and therefore surface a possible diagnosis.”

More engineers, please


As with any other industry, finding the right talent is a stumbling block for many industries, and there is a distinct lack of AI engineers and data scientists in Singapore, according to Liew. Here, AISG is running a 9-month apprenticeship programme that pairs fresh graduates or those with less than three years of working experience with 100E companies, he says.

“Companies are keen to use AI, machine learning, or data science to solve some of the business problems and challenges that they are facing; it’s just that they do not have enough people who are familiar with the tools,” Liew points out.

Hiring becomes even more challenging when tech giants such as Google and Facebook are attracting top talent, with top salaries to boot. “Some of our local companies are facing that crunch,” says Liew. “We hope that our apprentice programme will help resolve some of these problems.”
13 apprentices have just begun last week, and Liew hopes to train 200 over the next three years, he adds.

It has been almost a year since AISG began, and much has happened since then. Liew, who set up Singapore’s first AI lab in a collaboration with SingTel in the 90s, hopes to see more organisations introduce AI, but in a “meaningful way”.

He shares an anecdote of a local cleaning company that had recently introduced automation. Instead of firing workers, the company instead incentivised them to attend trainings and learn how to operate the machines, Liew shares.

Crucially, with the machines, the workers could clean much faster; the company could take on more jobs; and each worker received a salary increase of S$300. “I don’t believe that people cannot learn; you just need the right policies and incentives in place,” he remarks.

In a stunning display of man versus machine, Deep Blue beat a grandmaster at chess. But Singapore’s future could see man and machine, not as adversaries, but as teammates.