Wednesday, March 18, 2015

Banking on 'Watson' saves time for managers

MAR 14, 2015

BY AW CHENG WEI

IN THE depths of a DBS data centre, "Watson" goes about his job of remembering an estimated 800 reports of financial information that the bank produced for the year's quarter. Watson can ingest thousands of documents a day, at superhuman speed.

Watson is a cloud service that can understand human language and context, and deliver insights, saving bank relationship managers at least two hours of reading each day, says Mr Olivier Crespin, managing director and head of digital bank at DBS, who initiated the project last year.

Before this, bankers had to identify investment ideas from a wide body of financial reports.



On Monday, Manpower Minister Tan Chuan-Jin mentioned Watson in Parliament as an example of how automation is changing jobs at a faster pace than before.

While there has been a push for automating blue-collar jobs, technology is seeping into white-collar ones, reducing the reliance on brainpower. For the latter, the trend is towards deeper co-existence between man and machine as such jobs require discretion and judgment.

"You still need someone to input information or check the finished product," says Frost & Sullivan analyst Mark Koh. He added that in most industries, users make the final call after consulting the likes of Watson.

Globally, machines are moving into white-collar roles. Last year, news outlets such as the Associated Press and Los Angeles Times started using algorithms to write financial, disaster and sports articles.

Artificial intelligence is also used to scour legal documents for anomalies, draft contracts and collate data for lawyers.

In the region, oncologists at Bumrungrad International Hospital in Thailand began using Watson last year to devise treatment plans based on patient profiles, medical evidence and published research.

At home, the medical sector, too, appears to be taking the lead. Insight understands that a local cancer centre will begin to use big data in its research later this year. Oncologist Iain Tan, from the National Cancer Centre Singapore, says: "The accumulated knowledge from thousands of like patients becomes available as powerful support tools for doctors to choose the best treatment for each patient."

Currently, oncologists rely mostly on their previous experience in treating similar patients. At Tan Tock Seng Hospital, nurses upload pictures of wounds and sizes into a software called eWound Clinical Decision Support System.

Set up in 2013, it provides diagnosis and suggests treatment options. Previously, the nurses would wait for a specialist to decide on treatment.

Nurse manager Betty Khong says: "This process lowers the risk of wound infection or deterioration."

In education, to prepare the country's labour force for the growing emphasis on data analytics, the Infocomm Development Authority piloted a training course in data sciences last year with online course provider Coursera.

Mr Deepak Ravindran, partner and managing director of The Boston Consulting Group, says humans will need to become more skilled and educated in order to use, programme and co-exist with machines: "Though machines can recognise pattern, there is huge effort required by humans to programme them."




No comments: