16
July
2018
|
18:17
Asia/Singapore

AI transforms frontline operations

From left: Prof Tan Chorh Chuan, Executive Director, Ministry of Health’s Office for Healthcare Transformation; Asst Prof Ngiam; Asst Prof Feng; Prof Wong; and Prof Chen Tsu Han, NUS Deputy President (Research and Technology) and Chief Scientist of AISG at the expo

A new artificial intelligence (AI) platform that uses deep learning technologies will soon be able to harness the vast amounts of data available in hospitals for better healthcare delivery. The innovative “DISCOVERY AI” platform was launched on 6 July at Singapore’s first ever public Healthcare AI Datathon and Expo, jointly organised by NUS, the National University Health System (NUHS), AI Singapore (AISG) and the Massachusetts Institute of Technology.

“Challenges such as a rapidly ageing population, rising incidence of chronic diseases and finite resources are an opportunity for technology to better prevent and manage diseases with optimal outcomes and affordable costs,” said Isabel Chan Professor in Medical Sciences John Eu-Li Wong, NUS Senior Vice President (Health Affairs) and Chief Executive of NUHS.

Jointly developed by NUS and NUHS, “DISCOVERY AI” currently has three AI tools in place to augment the work of healthcare professionals. The first tool is able to read free text and provide a diagnosis for appendicitis with approximately 90 per cent accuracy. The second looks at better identifying high-risk patients, while the third tool is able to predict disease progression by analysing several factors of admission, such as a patient’s blood test results, length of hospital stay and social status.

Explaining the benefits of the platform, Assistant Professor Ngiam Kee Yuan from the NUS Yong Loo Lin School of Medicine and NUHS Group Chief Technology Officer said, “In my lifetime, I probably wouldn’t see all the variations of the disease. Having the machine gives you that edge…sometimes it’s very hard for doctors to make very accurate predictions of what might happen to a patient, we also rely on calculators and scoring systems, but this is a new paradigm using models that are based on past history and records.”

Assistant Professor Feng Mengling from the NUS Saw Swee Hock School of Public Health, who helped to develop “DISCOVERY AI”, likened the AI agent to an infant, saying, “We look at it like a baby, they learn and accumulate knowledge…over time, given more data, they will have a higher proficiency with other diseases as well.”

However, Asst Prof Ngiam was quick to point out that the platform was not built to replace doctors “who can do so much more with their experience and insights” and whose acute awareness and human touch are still necessary, particularly in medical situations such as identifying signs of depression in a patient. Rather, the smart tools would be able to reduce the amount of time spent by doctors and administrators on tedious and repetitive processes, freeing them to focus more on direct patient care.

“DISCOVERY AI” is currently undergoing clinical trials and is expected to be deployed in phases across NUHS institutions in the coming years.

ai-2.png

The new speech recognition system can recognise a mix of languages in one conversation

Another new tool designed to help organisations improve frontline operations is a first-of-its-kind speech recognition system able to interpret and process the unique vocabulary used by Singaporeans — including Singlish and dialects. The tool was developed by the newly launched AI Speech Lab, AISG’s first major collaboration with multiple government agencies to design an AI system that could be deployed government-wide and in future, nation-wide.

The new research laboratory, set up with $1.25 million funding by AISG and located at the NUS innovation 4.0 building, is led by NUS Engineering Professor Li Haizhou, a world-renowned expert in Speech, Text and Natural Language Processing, and Associate Professor Chng Eng Siong from Nanyang Technological University.

Speaking of the novel speech recognition system, AISG Executive Chairman Professor Ho Teck Hua, who is also NUS Senior Deputy President and Provost said, “This is a challenge that is unique to Singapore and the region which is currently not addressed by existing speech engines offered commercially or by major cloud-based AI providers.” He added that the system could also be customised to the business needs of private companies.

The AI Speech Lab has already secured its first collaborator — the Singapore Civil Defence Force — which plans to use the system to reduce the time needed to log in information during emergency calls, where every minute counts.