Is it possible for “cold” computers to ever understand us?
This is the question Dr Desmond Ong constantly asks himself. Yet he is hopeful that his research brings humanity one step closer to this possibility each day.
The 31-year-old is a research scientist at the A*STAR Artificial Intelligence Initiative and Institute of High Performance Computing. His research focuses on human reasoning with the aim of improving artificial intelligence.
But Dr Ong, who majored in physics and economics from Cornell University in 2011, became involved in this field through a chance development.
During his studies, he attended a lecture by a professor who showed people silent clips of election candidates. They were able to accurately predict the winners by observing each candidate’s attitude and charisma.
Dr Ong was intrigued by how such aspects of human behaviour could be studied scientifically. After completing his undergraduate degree, he went on to obtain a doctorate in cognitive psychology and a master’s degree in computer science from Stanford University in 2017.
“This might be an unusual choice for an A*STAR scholar,” he says. “But when I was applying for my PhD in 2011, I realised that as AI becomes a bigger part of our lives, more researchers who understand both psychology and AI would be needed to design better, more human-like AI.”
He also credits the National Science Scholarship (BS-PhD) — which he was awarded as both an undergraduate and a doctoral student — for setting the stage.
“During my first summer vacation, I did an eight-week research attachment with A*STAR, and gained a better understanding of research and the skills required.” Dr Ong is especially keen to correct the misconception that psychology is not related to science.
“Psychology is not just about marketing campaigns or dealing with clinical disorders,” he says. “There is so much more to it.”
To help computers “understand” human emotions, thoughts, facial expressions and gestures, Dr Ong uses equations and algorithms, and has to design scientific experiments.
“My job involves formulating precise research questions, designing experiments that can answer those questions, and building computational models that translate the answers into a language that computers can understand.
Giving back to society
While the word “robots” often brings to mind Hollywood humanoids and online “chatbots”, Dr Ong’s work is far more practical.
One of his projects is Tomo, a robot “pet” that runs away or approaches, depending on how it is called, which helps instil correct behaviour in children.
His long-term goal is to develop robots that can help with mental health issues or learning disabilities, such as healthcare robots that can detect pain and loneliness in patients, or emotionally-aware robot tutors to help improve learning.
His wife jokingly says that Dr Ong wants to build a “Baymax for emotions”, referring to the tirelessly loyal healthcare robot from the Disney animated movie Big Hero 6.
Dr Ong agrees. “I want to build emotionally-aware AI to help people achieve emotional well-being, and hence lead better lives.”
This is why he encourages students to explore social sciences like psychology, sociology and linguistics because they play a critical role in understanding what makes communities tick.
“Social scientists also play a crucial role in doing research that benefi ts society,” he says.
“Within A*STAR, we have a growing group of social scientists who are interested in tackling social issues using computational approaches.
“I am grateful that I can contribute to society as a computational social scientist.”