Dr. Anna Goldenberg is fully aware of public misperceptions of artificial intelligence, a field often conjured straight out of science fiction as a power-mad computer or an evil robot.
“I’ve never worked in robotics,” Goldenberg said, “and I’ve been in this field a long time.”
Her description of her new job is somewhat less heady: “I work specifically on developing methodology, which is really a combination of algorithms and statistics, to identify patterns in data and to try to make sense of data that’s too complex to make sense of by just human cognitive capacity.”
But don’t be fooled: Goldenberg’s new position is cutting-edge and pioneering.
On Jan. 15, Toronto’s renowned Hospital for Sick Children announced the appointment of Goldenberg as its first chair in biomedical informatics and artificial intelligence. It’s believed that SickKids is the first children’s hospital in North America to have such a position.
The post is partly funded by a $1.75-million donation from Amar Varma, a Toronto engineer and entrepreneur whose newborn son had surgery at SickKids six years ago. The hospital’s foundation will match the gift.
Goldenberg, 40, works in the relatively esoteric field of collecting data and using algorithms to improve health care outcomes across a variety of diseases and conditions.
“There are lots of variables we are trying to make sense of at the same time,” she told The CJN. “It’s a beautiful field with a lot of potential that can solve challenging problems in health care.”
Essentially, Goldenberg and her team try to change the old trial-and-error model to a more reliably predictive one. The idea is to gauge how likely it is that a patient will respond to a certain treatment, and the best time to administer it.
Artificial intelligence would also allow doctors to predict when and where in the body some malignancies could develop or recur in patients predisposed to cancer.
“We try to make the prediction to figure out which treatment to give, when to stop it and start a new one,” Goldenberg explained. “So we are trying to make (treatment) a lot more rigorous.”
The idea is that mining big data – in the case of SickKids, about two trillion data points collected since 2013 – will allow personalized predictions and, it’s hoped, prevent disease.
The field has applications to heart disease, cancer, mental health and chronic conditions.
SickKids has already built machine-learning models to predict cardiac arrest in its pediatric intensive care unit and the need for thyroid biopsies – all with a view toward prevention.
Born in the southwestern Russian town of Voronezh, Goldenberg said she did not know she was Jewish until early in grade school. Then, she found out the hard way.
“The kids started screaming, ‘Jew, go back to Palestine.’ I didn’t understand what they’re saying. My mother sat me down and said, ‘Yes, we are Jewish. This is how it is.’ ”
The family left Russia as Jewish refugees in 1995 and arrived in Kentucky to join family. Goldenberg completed her graduate work at Carnegie Mellon University in Pittsburgh, where she got to know Joyce Feinberg, one of the 11 victims killed in the massacre at Pittsburgh’s Tree of Life synagogue in October.
“We were very close,” Goldenberg recalled, noting that she had celebrated the High Holidays with Feinberg and her family. “Her death was really hard for me.”
Goldenberg came to Canada in 2008 as a post-doctoral fellow and held a series of positions at SickKids, the University of Toronto and the Vector Institute for Artificial Intelligence.
“My Jewish identity has always been very important to me,” she said. “I was told that I was Jewish in a very unpleasant way and I have realized it was a long journey for me to understand that I am Jewish, no matter what. This is part of my identity.”
As for work, Goldenberg said, “I love my job. It’s really fun.”