Israeli MIT Prof. Regina Barzilay first recipient of $1 m. AI award

A graduate of BGU, Brazilay is lauded in the field of AI and machine learning for creating an innovative breast-cancer predictive model and proving machines can decipher ancient languages.

MIT Professor Regina Barzilay (photo credit: MIT)
MIT Professor Regina Barzilay
(photo credit: MIT)
Israeli MIT Prof. Regina Barzilay is the first recipient of the $1 million Association for the Advancement of Artificial Intelligence (AAAI) award, MIT reported in a press release on Monday night.
Barzilay will be presented with the award in February during the 35th annual AAAI conference.
Meant to place a spotlight on the field of AI similar to the way the Turing Award focuses public attention on computing, the prize money was made possible due to the generous support of Squirrel AI, a Chinese educational AI company that developed a teaching system that offers to teach students how to get better score results, outperforming the success of human teachers.
“One of the misconceptions people have about AI is that it is allegedly everywhere,” Barzilay told The Jerusalem Post. “But the truth is that AI is not uniformly distributed across all the areas of our life the way it impacted e-commerce.” This is the AI that offers you a new product you might like to buy or a new television series to watch. “But when was the last time you went to a doctor’s office and an AI treated you?”
Barzilay developed an AI model that can detect breast cancer among women thanks to the unique quality of machine learning. The AI can examine hundreds of thousands of mammography results and predict that some of the white spots the scan reveals are likely to develop into a tumor. This is crucial as early detection and prevention are vital to save women’s lives.
As in all predictions, when asked to predict for the near future (one year), the AI reaches a stunning level of accuracy – roughly 70%. When asked to offer a prediction into the distant future – three or five years from now – the success rate declines, as there are many different factors to consider, such as individual health choices. Barzilay pointed out that this is also true for human predictions, not just AI.
“It’s difficult for medical doctors to accept the prediction of the AI because they don’t fully understand it and because they bear all the responsibility,” she said. “If the AI offers you a product and you don’t like, you can return it – but in health, the cost of a mistake is huge.”
“It’s not a lack of AI technology,” she said. “It’s a lack in understanding how to translate it into a clinical setting; it’s an issue of psychology and how the dynamics of an organization works. I hope that this award will be a stepping-stone to promote the technology further.”
She offered the example that AI could offer a prediction of which COVID-19 patients will need tube ventilation and which could be sent home to recover, saving valuable resources for any national health service.
BARZILAY IS lauded for her success in creating an AI that was able to “read” vast digital libraries and suggest a unique combination of chemicals to create a new antibody called halicin. The antibody can kill antibiotic-resistant bacteria, such as Acinetobacter baumannii, which could in turn offer humanity a last line of defense in a future health crisis.
No human could process so much data, retain it and “see” among billions of possibilities that such a structure would have these effects. The machine, if it is given good data and a good algorithm to work with, can see in the data pattern things hidden from human eyes.
To prove that an AI could get good results even if it is faced with a small amount of data, Barzilay developed one that was able to translate Ugaritic, an ancient Semitic language spoken roughly 8,000 years ago. The AI was able to do this by using Hebrew, a language with more available data that shares a proto-Semitic origin with Ugaritic.
“I have so many publications in the field of AI,” she said, laughing. “But this one thing is something I always get asked about. People love deciphering things. Nothing else I did, until now, got so much attention!
“I really loved studying at Ben-Gurion University,” she told the Post, “and attribute all my success to that place. It was a great university for me.”
Her advice to young academics in Israel and elsewhere is “pay close attention to what is important to you, keep your eye on the ball, and don’t pay attention to naysayers.”