Dr. Kira Radinsky.
(photo credit: COURTESY TECHNION)
Kira Radinsky is only 31, but she is already a well-known breakthrough researcher in the field of artificial intelligence (AI). In 2013, when she was 27 years old, she was chosen by MIT to be included in their annual “35 Innovators Under 35” list.
She currently works as the data science director at eBay. She also works as a researcher at the Technion-Israel Institute of Technology in Haifa, where she studied for all of her degrees. In June, Radinsky was appointed by Finance Minister Moshe Kahlon as a member of the Israel Securities Authority’s board of directors. She also serves as director and adviser to various corporations.
It all began for her in the years 2009 to 2012, when as part of her doctorate – and together with Eric Horvitz, who serves as technical fellow and managing director at Microsoft and was guiding her in her short internship at the company – she developed a Big Data system that tries to predict future occurrences using available data and algorithm-based analyses.
“The initial development of the system,” Radinsky told The Jerusalem Post
, “was based on news databases from 1850 until today, and on top of that we added extra sources from social media, search engines and encyclopedias.”
The system attracted attention after some successful predictions. Two of those are predictions of extreme violence in Sudan around the rising gas prices in the country in 2013 and an outbreak of cholera in Cuba, the first epidemic outbreak of the disease in the country in 130 years.
The system recognizes patterns, as Radinsky explains, and identifies when these patterns repeat themselves.
Another example Radinsky gave in a past interview is from the time when the UN asked to work with the software to predict probabilities of future genocides. The software managed to detect that when a senior leader calls a minority “microbes, mice, insects”, i.e., “creatures that are not recognized as mammals,” the probability that three years later there will be a genocide in that same place rises significantly.
The process is in fact teaching a machine to “learn” from past experience. In other words, the system doesn’t only gather information in great quantities over a vast stretch of time, but is also designed to draw conclusions and predictions for future events based on available databases.
At eBay she uses AI technology to “create a better experience for our customers,” as she explained. “The ability to build a system based on artificial intelligence is to create an understanding of the product that is more intact and whole.”
She is also involved with a number of pro bono activities. One such activity is her position at the Technion, where she uses artificial intelligence technology with a special focus on the medical field.
“The researcher’s role will have to change,” she explains about the future of AI. “Instead of having a doctor who will learn by heart, he or she will have systems that will hypothesize so that the doctor will be able to think of better ways to treat patients.”
In the medical field, a system like that can help predict a person’s future ailments and health problems and perhaps, by early diagnosis or even pre-diagnosis, prevent future problems from developing.
This is expected to be achieved by accumulating medical data and developing a system able to analyze the data and reach conclusions, trying to apply existing information from previous cases to a specific patient.
Asked of the potential dangers of this kind of technological development, Radinsky admits that in general, assessing probabilities to predict future events may entail dangers, but says that “for this there is a need of regulations,” adding, however, “We shouldn’t avoid certain technological developments just because there is a problem of assimilating them in society.”