Heraklion, Crete, 1976. A ten-year-old girl is fascinated by the world of numbers and dreams of becoming a mathematician when she grows up. Ten years later, she moved to Athens and studied at the School of Mathematics of the Greek Academy of Sciences. But her gaze looks far away, to the opposite coast of the Atlantic. After finishing her studies, she goes to America. She arrives in the state of Rhode Island, she begins graduate studies at Brown University “elective”, and in 1993 she graduates with a prestigious Ph.D. in Applied Mathematics and Business Research. And although her initial goal was to return to Greece, the “land of great opportunities” conquered her and in 1998 she began teaching at the Massachusetts Institute of Technology, better known as MIT.
Having traced a path that deservedly represents the peaks that Greece in the diaspora usually conquers (many times without us, in the “motherland”, knowing), tells us not only about her own academic and research journey, but also of that journey that we have all recently embarked on: the one that takes place in the still unknown waters of artificial intelligence. In addition, she is one of the authorities on the subject, currently holding two critical positions at MIT: director of the Center for Operations Research, but also vice dean of the newly built Schwarzman College of Computing.
“Thank you for your interest. So many years in the United States and it is the first time that the Greek press has taken an interest in my work and role here,” he says at the beginning of our conversation, with the frankness, kindness and humility that distinguishes the words of others. two leading MIT educators whom we were recently lucky enough to meet We recently spoke on behalf of “K”, David Kaiser and Marsha Bartusiak. Is this humility typical of someone who prefers to work in silence, of someone who speaks only with his works? “At MIT we must all be humble, because we know that for each one of us there is always someone much better. Those who have understood this are successful and happy.”
But how does a girl from the Greek countryside end up on the scientific top of the world? What starting points and what stops does an adventure like this have? As is so often the case, simple gestures from a parent or teacher can boost entire careers. “There were two people who inspired me to love mathematics,” he recalls, looking back to his childhood in Crete in the 1970s and 1980s. “One was my father, who using matches taught me as a child to see mathematics as something interesting and fun. The second was my middle and high school math teacher. When I was a teenager, he would spend hours in his office, where he would give me difficult math problems that were a challenge to me, a game. That’s what math was to me: a game I’d love to play for the rest of my life.”
A female role model
Of her early years in the United States, she vividly remembers her teacher: “Stella Dafermou was the first woman to teach Applied Mathematics at Brown. She became a role model for me, but unfortunately, at the end of my third year there, she died of cancer. So I found myself teaching in her department, which she needed an assistant, while she was simultaneously starting new doctoral research with a supervising professor from MIT. This was also my introduction to the Institute: one door closed, but another opened. My beloved teacher left, but her death brought me here, where I have been for 25 years.”
Is the “game” of mathematics that you dreamed of playing all your life still going on at MIT today? She answers in the affirmative, but explains that this is now hidden in complex algorithms.
“MIT has recognized that artificial intelligence should be at the center of its research. That’s why it set up the School of Computer Science.”
So what’s going on in that new College of Computer Science? What needs created it, what research does it carry out? “MIT has recognized that artificial intelligence must now be at the center of its research. That is why it set up the School of Computer Science three years ago, a “horizontal” academic structure that places great emphasis on artificial intelligence research and its interdisciplinary applications. , and that through technology brings together its five faculties, the School of Architecture and Urbanism, the Faculty of Engineering, the Faculty of Sciences, the Faculty of Humanities, Arts and Social Sciences and the Faculty of Administration”.
So academic fields that were traditionally considered distinct, today at MIT are fruitfully married through computer science and artificial intelligence. “In fact, that’s the trend on our show today. For example, our undergraduate students can choose “double” majors (double majors) such as “Decision-making and Artificial Intelligence” but also “blended” majors (blended majors) such as e.g. “Informatics and Molecular Biology”. Also, there is the concept of “common ground” where students take courses (such as Climate Change), which correspond simultaneously to many different MIT schools.”
life with algorithms
When we ask about his students, he speaks warmly. “We are like a family,” he stresses, and when we ask him if there are Greeks among them, he says that of his nine doctoral students, four are European (two Greek, one Spanish, one Romanian) and five from North America (four American women and one Canadian). “All of them, and of course the Greeks, make me proud.” But are these students training for professions that artificial intelligence will make redundant tomorrow? “This is something that has happened before. It is an inevitable part of technological development. But I don’t think (artificial intelligence) can, at least soon, operate successfully autonomously, without our help.
Consider how in your first steps the ChatGPT it was powered by auto-completion technology, like the kind Google uses when we type in keywords and guess what to do with our search. Along the way, however, human participation was deemed necessary for its evolution. In several large projects that we are working on, I see that in the end it is the human that helps artificial intelligence and not the other way around.”
Asking him to share some relevant examples with us gives us a glimpse into the future: “One example comes from the retail space and our collaboration with the famous Spanish clothing company Zara, which has been heavily accused of producing vast amounts of fabric waste. . So we use artificial intelligence to find exactly how many parts need to be produced so that there is no surplus. But it is not enough to analyze market data, it is also necessary to consult local managers. Another example comes from our partnership with a local public health center, UMass Memorial Hospital, which works closely with the University of Massachusetts Medical School.
So we’ve built a system that can, with the help of artificial intelligence, calculate how long a patient will wait in the hospital emergency room, before being accepted for examination and treatment by competent doctors. This, as you will understand, is very important, as some incidents are more urgent than others.” technologytherefore, at the service of the environment and health, but always with our own indications.
At this point, the MIT professor shares something very interesting with us: “Once it occurred to me that, as a woman, when I went to the emergency room, I waited longer than a man. Was this my paranoia or not? The answer was somewhere in between: when the incident was truly urgent, there was no distinction. But if it was middleweight, it existed.” So, algorithms, unlike us, don’t have biases? “They have as many as we give them the right to have,” he replies. “If we give them biased information, they will give us a result It goes on to remind us of the case of the recruitment algorithm created by Amazon in 2018: “Its results were not particularly “fair” for women. The reason was that it was simply an artificial intelligence system trained on a database containing only men”.
In order for the AI to help us properly, we must first help it properly. “It’s a two-way street, as is the case in any rewarding relationship. That’s why our educational goal at MIT has another important pillar that we call “SERC” (Social and Ethical Responsibility of Computing), social and ethical responsibility. It is necessary to train the new generation in the proper use of these tools, otherwise the results can be frightening I cite the example of nuclear power as an example: it can provide energy sufficiency (as is the case in France) but also drive to Hiroshima”.