After 25 years abroad, researcher Patrick Charbonneau returns to his native Montreal thanks to an Eddie Goldenberg Canada Research Chair with a budget of eight million dollars over eight years that he obtained at the University of Montreal. Formerly grouped together in the Impact+ program, these chairs were created to attract exceptional researchers to Canada who are pursuing careers outside the country.
With his family, he says he is very excited to return to Montreal – he grew up in Ahuntsic and his parents still live there – while the expertise that bridges chemistry and physics acquired during his international career is now highly sought after with the creation of the Courtois Institute.
Understanding generative AI and improving it
One of the major challenges on which his chair will work: understanding artificial intelligence (AI) using notions of statistical physics.
“One of the big problems is that we don’t really understand why generative AI models work so well,” explains the 47-year-old researcher. For example, we don't understand their limitations. By knowing better how an architecture works, we will know when to expect it to fail.”
His proposal is therefore to take simple models very well known in statistical physics to use them as tests. “By trying the generative approach to provide answers to the problems submitted, we will immediately see if the AI makes a mistake and why,” adds the man who did his doctorate at Harvard University. The objective is to push these tools to their limits in order to then improve them and be able to use them for more complex problems.”
What does this have to do with materials? “In statistical physics, we have little understanding of frustrated matter,” says the man who has made it a specialty. A bit like when we look at relationships between people, there are materials that have unfavorable interactions that influence the organization of matter. It is central in statistical physics to understand how these systems work and, at the moment, the tools we have to do this are not very efficient.
To meet this challenge, it will hire people from here and abroad to work in collaboration with Mila – the Quebec Institute of Artificial Intelligence –, Samsung AI Lab Montreal and the materials industry via PRIMA Quebec.
Collaborator of a Nobel Prize winner
La réputation de Patrick Charbonneau n’est plus à faire dans son domaine. Celui qui a obtenu un poste à l’Université Duke, en Caroline du Nord, en 2008 après un stage postdoctoral à Amsterdam a même été cité deux fois par le comité du prix Nobel qui a primé le physicien italien Giorgio Parisi en 2021.
“A specialist in complex systems since the late 1970s, he has designed very influential analysis tools in different fields of study, including mathematics, computer science, physics and materials,” says the researcher. He works with physicist Francesco Zamponi, with whom I have collaborated since my early years at Duke University, and we have combined our expertise to look at glasses and other disordered materials. We developed a numerical approach that quickly resolved several issues relating to blocking transition and glass dynamics.”
A fascination with materials
The man who was part of the Petits Chanteurs du Mont-Royal and obtained a college diploma in science, letters and arts has always been interested in all kinds of subjects. He thought he would study history until he discovered a fascination with materials. “I wanted to understand how matter was put together,” he remembers. Then, I had ease in science and mathematics.”
Patrick Charbonneau even won a competition from the Quebec Mathematical Association, which allowed him to participate in a summer camp at UdeM where different members of the teaching team gave mini-courses. He already saw himself becoming a university professor.
Having taken part in English summer camps as a child, he chose to study at McGill University because it offered him the best scholarship. He majored with honors in chemistry and minored in computer science.
Having held various summer jobs – he even shoveled manure on a pig farm – during his undergraduate studies he wanted to work on research projects. He received a grant from the Natural Sciences and Engineering Research Council of Canada to work in theoretical chemistry at McGill University, and the following summer he was hired by Henry F. Schaefer III at the University of Georgia in Athens (UGA) with funding from the National Science Foundation.
“In both cases, I was published,” he explains.
Patrick Charbonneau then saw the doors of prestigious American universities open. “It was the first of three refusals I had from the University of California at Berkeley,” he says with self-deprecation. On the other hand, I was accepted to MIT [Massachusetts Institute of Technology], Harvard and UGA. I drove with my father on a Thursday evening to visit the campuses in Boston and I chose Harvard.
Various research objects
The one who now gives courses mainly in physics, even if his training is in chemistry, also has a little frosty side. He has published a series of scientific articles in recent years on the history of confectionery, more particularly cream sugar.
The idea came to him when he was asked to write a book chapter on molecular gastronomy. Since he was teaching a cooking science class in which he really enjoyed doing a lab on cream sugar to illustrate fundamental concepts in physical chemistry, he suggested this topic.
“But it’s a little-known recipe around the world, since it’s a particularity of the French-speaking Northeast,” he emphasizes. So I had to explain what it is. I thought it would be easy to find the right reference to cite. Then I realized that a lot of people had come up with hypotheses about the origin of cream sugar, but none were supported by historical evidence.
He told himself he was going to take a few weeks to look into the matter. It took him six years to complete his research and publish the results.
Values that are irreconcilable with the American social climate
If he is fully fulfilled in his work, it was he who raised his hand to return to Montreal when he saw the federal research chairs program. This is not unrelated to the current social climate in the United States. He first clarifies that he has never had any particular fascination for this country.
“I went there for studies and Harvard is an extraordinary academic environment in terms of a platform to meet people,” he observes. Then, I loved the whole democratic ideal of the United States which is reflected in the importance we give to education, to public libraries, to the public information service including NPR and PBS, which are of incredible quality. All these things that I like and which are part of the American intellectual identity are systematically attacked by the current administration.
With his Quebec wife, he has two young children and the couple is increasingly sensitive to the changes occurring south of the border. Among other things, he noticed that the many children and their Latin American parents who normally wait for the school bus almost all disappeared at the start of the year.
“This anecdote clearly illustrates the discomfort we can feel,” he says, moved. This is not right what is happening. Do we want to witness this? And critics will say that anyone who doesn't act is complicit. But how to act without getting into too many problems? We vote in every election. Still, it’s hard to be in the middle of all this.”
It is therefore with great enthusiasm that the little family will settle in Montreal in June 2027. “I love public transport, so we commissioned a real estate agent to find us a house near a metro station,” he points out.
It is also, for him, a great turnaround in his career. Hoping to work at UdeM, he contacted Professor Normand Mousseau towards the end of his postdoctoral fellowship. “He told me that I would not be hired in physics even though I had been trained in chemistry,” he remembers. Twenty years later, I contacted him again and he told me that it was exactly my profile that we were looking for to bridge the two fields at the Courtois Institute, which specializes in materials. Today, it is probably one of the best places in the world to do my work. I am incredibly lucky.”