François Laviolette

François Laviolette (October 6, 1962 – December 26, 2021), Canada-CIFAR Chair in Artificial Intelligence (AI) and an NSERC/Intact Financial Corporation Industrial Research Chair, was an eminent researcher who made major contributions to the theoretical foundations of AI and its application in various contexts. He first distinguished himself by solving a 60-year-old graph theory problem posed by mathematician Paul Erdös, on the decomposition of infinite graphs. His doctoral dissertation was among the seven finalists at the 1998 Council of Graduate Schools/University Microfilms International Distinguished Dissertation Award of Washington. When he joined the Department of Computer Science at Université Laval in 2002, he began to shift his attention toward machine learning. A mere 10 years later, he had already made his mark through his contributions to the PAC-Bayesian bounds that determine the reliability of a learning algorithm, and to the analysis of genome sequencing/assembly into fragments. Using a graph structure to represent fragments, this developed algorithm could undergo massively parallel processing—a first at the time—which reduced the de novo human genome assembly time from three weeks to about 10 hours!

During his career, he wrote over 100 articles that have been cited more than 9,600 times. Among his most notable contributions to the foundations of AI is DANN, a domain-adaption learning algorithm that involves transferring learning achieved in one context to a different domain. He also made significant contributions to health research (including three articles in Nature journal Scientific Reports), insurance and aeronautics (as Quebec’s Head Scientist on the DEEL project).

While these are all of major importance, his contributions go beyond scientific advancement. Mr. Laviolette was a master builder of major projects. He was able to motivate, convince and collaborate with people. He created the Big Data Research Centre at Université Laval (CRDM_UL), bringing together over 60 researchers from different fields. Mr. Laviolette was committed to the ethical aspects of AI and its social acceptability. He participated in the development of the Montréal Declaration for a Responsible Development of Artificial Intelligence whose signatories commit to the responsible and ethical use of AI and digital technologies. He was a founder of the International Observatory on the Societal Impacts of AI and Digital Technologies (OBVIA), as well as the Institute Intelligence and Data (IID) at Université Laval. After being diagnosed with cancer, he acted as a patient-partner and expert in AI in several groups, such as Imagia’s DHDP project, Catalis, and the AI Health Committee led by Quebec’s Chief Scientist, as he believed that making health data available for research purposes will be a key factor in developing a healthy society.

Mr. Laviolette encouraged the creation and growth of an AI ecosystem in Quebec. Its industrial chair, major projects and the events organized by the CRDM_UL allow researchers, industries and companies to collaborate on emerging and traditional issues through AI. The DEEL (Dependable Explanation Learning) project, which aims to make AI certifiable as well as more robust and dependable, is another example of how Canadian and French researchers are collaborating with companies such as Thales, Bell, Bombardier and CAE.

Very recently, a group of colleagues and friends created the François Laviolette Scholarship Fund to support the work of graduate students, particularly on interpretability and the ethical and responsible use of data. Interpretability is essential in certain sectors, such as health and insurance, where a machine cannot be expected to make decisions on critical actions without it being able to provide a justification that humans can understand. He saw the Fund as a catalyst that could help create a responsible and transparent collective choice, connecting citizens, decision makers and organizations working with data and allowing data to be viewed as a natural resource that serves the good of all, thus initiating, perhaps, a new Quiet Revolution.

François Laviolette positioned Canada as a leader in the pivotal transformations that AI is bringing to our lives. Through his kindness and ability to bring people together, he created a friendly environment for all his colleagues.