by Nora Azouz
As announced on social media networks by Lyse Langlois, Executive Director of the International Observatory on the Societal Impacts of AI and Digital Technologies (OBVIA), the death of François Laviolette, a researcher in Artificial Intelligence, will leave a void that will be difficult to fill in the research community.
“Rest in peace my dear friend, we will continue the work,” are the few words left on Twitter by Lyse Langlois to announce the death of his colleague at OBVIA on December 26.
According to OBVIA’s executive director, François Laviolette had “great capacities for popularization”. He was an excellent teacher and generous with his time,” she says. “He worked until the end of his life, meeting with many researchers and always having a desire to help.”
“When he was on a committee, participating researchers could freely ask questions on his favorite subjects. He did not judge anyone. He gave a sincere point of view by maintaining a dialogue.” – Lyse Langlois, Executive Director of the OBVIA
A MAN OF COMMITMENT
François Laviolette’s imprint will certainly remain indelible. He published more than 180 scientific articles over the course of his career, including a hundred or so in the last ten years.
He was also involved in the development of the Montreal Declaration, hence his interest in charters and ethical standards that serve to better frame the field of Artificial Intelligence and AI for the common good.
Professor Laviolette was part of all initiatives promoting the ethical use of Artificial Intelligence and machine learning. He was Director of the Centre de données massives de l’Université Laval (CRDM); Co-director of the Groupe de recherche en apprentissage automatique de l’Université Laval (GRAAL), which focuses mainly on the theory and applications of machine learning; collaborating researcher at the Intelligence and Data Institute of Laval University and associate member of the Institute of Applied Ethics (IDEA), a multidisciplinary group of professors, researchers and students conducting research in the field of ethics; and researcher responsible for the AI ecosystem at the International Observatory on the Societal Impacts of AI and Digital Technologies (OBVIA).
A FRANÇOIS LAVIOLETTE SCHOLARSHIP
To thank him for his involvement, Professor Laviolette’s colleagues organized a tribute event in Quebec City on November 28, in the presence of several key people in the Artificial Intelligence sector.
“It was also during this tribute that we launched the François-Laviolette Scholarship to continue his work,” says Lyse Langlois. He was very touched by these gestures of tribute and scholarships and warmly thanked everyone participating, whether in person or virtually.”
$120,000 has been raised to date in donations to establish scholarships for graduate students in Francois Laviolette’s fields of interest. The scholarships aim to support interdisciplinary work on the ethics of AI and on the interpretability of algorithms. The fellowships will be launched next spring.
FROM RESEARCH TO CONCRETE APPLICATIONS
François Laviolette was a full professor in the Department of Computer Science and Software Engineering at Université Laval. His research on Artificial Intelligence has contributed to the development of PAC-Bayesian theory, a branch of learning theory aimed at better understanding machine learning algorithms. He was particularly interested in algorithms related to genomics, proteomics (the science that studies all the proteins in a cell) and drug discovery.
A prolific researcher, his fields of investigation continued to expand over the past twenty years. In addition to AI and learning theory, he was a specialist in graph theory, automated verification of bioinformatics and kernel methods. In the last few years, he was particularly interested in the field of health and the potential of AI for the benefit of patients.
BIG DATA FOR INDUSTRY
The insurance and health sectors as well as environmental metadata have directly benefited from the advances resulting from the work of François Laviolette and his teams.
His work on machine learning algorithms and the exploitation of big data is helping the insurance industry to turn to big data.
He participated in the development of the report, published in January 2020, “Artificial Intelligence, Solidarity and Insurance in Europe and Canada”, which proposes good practices related to AI in the insurance sector.
Also, as a co-investigator at the MDRC, he took part in a partnership with Two Hat Security to improve the conversation analysis algorithms and toxicity metrics of its Community Sift software. This partnership should lead to the creation of algorithms that will help study the nature of toxic behaviors, their impact on users and to propose mechanisms to combat them.
Another major research project in which he was involved is aimed at refining quality control and diagnostics in medicine through the use of very high throughput mass spectrometry (identification of molecules by measuring their mass).
The North Sentinel research program aims to determine the health of northern ecosystems by analyzing key microbial markers in real time. It is based on the EcoChip, an in situ microbial culture tool that assesses the growth of microorganisms in their natural environment while measuring environmental metadata in real time.
Using analytical approaches based on genomics, bioinformatics, localization, and machine learning, the team of researchers of which Professor Laviolette was a member is working to make better drugs by improving the process of searching for molecular structures using the building blocks of amino acids. The resulting models could create drugs against microbes. Other areas such as cancer, neurology, infectious and autoimmune diseases could benefit from these discoveries.
All of these exemplary works are only a part of the considerable achievements of this passionate man. More than anything else, it was important for him to support graduate students on work devoted to the ethics of AI and the interpretability of algorithms.
Lyse Langlois concluded: “In my heart, he is irreplaceable because he had this ability to rally people; he leaves an important legacy that will allow other researchers to continue his work by bringing their own color.”
We share the sadness of François Laviolette’s family and send them our sincere condolences.