The latest faculty arrival to the UBC Department of Computer Science is Dr. Mathias Lécuyer, who heralds from the area of privacy and security research. And we're telling the world.
Dr. Lécuyer moved to Vancouver and started his new role this summer as Assistant Professor at the University of British Columbia. It had been a 2-year deferral, thanks (no thanks) to COVID and like so many others, Dr. Lécuyer started his new job in the middle of a pandemic. Yet, he is already fully immersed in making a difference within the department.
Mathias has joined the Systopia Lab and is supervising students as they pursue research on applications that provide rigorous guarantees of robustness, privacy, and security. Mathias explains his endeavours, “My research focuses both on improving practical and theoretical tools like differential privacy, causal inference and reinforcement learning. This is all related to Machine Learning attacks/defenses, privacy preserving data management, and system decisions optimization.”
From the Big Apple to Lotusland
Although Mathias grew up in the suburbs of Paris, he most recently was in New York, where he was a postdoc researcher at Microsoft Research. Prior to that, he completed his PhD and Master of Science in Computer Science at Columbia University, New York City.
Mathias is currently pursuing two different avenues within his research. “One is privacy-preserving machine learning models,” Mathias explains. “I focus on how to better train them, and investigate what infrastructure might be needed to make it easier for industry and government to use ML tools more effectively.” he says. “The second avenue relates to causal inference in systems. This is about uncovering how we can leverage certain system knowledge, and build it into causal inference models. From that, we can explore how to better understand and optimize the systems.”
Mathias says the crux of the enjoyment he gains from his research stems from the more detailed parts of learnings that relate to new discoveries and those ‘eureka’ moments. “I like learning new things in depth and connecting them in interesting ways, such as theoretical parts leading to practical applications. I enjoy discovering a new connection or understanding a new detail that moves a project or goal forward in a significant way,” he explains.
Mathias points to one of his papers from the 2019 IEEE Symposium on Security and Privacy as an example of the type of research he conducts, which defends against adversarial examples with provable guarantees:
Certified Robustness to Adversarial Examples with Differential Privacy Mathias Lécuyer, Vaggelis Atlidakis, Roxana Geambasu, Daniel Hsu, Suman Jana (S&P'19) [Code]
In terms of working in the UBC Computer Science department, Mathias says he is excited to work and collaborate on a broad range of topics. “It’s a fantastic opportunity to be able to work with the really great, and accomplished faculty and students in this department. It’s exciting to know we have such a diverse and impressive level of expertise across so many areas of computer science.”
In his downtime, Mathias is either immersed in a book, a cookbook, or he’s in the garden trying to grow vegetables. Walks and time on the beach with his family as they all get more familiarized with Vancouver, is also high on Mathias’ list.
It’s no secret: the privacy and security research area of UBC CS is becoming more prominent all the time. A very warm welcome, Mathias.