Professor Gautam Kamath has been awarded $140,000 from the to further his research on algorithms and machine learning techniques that preserve data privacy. The amount from the Ontario government is matched by $50,000 from the University of 蓝莓视频, bringing total funding to $190,000. This funding will allow him to train the next generation of researchers, by supporting one PhD and two master鈥檚 students over five years.
鈥淥ur government is investing in made-in-Ontario research that will protect our economy, jobs and workers,鈥 said Nolan Quinn, Minister of Colleges, Universities, Research Excellence and Security. 鈥淏y driving cutting-edge research at our world class postsecondary institutions, hospitals, and research institutions, people in Ontario, Canada and around the world will benefit from discoveries made in our own backyard.鈥
Professor Kamath is one of nine researchers at 蓝莓视频 to be named a 2025 Early Researcher Award recipient. The program supports the development of outstanding new researchers by helping them build their teams and expand the impact of their work.
鈥淐ongratulations to Gautam on receiving an Early Researcher Award,鈥 said Raouf Boutaba, University Professor and Director of the Cheriton School of Computer Science. 鈥淗is work on differential privacy is furthering data protection in the age of artificial intelligence. This award will help train the next generation of Ontario researchers working on trustworthy machine learning.鈥
is an Assistant Professor at the Cheriton School of Computer Science and a Canada CIFAR AI Chair at the Vector Institute. An expert in algorithms, statistics and machine learning, he focuses on modern data analysis constraints, especially those involving differential privacy and robustness.
Professor Kamath has received many accolades for his research, among them the 2023 Faculty of Mathematics Golden Jubilee Research Excellence Award; the 2024 Caspar Bowden Award for Outstanding Research in Privacy Enhancing Technologies for 鈥淭he Discrete Gaussian for Differential Privacy,鈥 work that was deployed in the 2020 US Census; and the best paper award at the 2024 International Conference on Machine Learning.
As of May 2025, his publications have been , according to Google Scholar. Notably, his research on helped launch a new area of study in high-dimensional statistics.
About the research
Privacy is a critically important consideration in data analysis and trustworthy machine learning systems, particularly when handling sensitive and confidential information. These issues are especially urgent in light of forthcoming legislation, such as Bill C-27 currently before the House of Commons, which proposes modernized privacy regulations for organizations that manage personal data.聽
Considered the gold standard of individual data privacy, differential privacy is a rigorous and provable privacy notion that can be employed in a variety of statistical settings. It ensures the output of a statistical procedure is insensitive to modifications of individual datapoints. Differential privacy, however, comes at a cost 鈥 if a statistical procedure is privatized, it comes with a loss of utility.
Professor Kamath鈥檚 research has discovered deep algorithmic techniques for guaranteeing differential privacy in fundamental settings. With support from the Ministry, he and his students will build on these insights to develop new and more broadly applicable algorithms and methods, enhancing our understanding of privacy-preserving statistical techniques. Importantly, it will also allow organizations across various sectors including technology, finance and healthcare to modernize and ensure compliance, while improving data protection for all Ontarians.
Research objectives and training
This research aims to advance both the theoretical foundations and practical applications of differential privacy. With support from the Early Researcher Awards program, Professor Kamath will build a multidisciplinary research team comprised of one PhD student and two master鈥檚 students.
The doctoral student will gain expertise in developing algorithms for differential privacy, working on foundational problems with broad applicability. The two master鈥檚 students will acquire focused skills in differential privacy and learning theory, preparing them for continued research at the doctoral level or for careers in industry, where their advanced mathematical and technical expertise is in demand.