Professor Shai Ben-David has been appointed a University Research Chair in recognition of his outstanding contributions to machine learning theory, logic, the theory of distributed computation and complexity theory. This prestigious title may be held for up to seven years and is conferred to recognize the exceptional achievements of Waterloo faculty members and to acknowledge their pre-eminence in a field of knowledge.
“Shai is respected internationally for his research on machine learning theory,” said Mark Giesbrecht, Director of the Cheriton School of Computer Science. “Among his notable contributions are the pioneering steps he developed to analyze domain adaptation, learnability of real valued functions and change detection in streaming data. More recently, Shai has made groundbreaking contributions to unsupervised learning, clustering and the mathematical underpinnings of statistical machine learning.”
The most fundamental question in unsupervised machine learning concerns what reliable conclusions about an unknown data distribution can be made from finite samples drawn from that distribution. The two major directions of research on this topic are the complexity of finding a good approximation to the unknown distribution under some prior knowledge about that distribution, and what questions about the distribution can be answered without any such assumptions. In both areas, Professor Ben-David has conducted internationally recognized research.
To address the first question, Professor Ben-David developed a powerful novel modular tool for learning mixtures of distributions. He then applied the tool with his former PhD student Hassan Ashtiani, now an Assistant Professor at McMaster University, along with international colleagues Christopher Liaw, Abbas Mehrabian and Yaniv Plan, to fully resolve a long-standing open problem of the sample complexity of learning mixtures of Gaussian distributions, work that received a Best Paper Award at NeurIPS 2018, the pre-eminent and largest annual conference on machine learning.
To address the second line of research on unsupervised learning, Professor Ben-David and his international colleagues Pavel Hrubeš, Shay Moran, Amir Shpilka and Amir Yehudayoff made a fundamental discovery that no combinatorial dimension can characterize the learnability in Vapnik’s general model of statistical learning. This mathematically surprising and exceptional result was published in Nature Machine Intelligence and featured prominently as a front-page news and views article on Nature’s website.
In 2015, Professor Ben-David and his colleague Shai Shalev-Shwartz coauthored a book titled Understanding Machine Learning: From Theory to Algorithms, which provides a systematic exposition of the principles of modern machine learning and has become a standard course textbook around the world. In addition, his Foundations of Machine Learning channel on YouTube reaches tens of thousands of subscribers and viewers internationally.
Professor Ben-David earned his PhD in mathematics from the Hebrew University in Jerusalem and his first faculty appointment was as a Professor of Computer Science at the Technion in Haifa, Israel. He joined the Cheriton School of Computer Science in 2004. He has held visiting faculty positions at the Australian National University, Cornell University, ETH Zurich, TTI Chicago and at the Simons Institute for the Theory of Computing at the University of California, Berkeley.
Professor Ben-David is the ninth faculty member at the Cheriton School of Computer Science to be appointed a University Research Chair. His appointment to a seven-year term is effective July 1, 2020.
Other faculty members at the Cheriton School of Computer Science who have been appointed University Research Chairs are Professor Kate Larson (2019), Professor Raouf Boutaba (2018), Professor Lila Kari (2015), Professor Ian Goldberg (2012), Professor Bin Ma (2008), Professor Timothy Chan (2005), Professor Doug Stinson (2005), and Professor Tamer Özsu (2004).