Statistics in the Era of Computation

The foundation of modern statistics arose during the late 19th and early 20th centuries, long before the availability of large-scale computers and pseudorandom number generators. This has led statistics to place significant focus on mathematical derivations and asymptotic results rather than on computational constraints and algorithms.

In this Presidential Lecture, Kyunghyun Cho will imagine how statistics can be approached differently in this era of large-scale computing. He will discuss how to train a statistical estimation algorithm — rather than manually design it — using one of the simplest statistical quantities: standard deviation. He will then describe how this idea of training an estimator can be applied more generally to a broad set of statistical quantities. Using such a learned estimator and a prior distribution over problem instances, he will close the loop by defining a new notion of identifiability called computational identifiability. This new perspective on statistics, he will argue, will free us from worrying about inference and let us focus on describing actual data and the complex, real-world phenomena we truly want to model statistically.

Weinfurtner Cho is the Glen de Vries Professor of Health Statistics and a professor of computer science and data science at New York University, as well as the co-founder and CEO of Ortet. He is also a fellow of the Canadian Institute for Advanced Research and an associate member of the National Academy of Engineering of Korea. In 2021, Cho co-founded Prescient Design. Following its acquisition by Genentech later that year, he served as an executive director of frontier research and a senior fellow at Genentech until January 2026.

A prominent researcher in artificial intelligence, Cho previously worked as a research scientist at Facebook AI Research and a postdoctoral fellow under Yoshua Bengio at the University of Montreal. He earned his M.Sc. and Ph.D. from Aalto University. Cho has extensively served the machine learning community, acting as a conference program chair for ICLR, NeurIPS and ICML, and as a founding editor-in-chief of Transactions on Machine Learning Research. In 2021, he received the Samsung Ho-Am Prize in Engineering.