����ֱ��

Math & Stats Department Colloquium

Friday, March 22, 2024

Time: 2:30 p.m.  Place: Jeffery Hall, Room 234

Speaker: Xiao-Li Meng (Harvard)

Title: Multi-resolution Meandering: Personalized Treatments, Individual Privacy, Machine Unlearning, and aWorld without Randomness

Abstract: Data science revolutionizes the granularity of human inquiries and even offers the promise of personalized assessments. However, how can we assess individual treatment effect before treating the individual? Transitional Inference addresses this dilemma through the concept of “transfer to the similar,” a notion that has been pondered by philosophers since Galen of the Roman Empire. This talk presents a Multi-Resolution Framework (Li and Meng, 2021, JASA) for transitional inference, where similarity is prescribed probabilistically by concomitantly specifying the sameness — the shared distributional form — and the differences — the individual realizations. This framework avoids the concept of randomness and defines “individual probability” as a deterministic limit with infinite resolution. These conceptualizations help us operationalize the meaning of personalized treatments, clarify what individual privacy is protected by differential privacy, and anticipate the challenges of preserving an individual’s right to be forgotten through machine unlearning. Furthermore, it reveals a world that is resistant to overfitting when the resolutions of our data and (deep) learning far exceed the resolution necessary for pattern recognition.

Bio: Xiao-Li Meng, the Founding Editor-in-Chief of Harvard Data Science Review and the Whipple V. N. Jones Professor of Statistics at Harvard University, is well known for his depth and breadth in research, his innovation and passion in pedagogy, his vision and effectiveness in administration, as well as for his engaging and entertaining style as a speaker and writer. Meng was named the best statistician under the age of 40 by the Committee of Presidents of Statistical Societies (COPSS) in 2001, and he is the recipient of numerous awards and honours for his more than 150 publications. In 2020, he was elected to the American Academy of Arts and Sciences. Meng received his BS in mathematics from Fudan University in 1982 and his PhD in statistics from Harvard in 1990. He was on the faculty of the University of Chicago from 1991 to 2001 before returning to Harvard, where he served as the Chair of the Department of Statistics (2004–2012) and the Dean of the Graduate School of Arts and Sciences (2012–2017).