FEB 15: Health Data Science Distinguished Speaker Series: Dr. Tamara Broderick, MIT

“Toward a Taxonomy of Trust in Data Science Methods”

Feb 15, 2024 | 11:30AM-12:30PM.
Hybrid Event (BUMC & Zoom).
In-person:  Crosstown Center (801 Mass Ave), 2nd Floor, Room 2128.

Registration is now closed.

Abstract:  Probabilistic data analysis increasingly informs critical decisions in medicine, economics, politics, and beyond. A major concern is generalization: if we conclude that an economic or health intervention helps people based on a data analysis, we hope that it will indeed help people when deployed in the future. We detail a taxonomy of where generalization might break down: (i) in the translation of real-world goals to goals on a particular set of training data, (ii) in the translation of abstract goals on the training data to a concrete mathematical problem, (iii) in the use of an algorithm to solve the stated mathematical problem, and (iv) in the use of a particular code implementation of the chosen algorithm. We illustrate potential pitfalls and also describe mitigations at each step. In the second part of the talk, we will focus on one example check: is it possible to drop a tiny fraction of data and change substantive conclusions of a data analysis? For instance, our check reveals that we can drop 1 data point out of over 16,500 in a randomized controlled trial of microcredit and change the sign of the estimated effect.

Bio:  Dr. Broderick is an Associate Professor with tenure in the Department of Electrical Engineering and Computer Science at MIT. She is a member of the MIT Laboratory for Information and Decision Systems, the MIT Statistics and Data Science Center, and the Institute for Data, Systems, and Society. She completed her Ph.D. in Statistics at the University of California, Berkeley in 2014. Dr. Broderick is interested in understanding how we can quickly, easily, and reliably quantify uncertainty and robustness in modern data analysis procedures. To that end, she is particularly interested in Bayesian inference – with an emphasis on scalable, nonparametric, and unsupervised learning.  She has been awarded membership to the COPSS Leadership Academy, an Early Career Grant from the Office of Naval Research, an AISTATS Notable Paper Award, an NSF CAREER Award, a Sloan Research Fellowship, an Army Research Office Young Investigator Program award, Google Faculty Research Awards, an Amazon Research Award, the ISBA Lifetime Members Junior Researcher Award, the Savage Award, the Evelyn Fix Memorial Medal and Citation, the Berkeley Fellowship, an NSF Graduate Research Fellowship, a Marshall Scholarship, and the Phi Beta Kappa Prize.

Learn more about Dr. Broderick’s research here.