Michelle Ntampaka (Space Telescope Science Institute): AI+Science Schmidt Fellows Speaker Series
Organized by the University of Chicago’s Eric and Wendy Schmidt AI in Science Postdoctoral Fellowship Program.
Agenda
4:00pm – 4:45pm: Presentation
4:45pm – 5:00pm: Q&A
5:00pm – 5:30pm: Reception
Meeting location
William Eckhardt Research Center. Room 401
5640 S Ellis Avenue, Chicago, IL 60637
Map It
Abstract: ML methods for cosmology often hinge on one essential assumption: that the simulations sufficiently capture reality. But is this necessarily the case? Observations may be out of the domain of the simulations, with the data sets differing in subtle but important ways. These differences can be very difficult to detect and quantify, but they must be detected and quantified, or risk biasing the results. This is especially important in the scenario where one wants to train deep learning methods on simulations and then directly apply the models to observations as a black box. In this talk, I will highlight opportunities and challenges for creating credible ML to interpret galaxy cluster observations. I will expand on ML interpretability and domain adaptation as keystones for building models that give trustworthy results. And I will show examples of how machine learning can be used, not just as a tool for getting “better” results at the expense of understanding but also as a partner that can point us toward physical discovery.
Bio: Dr. Michelle Ntampaka is an Associate Astronomer at Space Telescope Science Institute, where she also serves as the Deputy Head of the Data Science Mission Office. She joined STScI after a first career in education, and has a decade of classroom teaching experience. She has run multiple teacher-training programs in Africa, teaching Rwandese high school educators how to use inexpensive, commonly available items as lecture demonstrations to enhance student learning. At her career transition from education to astronomy, Dr. Ntampaka received an American Physical Society Blewett Fellowship, a national award for women returning to careers in physics. After completing her Ph.D. at Carnegie Mellon University, she was a member of the inaugural cohort of Harvard Data Science Fellows. Dr. Ntampaka’s research focuses on ways to use machine learning as a tool for cosmological discovery. While the field has traditionally used machine learning techniques as a black box, Dr. Ntampaka’s work explores how to build trustworthy and interpretable ML methods as a partner for interpreting cosmological large-scale structure.
Parking
Campus North Parking
5505 S Ellis Ave
Map It