Please note: This seminar will take place in DC 1304 and online.
Ittai Rubinstein, PhD student
Electrical Engineering and Computer Science, MIT
Data attribution estimates the effect of removing a set of samples from a model’s training set without retraining the model from scratch and are used for interpretability, credit assignment, privacy and more. However, key approaches to data attribution significantly underestimate removal effects in the high-dimensional regime (#params >= Omega(#samples)), and existing theoretical analyses require strong convexity assumptions that rarely hold in practice, even for simple linear probes.
In this talk, we will present a correction to the leading approaches to data attribution that improve accuracy in the high-dimensional regime and the first theoretical analysis of these data attribution methods without strong convexity.
To attend this seminar in person, please go to DC 1304. You can also attend virtually on Zoom.