My research investigates foundational questions about responsible machine learning.
Much of this work aims to identify problematic behaviors that emerge in machine-learned models and to develop algorithmic tools that provably mitigate such behaviors.
More broadly, I am interested in how the theory of computation can provide insight into emerging societal and scientific challenges.
Prior to Cornell, I was a Miller Postdoctoral Fellow at UC Berkeley, hosted by Shafi Goldwasser.
I completed my Ph.D. in the Stanford Theory Group under the guidance of Omer Reingold.
The Sample Complexity of Multi-Class Omniprediction
Joey Rivkin, Princewill Okoroafor, MPK
preprint 2026
Publicly-Verifiable Certificates for Statistical Algorithms [arXiv]
Michael Ngo and MPK
ITCS 2027
Near-Optimal Algorithms for Omniprediction [arXiv]
Princewill Okoroafor, Robert Kleinberg, MPK
FOCS 2025
Planting Undetectable Backdoors in Machine Learning Models [arXiv] [Quanta]
Shafi Goldwasser, MPK, Vinod Vaikuntanathan, Or Zamir
FOCS 2022
Outcome Indistinguishability [arXiv]
Cynthia Dwork, MPK, Omer Reingold, Guy N. Rothblum, Gal Yona
STOC 2021
Multicalibration: Calibration for the (Computationally-Identifiable) Masses
[arXiv]
Úrsula Hébert-Johnson, MPK, Omer Reingold,
Guy N. Rothblum
ICML 2018
I am fortunate to work with a wonderful set students.
Current Students:
Joey Rivkin, Amanda Wang
Alumni: Michael Ngo (BS 2026 → MIT PhD)