Jake A. Soloff

I am a postdoctoral researcher at the University of Chicago, broadly interested in the foundations of data science and machine learning. In my research, I develop theoretical frameworks that enable strong statistical guarantees without restrictive assumptions, allowing practitioners to apply rigorous tools with minimal tuning and maximal flexibility.

Starting this fall, I will join the University of Michigan as an assistant professor of statistics.

Current Position

Postdoctoral Researcher in Statistics

University of Chicago

Advisors: Rina Foygel Barber & Rebecca Willett

Education

Ph.D. in Statistics (2022)

University of California, Berkeley

Advisors: Aditya Guntuboyina & Michael I. Jordan

Select Publications

Can a calibration metric be both testable and actionable?

R. Rossellini, J. A. Soloff, R. F. Barber, Z. Ren, and R. Willett

To appear in COLT 2025

Building a stable classifier with the inflated argmax

J. A. Soloff, R. F. Barber, and R. Willett

Advances in Neural Information Processing Systems 37 (NeurIPS 2024)

preprint paper experiments

Bagging provides assumption-free stability

J. A. Soloff, R. F. Barber, and R. Willett

Journal of Machine Learning Research, 25(131), 1-35

preprint paper experiments

The edge of discovery: Controlling the local false discovery rate at the margin

J. A. Soloff, D. Xiang, and W. Fithian

Annals of Statistics, 52(2), 580-601

preprint paper experiments

Multivariate, heteroscedastic empirical Bayes via nonparametric maximum likelihood

J. A. Soloff, A. Guntuboyina, and B. Sen

Journal of the Royal Statistical Society: Series B, 87(1), 1-32

preprint paper experiments

Distribution-free properties of isotonic regression

J. A. Soloff, A. Guntuboyina, and J. Pitman

Electronic Journal of Statistics, 13(2), 3243-3253

Preprints

A frequentist local false discovery rate

D. Xiang, J. A. Soloff, and W. Fithian

Cross-validation with antithetic Gaussian randomization

S. Liu, S. Panigrahi, and J. A. Soloff

preprint

Stabilizing black-box model selection with the inflated argmax

M. Adrian, J. A. Soloff, and R. Willett

preprint

Testing conditional independence under isotonicity

R. Hore, J. A. Soloff, R. F. Barber, and R. J. Samworth

preprint

Stability via resampling: Statistical problems beyond the real line

J. A. Soloff, R. F. Barber, and R. Willett

preprint

Covariance estimation with nonnegative partial correlations

J. A. Soloff, A. Guntuboyina, and M. I. Jordan

Incentive-theoretic Bayesian inference for collaborative science

S. Bates, M. I. Jordan, M. Sklar, and J. A. Soloff

Principal-agent hypothesis testing

S. Bates, M. I. Jordan, M. Sklar, and J. A. Soloff

Software

NPEB

A Python package for empirical Bayes using nonparametric maximum likelihood