Jake Soloff

Jake Soloff is an Assistant Professor of Statistics at the University of Michigan. He earned his Ph.D. in Statistics from UC Berkeley in 2022 under advisors Adityanand Guntuboyina and Michael Jordan, followed by a postdoctoral fellowship at the University of Chicago with Rina Foygel Barber and Rebecca Willett.

• Research Focus: Statistical machine learning, algorithmic stability, empirical Bayes methods, and theoretical analysis of calibration and false discovery rate control.
• Key Contributions: Developed the Inflated Argmax for stable classification, proposed Cutoff Calibration Error metrics, analyzed distribution-free properties of isotonic regression, and created resampling techniques for stability guarantees.

Publication Trends: Recent work emphasizes stability in machine learning pipelines, empirical Bayes optimization for high-dimensional problems, and novel approaches to hypothesis testing with stochastic monotonicity constraints.

Software Development: Maintains the NPEB Python package for nonparametric empirical Bayes estimation.