Research | Sasha's webpage

Agenda

I am broadly interested in the following topics:

Post-hoc predictive uncertainty quantification (conformal prediction, calibration)
Sequential/batch nonparametric inference (two-sample, indepedendece testing)
Practical methods for valid statistical inference under distribution shifts

Recent publications

Aleksandr Podkopaev, Darren Xu, Kuang-Chih Lee (2024). Adaptive conformal inference by betting. In ICML 2024.

Aleksandr Podkopaev, Aaditya Ramdas (2023). Sequential predictive two-sample and independence testing. In NeurIPS 2023.

Aleksandr Podkopaev, Patrick Bloebaum, Shiva Prasad Kasiviswanathan, Aaditya Ramdas (2023). Sequential kernelized independence testing. In ICML 2023.

Aleksandr Podkopaev, Aaditya Ramdas (2021). Distribution-free uncertainty quantification for classification under label shift. In UAI 2021 (Long oral talk).