Computationally Efficient Reductions Between Some Statistical Models

Ashwin Pananjady's Official Webpage

ASHWIN PANANJADY – GEORGIA INSTITUTE OF TECHNOLOGY

ABSTRACT

Can a sample from one parametric statistical model (the source) be transformed into a sample from a different (target) model? Versions of this question were asked as far back as 1950, and a beautiful asymptotic theory of equivalence between experiments emerged in the latter half of the 20th century. Motivated by problems spanning information-computation gaps and differentially private data analysis, we address the analogous non-asymptotic question in high-dimensional problems and with algorithmic considerations. We show how a single observation from some source models can be approximately transformed to a single observation from a large class of target models by computationally efficient algorithms. I will present several such reductions and discuss their applications to the aforementioned problems.

This is joint work with Mengqi Lou and Guy Bresler.

Bio:
Ashwin Pananjady is the Gerald D. McInvale Early Career Assistant Professor at Georgia Tech, with a joint appointment between the H. Milton Stewart School of Industrial and Systems Engineering and the School of Electrical and Computer Engineering. Pananjady received his Ph.D. from the Department of Electrical Engineering and Computer Sciences (EECS) at the University of California Berkeley. His accolades include the David J. Sakrison Memorial Prize from UC Berkeley for the best dissertation in EECS, a Simons-Berkeley Research Fellowship, early career recognitions from the Bernoulli Society, Institute of Mathematical Statistics and Google Research, paper prizes from the Mathematical Optimization Society and Algorithmic Learning Theory conference, and the R. Narasimhan Memorial Lecture Award from TIFR.