Landscape Complexity for the Empirical Risk of Generalized Linear Models

arXiv1912.02143MaRDI QIDQ6330441FDOQ6330441

Authors: Antoine Maillard, Gerard Ben Arous, Giulio Biroli 

Publication date: 4 December 2019

Abstract: We present a method to obtain the average and the typical value of the number of critical points of the empirical risk landscape for generalized linear estimation problems and variants. This represents a substantial extension of previous applications of the Kac-Rice method since it allows to analyze the critical points of high dimensional non-Gaussian random functions. Under a technical hypothesis, we obtain a rigorous explicit variational formula for the annealed complexity, which is the logarithm of the average number of critical points at fixed value of the empirical risk. This result is simplified, and extended, using the non-rigorous Kac-Rice replicated method from theoretical physics. In this way we find an explicit variational formula for the quenched complexity, which is generally different from its annealed counterpart, and allows to obtain the number of critical points for typical instances up to exponential accuracy.

This page was built for publication: Landscape Complexity for the Empirical Risk of Generalized Linear Models

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6330441)