24
0
v1v2 (latest)

A Concentration of Measure Framework to study convex problems and other implicit formulation problems in machine learning

Abstract

This paper provides a framework to show the concentration of solutions YY^* to convex minimizing problem where the objective function ϕ(X)(Y)\phi(X)(Y) depends on some random vector XX satisfying concentration of measure hypotheses. More precisely, the convex problem translates into a contractive fixed point equation that ensure the transmission of the concentration from XX to YY^*. This result is of central interest to characterize many machine learning algorithms which are defined through implicit equations (e.g., logistic regression, lasso, boosting, etc.). Based on our framework, we provide precise estimations for the first moments of the solution YY^*, when X=(x1,,xn)X= (x_1,\ldots, x_n) is a data matrix of independent columns and ϕ(X)(y)\phi(X)(y) writes as a sum 1ni=1nhi(xiTY)\frac{1}{n}\sum_{i=1}^n h_i(x_i^TY). That allows to describe the behavior and performance (e.g., generalization error) of a wide variety of machine learning classifiers.

View on arXiv
Comments on this paper