Exponential weighting and oracle inequalities for projection methods

TL;DR

This paper introduces an exponential weighting aggregation method for projection estimates to recover unknown vectors from noisy data, providing a new risk bound for improved estimation accuracy.

Contribution

It presents a novel exponential weighting approach for projection methods and derives a new upper bound on the mean square risk, advancing theoretical understanding.

Findings

New upper bound for mean square risk of the aggregation method

Improved theoretical guarantees for projection estimate aggregation

Enhanced understanding of risk minimization in noisy data recovery

Abstract

We consider the problem of recovering an unknown vector from noisy data with the help of projection estimates. The goal is to find a convex combination of these estimates with the minimal risk. We study an aggregation method based on the so-called exponential weighting and provide a new upper bound for the mean square risk of this method.