39
0

Reduced Label Complexity For Tight 2\ell_2 Regression

Abstract

Given data XRn×d{\rm X}\in\mathbb{R}^{n\times d} and labels yRn\mathbf{y}\in\mathbb{R}^{n} the goal is find wRd\mathbf{w}\in\mathbb{R}^d to minimize Xwy2\Vert{\rm X}\mathbf{w}-\mathbf{y}\Vert^2. We give a polynomial algorithm that, \emph{oblivious to y\mathbf{y}}, throws out n/(d+n)n/(d+\sqrt{n}) data points and is a (1+d/n)(1+d/n)-approximation to optimal in expectation. The motivation is tight approximation with reduced label complexity (number of labels revealed). We reduce label complexity by Ω(n)\Omega(\sqrt{n}). Open question: Can label complexity be reduced by Ω(n)\Omega(n) with tight (1+d/n)(1+d/n)-approximation?

View on arXiv
Comments on this paper

We use cookies and other tracking technologies to improve your browsing experience on our website, to show you personalized content and targeted ads, to analyze our website traffic, and to understand where our visitors are coming from. See our policy.