Combinatorial QPs via a low-dimensional subspace sampling | IEEE Conference Publication | IEEE Xplore

Combinatorial QPs via a low-dimensional subspace sampling


Abstract:

Several large scale data processing problems can be formulated as quadratic programs subject to combinatorial constraints. We present a novel low-rank approximation frame...Show More

Abstract:

Several large scale data processing problems can be formulated as quadratic programs subject to combinatorial constraints. We present a novel low-rank approximation framework for problems such as sparse PCA, nonnegative PCA, or finding the k-densest submatrix. Our framework comes with provable approximation guarantees, that dependent on the spectrum of the data-set matrix. Our algorithm operates by solving a number of QP instances, which are randomly sampled from a low-dimensional subspace of the input matrix.
Date of Conference: 19-21 March 2014
Date Added to IEEE Xplore: 12 May 2014
Electronic ISBN:978-1-4799-3001-2
Conference Location: Princeton, NJ, USA

References

References is not available for this document.