Algorithm to find a maximum of a multilinear map over a product of spheres
arXiv:1110.62174 citationsh-index: 9
Originality Synthesis-oriented
AI Analysis
It provides a computational method for a fundamental problem in multilinear algebra, with potential impact on quantum physics and tensor approximation.
The paper presents an algorithm for computing the 2-norm maximum of a multilinear map over a product of spheres, with applications to singular value computation and finding the closest rank-one tensor.
We provide an algorithm to compute the 2-norm maximum of a multilinear map over a product of spheres. As a corollary we give a method to compute the first singular value of a linear map and an application to the theory of entangled states in quantum physics. Also, we give an application to find the closest rank-one tensor of a given one.