Approximate Real Symmetric Tensor Rank

2022/07/25 by Ergür, Alperen A., Bueno, Jesus Rebollo, Valettas, Petros
#Commutative Algebra (math.AC) #FOS: Computer and information sciences #FOS: Mathematics #Machine Learning (cs.LG) #Numerical Analysis (math.NA) #Optimization and Control (math.OC) #Symbolic Computation (cs.SC)

paper · doi:10.48550/arxiv.2207.12529

Abstract

We investigate the effect of an ε-room of perturbation tolerance on symmetric tensor decomposition. To be more precise, suppose a real symmetric d-tensor f, a norm ||.|| on the space of symmetric d-tensors, and ε >0 are given. What is the smallest symmetric tensor rank in the ε-neighborhood of f? In other words, what is the symmetric tensor rank of f after a clever ε-perturbation? We prove two theorems and develop three corresponding algorithms that give constructive upper bounds for this question. With expository goals in mind; we present probabilistic and convex geometric ideas behind our results, reproduce some known results, and point out open problems.

Related

• Remarks on the symmetric rank of symmetric tensors
• Symmetric tensors and symmetric tensor rank
• On the ranks and border ranks of symmetric tensors
• Rank decomposition and symmetric rank decomposition over arbitrary fields
• Symmetric tensors: rank, Strassen's conjecture and e-computability