Strongly Positive Semi-Definite Tensors and Strongly SOS Tensors
Liqun Qi, Chunfeng Cui

TL;DR
This paper introduces and explores classes of strongly positive semi-definite and sum-of-squares tensors, examining their properties, relationships, and examples, especially in odd-order cases, expanding the understanding of tensor positivity.
Contribution
It defines strongly PSD and strongly SOS tensors for odd orders, introduces strict Hankel and barren tensors, and studies their properties and distinctions from other tensor classes.
Findings
Strongly SOS tensors are a subset of strongly PSD tensors.
Odd-order barren tensors have no H-eigenvalues.
Examples show differences between various PSD-like tensor classes.
Abstract
We introduce {odd-order} strongly PSD (positive semi-definite) tensors which map real vectors to nonnegative vectors. We then introduce odd-order strongly SOS (sum-of-squares) tensors. A strongly SOS tensor maps real vectors to nonnegative vectors whose components are all SOS polynomials. Strongly SOS tensors are strongly PSD tensors. Odd order completely positive tensors are strongly SOS tensors. We also introduce strict Hankel tensors, which are also strongly SOS tensors. Odd order Hilbert tensors are strict Hankel tensors. However, the Laplacian tensor of a uniform hypergraph may not be strongly PSD. This motivates us to study wider PSD-like tensors. A cubic tensor is said to be a generalized PSD tensor if its corresponding symmetrization tensor has no negative H-eigenvalue. In the odd order case, this extension contains a peculiar tensor class, whose members have no H-eigenvalues at…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsTensor decomposition and applications · Elasticity and Material Modeling · Elasticity and Wave Propagation
