Mathlib.Linear­Algebra.Tensor

Tensor methods for finite symmetric rank-one decompositions, centered on contraction pencils, their simultaneous diagonalization and conditioning, and quantitative factor recovery from tensors of order d + d + q.

Symmetric­Pencil 40 core · 27 supporting · 7 submodules Recovery of the rank-one factors of a symmetric tensor from its pencil of contracted slices: spectral projectors, quantitative conditioning of the eigendecomposition, matching of eigenvalues across slices, and trace-coor