{"id":"arxiv/2609.11369","url":"https://arxiv.org/abs/2609.11369","tags":["math.NA"],"score":0,"title":"The tubal Arnoldi method for tensor function approximation","domain":"math.NA","source":"arxiv","authors":["Fatima Bouyghf","Mohamed El Ghomari","Alaa El Ichi"],"reviews":0,"subject":"math.NA","summary":"This paper describes Krylov subspace methods based on the tensor t-product for approximating quantities associated with third-order tensor functions. Specifically, it introduces the tensor-tubal Arnoldi method, which projects highdimensional tensor problems onto lower-dimensional tensor Krylov subspaces. The derived method exploits novel algebraic properties of the t-product to address large-scale tensor computations efficiently. Applications include the evaluation of parameter-dependent tensor functions and the solution of multidimensional ordinary differential equations. Numerical…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-09-10","reactions":{}}