Tensor Derivatives, Unified Tensor-Form Differential Equations, and Model Reduction via Partial Tucker Decomposition

This paper develops a unified tensor calculus for matrix-valued functions and their derivatives, and leverages this framework to construct efficient model reduction techniques for high-dimensional tensor differential equations. We first establish a systematic theory of tensor differentiation, wherei