Matrix Factorization Challenges Associated with Tensor and Tensor-like Computation
Charles Van Loan
Cornell University
Through reshaping, tensor computations reduce to matrix computations. If the underlying tensor is structured, then interesting structured matrix factorization problems can arise. Two examples will be discussed associated with order-4 tensors. We will also present a generalization of the generalized SVD that can be applied to collections of matrices that have the same column dimension.