Abstract:In this paper we propose efficient randomized fixed-precision techniques for low tubal rank approximation of tensors. The proposed methods are faster and more efficient than the existing fixed-precision algorithms for approximating the truncated tensor SVD (T-SVD). Besides, there are a few works on randomized single-pass algorithms for computing low tubal rank approximation of tensors, none of them experimentally reports the robustness of such algorithms for low-rank approximation of real-world data tensors e.g., images and videos. The current single-pass algorithms for tensors are generalizations of those developed for matrices to tensors. However, the single-pass randomized algorithms for matrices have been recently improved and stabilized. Motivated by this progress, in this paper, we also generalize them to the tensor case based on the tubal product (T-product). We conduct extensive simulations to study the robustness of them compared with the existing single-pass randomized algorithms. In particular, we experimentally found that single-pass algorithms with the sketching parameters of equal sizes usually lead to ill-conditioned tensor least-squares problems and inaccurate results. It is experimentally shown that our proposed single-pass algorithms are robust in this sense. Numerical results demonstrate that under the same conditions (setting the same hyper-parameters), our proposed algorithms provide better performance. Three applications to image compression, super-resolution problem and deep learning are also presented.

Best subspace tensor approximations

Decomposing tensors via rank-one approximations

On The Tensor Svd And The Optimal Low Rank Orthogonal Approximation Of Tensors

The number of singular vector tuples and uniqueness of best rank one approximation of tensors

On the rank and the approximation of symmetric tensors

Optimum Subspace Learning And Error Correction For Tensors

Better than best low-rank approximation with the singular value decomposition

Exact Tensor Completion Using T-Svd

A Randomized Algorithm for Tensor Singular Value Decomposition using an Arbitrary Number of Passes

On the Uniqueness and Perturbation to the Best Rank-One Approximation of a Tensor

Tensor Golub Kahan based on Einstein product

Optimal Sparse Singular Value Decomposition for High-dimensional High-order Data

Subtracting a best rank-1 approximation may increase tensor rank

The bounds for the best rank-1 approximation ratio of a finite dimensional tensor space

Efficient Algorithms for Low Tubal Rank Tensor Approximation with Applications to Image Compression, Super-Resolution and Deep Learning

Low Rank Tensor Decompositions and Approximations

Uniqueness of Nonnegative Tensor Approximations

Efficient Alternating Least Squares Algorithms for Low Multilinear Rank Approximation of Tensors

Very Large-Scale Singular Value Decomposition Using Tensor Train Networks

Guaranteed Functional Tensor Singular Value Decomposition

Singular Value Decomposition Approximation via Kronecker Summations for Imaging Applications