Abstract:This paper is concerned with the approximation of tensors using tree-based tensor formats, which are tensor networks whose graphs are dimension partition trees. We consider Hilbert tensor spaces of multivariate functions defined on a product set equipped with a probability measure. This includes the case of multidimensional arrays corresponding to finite product sets. We propose and analyse an algorithm for the construction of an approximation using only point evaluations of a multivariate function, or evaluations of some entries of a multidimensional array. The algorithm is a variant of higher-order singular value decomposition which constructs a hierarchy of subspaces associated with the different nodes of the tree and a corresponding hierarchy of interpolation operators. Optimal subspaces are estimated using empirical principal component analysis of interpolations of partial random evaluations of the function. The algorithm is able to provide an approximation in any tree-based format with either a prescribed rank or a prescribed relative error, with a number of evaluations of the order of the storage complexity of the approximation format. Under some assumptions on the estimation of principal components, we prove that the algorithm provides either a quasi-optimal approximation with a given rank, or an approximation satisfying the prescribed relative error, up to constants depending on the tree and the properties of interpolation operators. The analysis takes into account the discretization errors for the approximation of infinite-dimensional tensors. Several numerical examples illustrate the main results and the behavior of the algorithm for the approximation of high-dimensional functions using hierarchical Tucker or tensor train tensor formats, and the approximation of univariate functions using tensorization.

Adaptive quarklet tree approximation

Multirate Shading with Piecewise Interpolatory Approximation

Adaptive and non-adaptive randomized approximation of high-dimensional vectors

Almost-sure quasi-optimal approximation in reproducing kernel Hilbert spaces

Shorter quantum circuits via single-qubit gate approximation

Higher-order principal component analysis for the approximation of tensors in tree-based low-rank formats

Approximate Evaluation of Quantitative Second Order Queries

On the reconstruction of functions from values at subsampled quadrature points

On approximation by tight wavelet frames on the field of $p$-adic numbers

Efficient Large-scale Approximate Nearest Neighbor Search on the GPU

Geometric and arithmetic aspects of approximation vectors

On converses to the polynomial method

Diophantine Approximation in local function fields via Bruhat-Tit trees

Efficient Approximation of Fractional Hypertree Width

An exact tree projection algorithm for wavelets

Error Bounds for Compositions of Piecewise Affine Approximations

An adaptive partition of unity method for multivariate Chebyshev polynomial approximations

Multiscale interpolative construction of quantized tensor trains

A Geometric Approach to Polynomial and Rational Approximation

Uniform Approximation by Polynomials with Integer Coefficients via the Bernstein Lattice

A sparse approximation for fractional Fourier transform