Tensor train construction from tensor actions, with application to compression of large high order derivative tensors
From MaRDI portal
(Redirected from Publication:5146679)
Multilinear algebra, tensor calculus (15A69) PDEs in connection with statistics (35Q62) PDEs with randomness, stochastic partial differential equations (35R60) Bayesian problems; characterization of Bayes procedures (62C10) Probabilistic models, generic numerical methods in probability and statistics (65C20) Numerical linear algebra (65F99) Randomized algorithms (68W20)
Abstract: We present a method for converting tensors into tensor train format based on actions of the tensor as a vector-valued multilinear function. Existing methods for constructing tensor trains require access to "array entries" of the tensor and are therefore inefficient or computationally prohibitive if the tensor is accessible only through its action, especially for high order tensors. Our method permits efficient tensor train compression of large high order derivative tensors for nonlinear mappings that are implicitly defined through the solution of a system of equations. Array entries of these derivative tensors are not directly accessible, but actions of these tensors can be computed efficiently via a procedure that we discuss. Such tensors are often amenable to tensor train compression in theory, but until now no efficient algorithm existed to convert them into tensor train format. We demonstrate our method by compressing a Hilbert tensor of size , and by forming high order (up to order derivatives/ order tensors) Taylor series surrogates of the noise-whitened parameter-to-output map for a stochastic partial differential equation with boundary output.
Recommendations
Cites work
- \(O(d \log N)\)-quantics approximation of \(N\)-\(d\) tensors in high-dimensional numerical modeling
- A computational framework for infinite-dimensional Bayesian inverse problems. I: The linearized case, with application to global seismic inversion
- A computational framework for infinite-dimensional Bayesian inverse problems. II: stochastic Newton MCMC with application to ice sheet flow inverse problems
- A literature survey of low-rank tensor approximation techniques
- A tensor-train accelerated solver for integral equations in complex geometries
- Approximation and sampling of multivariate probability distributions in the tensor train decomposition
- Approximation of 2d²^d matrices using tensor decomposition
- Approximation of matrices with logarithmic number of parameters
- Automated calculation of higher order partial differential equation constrained derivative information
- Automated solution of differential equations by the finite element method. The FEniCS book
- Black box approximation of tensors in hierarchical Tucker format
- Dimension-independent likelihood-informed MCMC
- Fast construction of hierarchical matrix representation from matrix-vector multiplication
- Fast solution of parabolic problems in the tensor train/quantized tensor train format with initial application to the Fokker-Planck equation
- Finding structure with randomness: probabilistic algorithms for constructing approximate matrix decompositions
- Hessian-based adaptive sparse quadrature for infinite-dimensional Bayesian inverse problems
- Hessian-based model reduction for large-scale systems with initial-condition inputs
- HESSIAN-BASED SAMPLING FOR HIGH-DIMENSIONAL MODEL REDUCTION
- Low-rank tensor structure of linear diffusion operators in the TT and QTT formats
- Matrix product state representations
- Mean-variance risk-averse optimal control of systems governed by PDEs with random parameter fields using quadratic approximations
- Most tensor problems are NP-hard
- Optimal low-rank approximations of Bayesian linear inverse problems
- Perturbation analysis for the Darcy problem with log-normal permeability
- Randomized algorithms for the approximations of Tucker and the tensor train decompositions
- Scalable and efficient algorithms for the propagation of uncertainty from data through inference to prediction for large-scale problems, with application to flow of the antarctic ice sheet
- Shifted power method for computing tensor eigenpairs
- Taylor approximation and variance reduction for PDE-constrained optimal control under uncertainty
- Tensor Decompositions and Applications
- Tensor networks for dimensionality reduction and large-scale optimization. I: Low-rank tensor decompositions
- Tensor train approximation of moment equations for elliptic equations with lognormal coefficient
- Tensor-train decomposition
- Tensor-train ranks for matrices and their inverses
- The alternating linear scheme for tensor optimization in the tensor train format
- The art of differentiating computer programs. An introduction to algorithmic differentiation.
- TT-cross approximation for multidimensional arrays
Cited in
(16)- Derivative-informed projected neural networks for high-dimensional parametric maps governed by PDEs
- Optimal design of acoustic metamaterial cloaks under uncertainty
- MERACLE: constructive layer-wise conversion of a tensor train into a MERA
- Efficient construction of tensor ring representations from sampling
- Multiresolution low-rank tensor formats
- Adaptive hierarchical subtensor partitioning for tensor compression
- Taylor approximation for chance constrained optimization problems governed by partial differential equations with high-dimensional random parameters
- Projected Wasserstein Gradient Descent for High-Dimensional Bayesian Inference
- An Offline-Online Decomposition Method for Efficient Linear Bayesian Goal-Oriented Optimal Experimental Design: Application to Optimal Sensor Placement
- Randomized Algorithms for Rounding in the Tensor-Train Format
- Streaming Tensor Train Approximation
- Derivative-informed neural operator: an efficient framework for high-dimensional parametric derivative learning
- Sketch-based multiplicative updating algorithms for symmetric nonnegative tensor factorizations with applications to face image clustering
- Efficient randomized algorithms for computing an approximation of the tensor train decomposition
- Accurate, scalable, and efficient Bayesian optimal experimental design with derivative-informed neural operators
- Taylor approximation variance reduction for approximation errors in PDE-constrained Bayesian inverse problems
This page was built for publication: Tensor train construction from tensor actions, with application to compression of large high order derivative tensors
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5146679)
