Abstract:The convergence of the conjugate gradient method for solving large-scale and sparse linear equation systems depends on the spectral properties of the system matrix, which can be improved by preconditioning. In this paper, we develop a computationally efficient data-driven approach to accelerate the generation of effective preconditioners. We, therefore, replace the typically hand-engineered preconditioners by the output of graph neural networks. Our method generates an incomplete factorization of the matrix and is, therefore, referred to as neural incomplete factorization (NeuralIF). Optimizing the condition number of the linear system directly is computationally infeasible. Instead, we utilize a stochastic approximation of the Frobenius loss which only requires matrix-vector multiplications for efficient training. At the core of our method is a novel message-passing block, inspired by sparse matrix theory, that aligns with the objective of finding a sparse factorization of the matrix. We evaluate our proposed method on both synthetic problem instances and on problems arising from the discretization of the Poisson equation on varying domains. Our experiments show that by using data-driven preconditioners within the conjugate gradient method we are able to speed up the convergence of the iterative procedure. The code is available at <a class="link-external link-https" href="https://github.com/paulhausner/neural-incomplete-factorization" rel="external noopener nofollow">this https URL</a>.

Numerical Study on Incomplete Orthogonal Factorization Preconditioners

CIMGS: an Incomplete Orthogonal FactorizationPreconditioner

Incomplete Factorization Preconditioning for Linear Least Squares Problems

A Class of Inexact Block Factorization Preconditioners for Indefinite Matrices with a Three-by-three Block Structure.

Parallel Incomplete Factorization Preconditioning of Block Tridiagonal Linear Systems with 2-D Domain Decomposition

A Hybrid Recursive Multilevel Incomplete Factorization Preconditioner for Solving General Linear Systems

Indefinite Preconditioning Techniques for Symmetric Indefinite Linear Systems

On Some New Approximate Factorization Methods for Block Tridiagonal Matrices Suitable for Vector and Parallel Processors

Parallelization of incomplete factorization preconditioning of block tridiagonal linear systems with 1-D domain decomposition

Flexible Incomplete Cholesky Factorization with Multi-Parameters to Control the Number of Nonzero Elements in Preconditioners.

Neural incomplete factorization: learning preconditioners for the conjugate gradient method

Learning incomplete factorization preconditioners for GMRES

Some Preconditioners for Block Pentadiagonal Linear Systems Based on New Approximate Factorization Methods

Effective Preconditioning Through Minimum Degree Ordering Interleaved with Incomplete Factorization

Avoiding breakdown in incomplete factorizations in low precision arithmetic

On Relaxed Nested Factorization and Combination Preconditioning

Incomplete Cholesky Factorization with Sparsity Pattern Modi cation

On incomplete Cholesky factorizations in half precision arithmetic

High Performance Block Incomplete LU Factorization