Saved in:
Bibliographic Details
Main Authors: Harb, Elfarouk, Lam, Ho Shan
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2501.16130
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909469865148416
author Harb, Elfarouk
Lam, Ho Shan
author_facet Harb, Elfarouk
Lam, Ho Shan
contents Efficiently solving sparse linear systems $Ax=b$, where $A$ is a large, sparse, symmetric positive semi-definite matrix, is a core challenge in scientific computing, machine learning, and optimization. A major bottleneck in Gaussian elimination for these systems is fill-in, the creation of non-zero entries that increase memory and computational cost. Minimizing fill-in is NP-hard, and existing heuristics like Minimum Degree and Nested Dissection offer limited adaptability across diverse problem instances. We introduce \textit{ReFill}, a reinforcement learning framework enhanced by Graph Neural Networks (GNNs) to learn adaptive ordering strategies for fill-in minimization. ReFill trains a GNN-based heuristic to predict efficient elimination orders, outperforming traditional heuristics by dynamically adapting to the structure of input matrices. Experiments demonstrate that ReFill outperforms strong heuristics in reducing fill-in, highlighting the untapped potential of learning-based methods for this well-studied classical problem.
format Preprint
id arxiv_https___arxiv_org_abs_2501_16130
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ReFill: Reinforcement Learning for Fill-In Minimization
Harb, Elfarouk
Lam, Ho Shan
Machine Learning
Efficiently solving sparse linear systems $Ax=b$, where $A$ is a large, sparse, symmetric positive semi-definite matrix, is a core challenge in scientific computing, machine learning, and optimization. A major bottleneck in Gaussian elimination for these systems is fill-in, the creation of non-zero entries that increase memory and computational cost. Minimizing fill-in is NP-hard, and existing heuristics like Minimum Degree and Nested Dissection offer limited adaptability across diverse problem instances. We introduce \textit{ReFill}, a reinforcement learning framework enhanced by Graph Neural Networks (GNNs) to learn adaptive ordering strategies for fill-in minimization. ReFill trains a GNN-based heuristic to predict efficient elimination orders, outperforming traditional heuristics by dynamically adapting to the structure of input matrices. Experiments demonstrate that ReFill outperforms strong heuristics in reducing fill-in, highlighting the untapped potential of learning-based methods for this well-studied classical problem.
title ReFill: Reinforcement Learning for Fill-In Minimization
topic Machine Learning
url https://arxiv.org/abs/2501.16130