Optimal Krylov On Average

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Luo, Qi, Schäfer, Florian
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909566927634432
author Luo, Qi
Schäfer, Florian
author_facet Luo, Qi
Schäfer, Florian
contents We propose an adaptive randomized truncation estimator for Krylov subspace methods that optimizes the trade-off between the solution variance and the computational cost, while remaining unbiased. The estimator solves a constrained optimization problem to compute the truncation probabilities on the fly, with minimal computational overhead. The problem has a closed-form solution when the improvement of the deterministic algorithm satisfies a diminishing returns property. We prove that obtaining the optimal adaptive truncation distribution is impossible in the general case. Without the diminishing return condition, our estimator provides a suboptimal but still unbiased solution. We present experimental results in GP hyperparameter training and competitive physics-informed neural networks problem to demonstrate the effectiveness of our approach.
format Preprint
id arxiv_https___arxiv_org_abs_2504_03914
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Optimal Krylov On Average
Luo, Qi
Schäfer, Florian
Numerical Analysis
65F10, 68W20, 65B99, 65C05
We propose an adaptive randomized truncation estimator for Krylov subspace methods that optimizes the trade-off between the solution variance and the computational cost, while remaining unbiased. The estimator solves a constrained optimization problem to compute the truncation probabilities on the fly, with minimal computational overhead. The problem has a closed-form solution when the improvement of the deterministic algorithm satisfies a diminishing returns property. We prove that obtaining the optimal adaptive truncation distribution is impossible in the general case. Without the diminishing return condition, our estimator provides a suboptimal but still unbiased solution. We present experimental results in GP hyperparameter training and competitive physics-informed neural networks problem to demonstrate the effectiveness of our approach.
title Optimal Krylov On Average
topic Numerical Analysis
65F10, 68W20, 65B99, 65C05
url https://arxiv.org/abs/2504.03914