A Preliminary Agentic Framework for Matrix Deflation
Abstract
Can a small team of agents peel a matrix apart, one rank-1 slice at a time? We propose an agentic approach to matrix deflation in which a solver Large Language Model (LLM) generates rank-1 Singular Value Decomposition (SVD) updates and a Vision Language Model (VLM) accepts or rejects each update and decides when to stop, eliminating fixed norm thresholds. Solver stability is improved through in-context learning (ICL) and types of row/column permutations that expose visually coherent structure. We evaluate on Digits (), CIFAR-10 ( grayscale), and synthetic () matrices with and without Gaussian noise. In the synthetic noisy case, where the true construction rank is known, numerical deflation provides the noise target and our best agentic configuration differs by only RMSE of the target. For Digits and CIFAR-10, targets are defined by deflating until the Frobenius norm reaches of the original. Across all settings, our agent achieves competitive results, suggesting that fully agentic, threshold-free deflation is a viable alternative to classical numerical algorithms.
Cite
@article{arxiv.2601.08219,
title = {A Preliminary Agentic Framework for Matrix Deflation},
author = {Paimon Goulart and Evangelos E. Papalexakis},
journal= {arXiv preprint arXiv:2601.08219},
year = {2026}
}