中文

Manifold-based Algorithms for the Hadamard Decomposition

最优化与控制 2026-05-29 v1 机器学习 数值分析 信号处理 数值分析

摘要

Given a matrix XX, and two ranks r1r_1 and r2r_2, the Hadamard decomposition (HD) looks for two low-rank matrices, X1X_1 of rank r1r_1 and X2X_2 of rank r2r_2, both of the same size as XX, such that XX1X2X\approx X_1\circ X_2, where \circ is the Hadamard (element-wise) product. In most cases, HD is more expressive than standard low-rank approximations such as the truncated singular value decomposition (TSVD), as it can represent higher-rank matrices with the same number of parameters; this is because the rank of X1X2X_1 \circ X_2 is generically equal to r1r2r_1 r_2. In this paper, we first present some theoretical insights for HD, in particular a useful reformulation XWHX\approx WH^\top where WW and HH have r1r2r_1 r_2 columns and belong to certain manifolds. These allow us to develop three new algorithms for computing HD. The first one uses the representation XX1X2X\approx X_1\circ X_2 and relies on the Manopt toolbox. The other two rely on the reformulation XWHX\approx WH^\top: one is a block projected gradient method, and the other is a manifold-based gradient descent algorithm that does not require projection onto the feasible set. The last two algorithms are particularly effective for handling large sparse data. We also propose new initializations that allow us to improve the accuracy of the HD. We compare our algorithms and initialization strategies with the TSVD and with the state of the art. Numerical results show that the new methods are efficient and competitive on both synthetic and real data.

关键词

引用

@article{arxiv.2605.28980,
  title  = {Manifold-based Algorithms for the Hadamard Decomposition},
  author = {Nicolas Gillis and Subhayan Saha and Stefano Sicilia and Arnaud Vandaele},
  journal= {arXiv preprint arXiv:2605.28980},
  year   = {2026}
}

备注

27 pages, code available from https://github.com/StefanoSicilia/Hadamard-Decomposition