Computational Equivalence of Spiked Covariance and Spiked Wigner Models via Gram-Schmidt Perturbation

Abstract

In this work, we show the first average-case reduction transforming the sparse Spiked Covariance Model into the sparse Spiked Wigner Model and as a consequence obtain the first computational equivalence result between two well-studied high-dimensional statistics models. Our approach leverages a new perturbation equivariance property for Gram-Schmidt orthogonalization, enabling removal of dependence in the noise while preserving the signal.

Cite

Text

Bresler and Harbuzova. "Computational Equivalence of Spiked Covariance and Spiked Wigner Models via Gram-Schmidt Perturbation." Proceedings of Thirty Eighth Conference on Learning Theory, 2025.

Markdown

[Bresler and Harbuzova. "Computational Equivalence of Spiked Covariance and Spiked Wigner Models via Gram-Schmidt Perturbation." Proceedings of Thirty Eighth Conference on Learning Theory, 2025.](https://mlanthology.org/colt/2025/bresler2025colt-computational/)

BibTeX

@inproceedings{bresler2025colt-computational,
  title     = {{Computational Equivalence of Spiked Covariance and Spiked Wigner Models via Gram-Schmidt Perturbation}},
  author    = {Bresler, Guy and Harbuzova, Alina},
  booktitle = {Proceedings of Thirty Eighth Conference on Learning Theory},
  year      = {2025},
  pages     = {594-595},
  volume    = {291},
  url       = {https://mlanthology.org/colt/2025/bresler2025colt-computational/}
}