← Back to all news
arXiv Machine Learning By Pierre Agui\'e, Mathieu Even, Laurent Massouli\'e

Improved Analysis of the Accelerated Noisy Power Method with Applications to Decentralized PCA

Read the original on arXiv Machine Learning →

arXiv:2602. 03682v2 Announce Type: replace-cross Abstract: We analyze the Accelerated Noisy Power Method, an algorithm for Principal Component Analysis in the setting where only inexact matrix-vector products are available, which can arise for instance in decentralized PCA.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

Related stories

arXiv Machine Learning
Sep 23

Gap-Free Streaming PCA Beyond Rank-One Updates: Near-Optimal Rates and Applications to Differential Privacy

The paper presents a new analysis of Oja's algorithm for streaming principal component analysis (PCA) that works without any eigengap assumptions, achieving near‑optimal rates and matching lower bounds. It extends the results to a Rayleigh quotient notion of approximate PCA, resolving an open question, and applies the findings to provide gap‑free differentially private PCA guarantees for sub‑Gaussian data. The analysis relies solely on a second‑moment bound of stochastic updates, avoiding the almost‑sure bounds used in previous work.

By Anming Gu, Syamantak Kumar, Kevin Tian, Chutong Yang
safety
More like this →
arXiv Machine Learning
Sep 3

Robust Streaming PCA

The paper studies streaming principal component analysis under a robust setting where the covariance matrix can vary within a temporal uncertainty set, rather than being fixed. It establishes fundamental convergence limits for any algorithm that recovers principal components and analyzes the noisy power method and Oja's algorithm, showing that the noisy power method achieves rate‑optimal convergence in this setting. Numerical experiments on synthetic and real‑world data confirm the theoretical findings.

By Daniel Bienstock, Minchan Jeong, Apurv Shukla, Se-Young Yun
More like this →
arXiv Machine Learning
Sep 23

SuperPCA: subspace analysis and an efficient algorithm for high-dimensional PCA

SuperPCA is a new algorithm for high‑dimensional principal component analysis that exploits an approximate eigenspace of the sample covariance matrix. The authors show that the subspace spanned by several leading eigenvectors contains useful signal information long before individual eigenvectors converge, and they derive posteriori bounds on the angle between this subspace and the true signal subspace. By using only a small number of subsampled coordinates, SuperPCA can achieve up to a ten‑fold improvement in accuracy over classical PCA while reducing data acquisition costs, especially when the signals are approximately sparse.

By Irina-Beatrice Haas, Maike Meier, Yuji Nakatsukasa, Taejun Park
More like this →