Research
Gap-free Differentially Private PCA for Gaussian Data
The paper “Gap‑free Differentially Private PCA for Gaussian Data” by Alina Ene and Huy L. Nguyen proposes a differentially private algorithm for principal component analysis that does not rely on a sp
The paper “Gap‑free Differentially Private PCA for Gaussian Data” by Alina Ene and Huy L. Nguyen proposes a differentially private algorithm for principal component analysis that does not rely on a spectral gap assumption, thereby extending existing methods to more general Gaussian datasets. Submitted to arXiv on 25 September 2026 (arXiv:2609.31614), the work is situated at the intersection of data‑structuring and machine‑learning research (cs.DS and cs.LG) and can be accessed via its DOI 10.48550/arXiv.2609.31614.
Source: arXiv cs.LG | 2026-09-28