AI Paper Atlas

Research map / Learning theory and algorithms / Matrix and tensor methods

Single Pass PCA of Matrix Products

Shanshan Wu, Srinadh Bhojanapalli, Sujay Sanghavi, Alexandros G Dimakis

NeurIPS 2016 · Learning theory and algorithms › Matrix and tensor methods · approach: Matrix completion and factorization · 7 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use rescaled JL estimates based on actual column norms instead of standard sketched dot products (ablation)SMP-PCA · Unit-vector dot-product simulationmean squared error: 0.129 -> 0.053Improved
Use SMP-PCA instead of SVD on the separately sketched matricesRank-r approximation of A^T B · SIFT10Kspectral norm error: 1.8Improved
Use SMP-PCA instead of SVD on the separately sketched matricesRank-r approximation of A^T B · NIPS-BWspectral norm error: 1.1Improved
Use SMP-PCA instead of SVD on separately sketched matricesRank-r approximation of A^T B · Columns drawn from a conespectral norm errorImproved
Increase sketch size k (ablation)SMP-PCA · SIFT10K and NIPS-BWspectral norm errorImproved
Use one-pass SMP-PCA instead of two-pass LELALow-rank approximation of A^T B · Synthetic, URL-malicious, and URL-benignspectral norm error: 0.0274 -> 0.0280; 0.0182 -> 0.0188; 0.0105 -> 0.0117Hurt
Use one-pass SMP-PCA instead of two-pass LELASpark implementation · 150GB synthetic datasetrunning time: 56 mins versus 34 minsImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy