AI Paper Atlas

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

Learning Feature Sparse Principal Subspace

Lai Tian, Feiping Nie, Rong Wang, Xuelong Li

NeurIPS 2020 · Learning theory and algorithms › Matrix and tensor methods · approach: Sparse and robust PCA

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 IPU instead of SOAPSOAP · Synthetic scheme A, random initIntersection Ratio: 0.73 -> 0.97Improved
use Go instead of SRTSRT · Synthetic scheme B, random initIntersection Ratio: 0.59 -> 0.99Improved
use IPU instead of GoGo · Synthetic scheme D, random initIntersection Ratio: 0.91 -> 0.83Hurt
use IPU instead of GoGo · Synthetic scheme F, random initIntersection Ratio: 0.83 -> 0.62Hurt
use IPU local ascent after GoGo · Synthetic scheme A, random initHit Frequency: 0.74 -> 1.00Improved
use IPU instead of GoGo · Synthetic scheme E, random initHit Frequency: 0.48 -> 0.89Improved
use Go instead of CSSPCSSP · Synthetic scheme A, random initRelative Error: 0.88 -> 0.01Improved
use Low Rank Approx. (Go) init instead of Convex Relaxation init (ablation)all methods · Synthetic dataexplained varianceNo clear effect
use best of 20 random initializations instead of Low Rank Approx. init (ablation)all methods · Synthetic dataobjective valueImproved
use Go on low-rank covarianceGo · Synthetic scheme CIntersection Ratio: 0.77 -> 1.00Improved

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