AI Paper Atlas

Research map / Learning theory and algorithms / Clustering

Provable Variable Selection for Streaming Features

Jing Wang, Jie Shen, Ping Li

ICML 2018 · Learning theory and algorithms › Clustering · approach: k-means++ seeding analysis · 4 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 online leverage-score feature selection (OFS) instead of Laplacian scoreLapS · CIFAR-10CPU time: 2 mins -> 2.28Improved
use OFS instead of Alpha-investingAlpha · CIFAR-10CPU time: 3 mins -> 2.28Improved
use OFS instead of Alpha-investingAlpha · ARCPU time: 26.99 -> 1.74Improved
use OFS instead of LevSLevS · ARCPU time: 7.76 -> 1.74Improved
use OFS instead of LapSLapS · MNISTCPU time: 19.38 -> 1.32Improved
use OFS instead of LevSLevS · USPSCPU time: 0.94 -> 0.15Improved
use OFS instead of SECSEC · all six datasetsCPU time: 0.002 to 0.005 versus 0.05 to 2.28Hurt
use OFS instead of LapS on ORLLapS · ORLCPU time: 0.24 -> 0.05Improved
use OFS instead of Alpha on ORLAlpha · ORLCPU time: 0.42 -> 0.05Improved
cluster on OFS-selected feature subset (10 to 500 features) instead of all featuresk-means on all features · COIL20, CIFAR-10, ORLclustering accuracyImproved
use OFS instead of baseline feature selection methods across 10 to 500 selected featuresLapS, LevS, Alpha, SEC · COIL20, CIFAR-10, ORLclustering accuracyImproved
process features one at a time online using a skinny sketch matrix instead of the full data matrixbatch methods · CIFAR-10CPU timeImproved

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