AI Paper Atlas

Research map / Graph learning / Graph algorithms and clustering

Stars: Tera-Scale Graph Building for Clustering and Learning

CJ Carey, Jonathan Halcrow, Rajesh Jayaram, Vahab Mirrokni, Warren Schudy, Peilin Zhong

NeurIPS 2022 · Graph learning › Graph algorithms and clustering · approach: Neural and local graph clustering · 1 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 Stars star-graph construction instead of all-pairs comparisons within LSH bucketsLSH graph building · All evaluated datasetspairwise similarity comparisons: at least a ∼10-fold improvementImproved
Use LSH+Stars instead of non-Stars LSHAmazon2m graph building · Amazon2mtotal running time: 10× fasterImproved
Use Stars-based graph building with neural-network similarity rather than non-Stars graph buildingAmazon2m · Amazon2mVMeasure clustering qualityNo clear effect
Use LSH+Stars instead of LSHGraph building · Random10Btotal running time: 100×Improved
Use SortingLSH+Stars instead of SortingLSHGraph building · Random10Btotal running time: 10×Improved
Use Stars instead of non-Stars LSH or SortingLSHGraph building · Random10Bpairwise similarity comparisons: more than 95%Improved
Reduce the number of Stars leaders to 1 (ablation)Stars · Wikipedia, Amazon2m, MNISTpairwise similarity comparisons: ∼ 100-fold improvementImproved
Reduce the number of Stars leaders to 5 (ablation)Stars · Wikipedia, Amazon2m, MNISTpairwise similarity comparisons: ∼ 100-fold improvementImproved
Use neural-network learned similarity instead of the cosine/Jaccard mixtureAmazon2m graph building and clustering · Amazon2mVMeasure clustering qualityImproved
Use neural-network learned similarity instead of the cosine/Jaccard mixtureAmazon2m graph building · Amazon2mtotal running time: 5×∼ 10× slowerHurt

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