AI Paper Atlas

Research map / Learning theory and algorithms / Clustering

A Theoretical Analysis of DeepWalk and Node2vec for Exact Recovery of Community Structures in Stochastic Blockmodels

Yichi Zhang, Minh Tang

TPAMI 2023 · Learning theory and algorithms › Clustering · approach: Correlation clustering · 34 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
increase n (1500 vs 600) for SVD-based node2vecSVD-based node2vec (t_U=5) · SBM B, rho=3n^-1/2proportion of perfect recovery: 0.30 -> 0.86Improved
increase n (1500 vs 600) for original node2vecOriginal node2vec (t_U=5) · SBM B, rho=3n^-1/2proportion of perfect recovery: 0.01 -> 0.57Improved
make graph denser (rho=3n^-1/3 instead of 3n^-1/2)SVD-based node2vec · SBM B, n=600proportion of perfect recovery: 0.30 -> 1.00Improved
increase window size t_U from 5 to 8Original node2vec · SBM B-natural, rho=3n^-1/3, n=600proportion of perfect recovery: 1.00 -> 0.45Hurt
use SVD-based instead of SGD-based node2vecnode2vec · SBM B-natural, rho=3n^-1/2, n=900proportion of perfect recovery: 0.00 -> 0.58Improved
replace K-means with GMM clustering on original node2vec embeddings (ablation)Original node2vec · SBM B1 n=600, rho=3n^-1/2clustering accuracy: 0.58 -> 0.84Improved
increase t_U from 5 to 6 for SVD-based node2vecSVD-based node2vec · SBM B1 n=600, rho=3n^-1/2accuracy: 0.61 -> 0.87Improved
use node2vec instead of ASEnode2vec (SVD-based) · PB (political blogs)accuracy: 0.61 -> 0.96Improved
use node2vec instead of LSEnode2vec (SVD-based) · PB (political blogs)accuracy: 0.51 -> 0.96Improved
use node2vec instead of ASEnode2vec (SVD-based) · ZK (Zachary karate)accuracy: 1.00 -> 0.97Hurt
use node2vec instead of ASEnode2vec (SVD-based) · WIKIARI: 0.03 -> 0.09Improved

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