AI Paper Atlas

Research map / Learning theory and algorithms / Clustering

Parametrized Power-Iteration Clustering for Directed Graphs

Gwendal Debaussart-Joniec, Harry Sevi, Matthieu Jonckheere, Argyris Kalogeratos

ICML 2026 · Learning theory and algorithms › Clustering · approach: Spectral clustering algorithms

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 parametrized reversible operator instead of symmetrized walk (S-PIC)power-iteration clustering · DiSBM-ChainAMI: 0.57 -> 0.93Improved
use parametrized operator instead of symmetrized walk (S-PIC)power-iteration clustering · DiSBM-CPAMI: 0.59 -> 1.00Improved
use parametrized operator instead of symmetrized walk (S-PIC)power-iteration clustering · PolBlogsAMI: 0.01 -> 0.39Improved
use parametrized operator instead of natural random walk (PIC)power-iteration clustering · DiSBM-ChainAMI: 0.28 -> 0.93Improved
use parametrized operator instead of natural walk (PIC)power-iteration clustering · Email-EuAMI: 0.01 -> 0.48Improved
use parametrized operator instead of symmetrized walk (S-PIC)power-iteration clustering · Email-EuAMI: 0.49 -> 0.48No clear effect
use parametrized operator instead of symmetrized walk (S-PIC)power-iteration clustering · K-NN digraphs (10 datasets)PRB: 0.96 -> 0.96No clear effect
use parametrized operator instead of symmetrized walk (S-PIC)power-iteration clustering · Directed networks (5 graphs)PRB: 0.65 -> 1.00Improved
vertex measure parameter gamma set to 0 instead of at least 0.25 (ablation)ParPIC · DiSBM-ChainAMI: gamma >= 0.25 perfect clustering -> 0 AMI at gamma=0Hurt
increase vertex measure parameter gamma close to 1 (ablation)ParPIC · K-NN digraphs (Iris, Seeds, Vertebral, Control Chart, Glass)AMIHurt
lower gamma on Email-Eu (ablation)ParPIC · Email-EuAMIImproved
select diffusion time by entropy instead of acceleration-based criterion (ablation)ParPIC · diffusion time selectionAMIImproved

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