Interferometric Graph Transform for Community Labeling

TL;DR

This paper introduces an extended interferometric graph transform method for unsupervised community labeling in graphs, leveraging non-linear feature extraction to improve node classification accuracy on standard datasets.

Contribution

The paper extends the Interferometric Graph Transform to community labeling, introducing a non-linear operator that captures graph topology for unsupervised node representation learning.

Findings

Achieves state-of-the-art performance on Cora, Citeseer, Pubmed, and WikiCS datasets.

Demonstrates the effectiveness of the IGT extension in community detection tasks.

Shows theoretical connection between IGT and E-IGT through ergodicity properties.

Abstract

We present a new approach for learning unsupervised node representations in community graphs. We significantly extend the Interferometric Graph Transform (IGT) to community labeling: this non-linear operator iteratively extracts features that take advantage of the graph topology through demodulation operations. An unsupervised feature extraction step cascades modulus non-linearity with linear operators that aim at building relevant invariants for community labeling. Via a simplified model, we show that the IGT concentrates around the E-IGT: those two representations are related through some ergodicity properties. Experiments on community labeling tasks show that this unsupervised representation achieves performances at the level of the state of the art on the standard and challenging datasets Cora, Citeseer, Pubmed and WikiCS.