Particle Competition and Cooperation for Robust Graph Convolutional Networks (PCC+GCN)

Forum topic · 小凯 · 2026-09-22

Summary

PCC+GCN is a hybrid framework that improves the robustness of Graph Convolutional Networks (GCNs) against label noise. It applies a Particle Competition and Cooperation (PCC) label-refinement stage before GCN training: particle domination dynamics identify suspicious labeled nodes, whose labels are then preserved, removed, or reassigned. The PCC graph can be augmented with feature-based k-nearest-neighbor edges, while the GCN trains on the original graph structure and node features. Evaluated on ten datasets from the NoisyGL benchmark under uniform, pair, random, and instance-dependent label noise, PCC+GCN achieved the best overall average accuracy and rank under conventional noise, outperforming baseline GCN by 1.67 percentage points on average across clean and noisy settings. Under instance-dependent noise it remained competitive with the most robust methods while being significantly faster, the fastest robust method on eight of ten datasets. The paper is available at arXiv:2609.22053.

Paper Overview

Field: Machine Learning
Author: Fabricio Breve
Published: 2026-09-18
arXiv: 2609.22053

Abstract

Graph Convolutional Networks (GCNs) are highly sensitive to label noise, since corrupted supervision can propagate through the graph and degrade learned node representations. This work proposes PCC+GCN, a hybrid framework that uses Particle Competition and Cooperation (PCC) as a graph-based label-refinement stage before GCN training. PCC identifies suspicious labeled nodes through particle domination dynamics and determines whether their labels should be preserved, removed, or reassigned before GCN training. The framework also allows the graph used by PCC to be augmented with feature-based $$k$$ -nearest-neighbor edges, while the GCN itself is trained on the original graph structure and node features.

The proposed method was evaluated on ten graph datasets from the NoisyGL benchmark under conventional Uniform, Pair, and Random label noise as well as instance-dependent label noise, with a detailed hyperparameter analysis on Cora, CiteSeer, and PubMed.

Key Results

Under conventional noise, PCC+GCN achieved the highest overall average accuracy and best average rank among all evaluated methods, outperforming the baseline GCN by 1.67 percentage points on average across clean and all noisy settings.
Under instance-dependent noise, PCC+GCN remained competitive with the most robust methods while requiring significantly less execution time — it was the fastest robust method on eight of the ten datasets.
The results indicate that PCC-based label refinement is an effective and computationally efficient preprocessing strategy for improving GCN robustness under noisy supervision.