Uniendo la Academia y la Industria: Un Benchmark Integral para la Agrupación de Grafos con Atributos
Bridging Academia and Industry: A Comprehensive Benchmark for Attributed Graph Clustering
February 9, 2026
Autores: Yunhui Liu, Pengyu Qiu, Yu Xing, Yongchao Liu, Peng Du, Chuntao Hong, Jiajun Zheng, Tao Zheng, Tieke He
cs.AI
Resumen
La Agrupación de Grafos con Atributos (AGC) es una tarea fundamental no supervisada que integra la topología estructural y los atributos de nodos para descubrir patrones latentes en datos con estructura de grafo. A pesar de su importancia en aplicaciones industriales como la detección de fraude y la segmentación de usuarios, persiste una brecha significativa entre la investigación académica y el despliegue en entornos reales. Los protocolos de evaluación actuales adolecen del uso de conjuntos de datos de citas a pequeña escala y con alta homofilia, paradigmas de entrenamiento por lotes completos no escalables, y una dependencia de métricas supervisadas que no reflejan el rendimiento en entornos con escasez de etiquetas. Para salvar estas brechas, presentamos PyAGC, un benchmark y biblioteca integral listo para producción, diseñado para someter a prueba métodos de AGC en diversas escalas y propiedades estructurales. Unificamos las metodologías existentes en un marco modular de Codificar-Agrup
English
Attributed Graph Clustering (AGC) is a fundamental unsupervised task that integrates structural topology and node attributes to uncover latent patterns in graph-structured data. Despite its significance in industrial applications such as fraud detection and user segmentation, a significant chasm persists between academic research and real-world deployment. Current evaluation protocols suffer from the small-scale, high-homophily citation datasets, non-scalable full-batch training paradigms, and a reliance on supervised metrics that fail to reflect performance in label-scarce environments. To bridge these gaps, we present PyAGC, a comprehensive, production-ready benchmark and library designed to stress-test AGC methods across diverse scales and structural properties. We unify existing methodologies into a modular Encode-Cluster-Optimize framework and, for the first time, provide memory-efficient, mini-batch implementations for a wide array of state-of-the-art AGC algorithms. Our benchmark curates 12 diverse datasets, ranging from 2.7K to 111M nodes, specifically incorporating industrial graphs with complex tabular features and low homophily. Furthermore, we advocate for a holistic evaluation protocol that mandates unsupervised structural metrics and efficiency profiling alongside traditional supervised metrics. Battle-tested in high-stakes industrial workflows at Ant Group, this benchmark offers the community a robust, reproducible, and scalable platform to advance AGC research towards realistic deployment. The code and resources are publicly available via GitHub (https://github.com/Cloudy1225/PyAGC), PyPI (https://pypi.org/project/pyagc), and Documentation (https://pyagc.readthedocs.io).