arXiv:2605.18387v1 Announce Type: new 
Abstract: Graph Neural Networks (GNNs) and Graph Transformers (GTs) are now a fundamental paradigm for graph learning, combining the representation-learning capabilities of deep models with the sample efficiency induced by their inductive biases. Despite their effectiveness, a large body of work has shown that these models still face fundamental limitations in tasks that require capturing correlations between distant regions of a graph. To address this issue, we introduce Graph Hierarchical Recurrence (GHR), a novel framework that operates jointly on the input graph and on a hierarchical abstraction obtained through pooling. We also show that the limitations of existing models are even more pronounced in out-of-range generalization, where test instances involve interactions over distances longer than those observed during training. By contrast, despite its simple design, GHR provides three key advantages: strong performance on long-range dependencies, improved out-of-range generalization, and high parameter efficiency. To corroborate these claims, we show that across a broad set of long-range benchmarks, GHR consistently outperforms existing graph models while using as little as 1% of the parameters of current state-of-the-art models. These results suggest a complementary direction to the current trend of scaling architectures to obtain graph foundation models, indicating that increased model capacity alone may not be sufficient for generalization.

تم تقديم إطار جديد يسمى Graph Hierarchical Recurrence (GHR) لتحسين قدرات الشبكات العصبية الرسومية (GNNs) والمحولات الرسومية (GTs) في التقاط الاعتمادات على المدى الطويل داخل الرسوم البيانية. يعمل هذا الإطار على كل من الرسم البياني المدخل وملخص هرمي تم الحصول عليه من خلال التجميع، مما يعالج القيود المفروضة على النماذج الحالية، خاصة في مهام التعميم خارج النطاق.

Se ha introducido un nuevo marco llamado Graph Hierarchical Recurrence (GHR) para mejorar las capacidades de las Redes Neuronales de Grafos (GNNs) y los Transformadores de Grafos (GTs) en la captura de dependencias a largo plazo dentro de los grafos. Este marco opera tanto en el grafo de entrada como en una abstracción jerárquica derivada a través de pooling, abordando las limitaciones de los modelos existentes, especialmente en tareas de generalización fuera de rango.

Un nouveau cadre appelé Graph Hierarchical Recurrence (GHR) a été introduit pour améliorer les capacités des Graph Neural Networks (GNNs) et des Graph Transformers (GTs) à capturer les dépendances à long terme au sein des graphes. Ce cadre fonctionne à la fois sur le graphe d'entrée et sur une abstraction hiérarchique dérivée par le biais de pooling, abordant les limitations des modèles existants, en particulier dans les tâches de généralisation hors de portée.

A novel framework called Graph Hierarchical Recurrence (GHR) has been introduced to enhance the capabilities of Graph Neural Networks (GNNs) and Graph Transformers (GTs) in capturing long-range dependencies within graphs. This framework operates on both the input graph and a hierarchical abstraction derived through pooling, addressing limitations in existing models, particularly in out-of-range generalization tasks.