ICLR 2026 Orals

One for Two: A Unified Framework for Imbalanced Graph Classification via Dynamic Balanced Prototype

Guanjun Wang, Binwu Wang, Jiaming Ma, Zhengyang Zhou, Pengkun Wang, Xu Wang, Yang Wang

Graph Learning Thu, Apr 23 · 10:30 AM–10:40 AM · 203 A/B Avg rating: 5.50 (4–6)

Abstract

Graph Neural Networks (GNNs) have advanced graph classification, yet they remain vulnerable to graph-level imbalance, encompassing class imbalance and topological imbalance. To address both types of imbalance in a unified manner, we propose UniImb, a Unified framework for Imbalanced graph classification. Specifically, UniImb first captures multi-scale topological features and enhances data diversity via learnable personalized graph perturbations. It then employs a dynamic balanced prototype module to learn representative prototypes from graph instances, improving the quality of graph representations. Concurrently, a prototype load-balancing optimization term mitigates dominance by majority samples to equalize sample influence during training. We justify these design choices theoretically using the Information Bottleneck principle. Extensive experiments on 19 datasets-including a large-scale imbalanced air pollution graph dataset AirGraph released by us and 23 baselines demonstrate that UniImb has achieved dominant performance across various imbalanced scenarios. Our code is available at GitHub.

One-sentence summary·Auto-generated by claude-haiku-4-5-20251001(?)

Unified framework for imbalanced graph classification using dynamic balanced prototypes and prototype load-balancing optimization.

Contributions·Auto-generated by claude-haiku-4-5-20251001(?)
  • Dynamic balanced prototype module to learn representative prototypes from graph instances under imbalanced settings
  • Prototype load-balancing optimization term to equalize sample influence during training
  • Theoretical justification using Information Bottleneck principle for design choices
Methods used·Auto-generated by claude-haiku-4-5-20251001(?)
  • Graph Neural Networks
  • dynamic prototypes
  • learnable graph perturbations
  • feature mixup
Datasets used·Auto-generated by claude-haiku-4-5-20251001(?)
  • AirGraph
  • 19 benchmark datasets
Limitations (author-stated)·Auto-generated by claude-haiku-4-5-20251001(?)
  • Current datasets primarily consist of homogeneous graphs, lacking diversity and complexity of real-world heterogeneous networks
    from the paper
Future work (author-stated)·Auto-generated by claude-haiku-4-5-20251001(?)
  • Evaluate performance on heterogeneous graphs with multiple types of nodes and edges
    from the paper
  • Extend comparison of imbalanced graph learning algorithms beyond classification to few-shot learning, dynamic graph learning, and anomaly detection
    from the paper

Author keywords

  • Graph classification; graph imbalance learning; graph neural networks; Graph data mining; long-tail learning

Related orals

Something off? Let us know →