Khan-GCL: Kolmogorov-Arnold Network Based Graph Contrastive Learning with Hard Negatives

2025-05-21

AI Summary

Traditional graph contrastive learning (GCL) suffers from two key bottlenecks: limited expressiveness of MLP-based encoders and semantically uninformative negative samples—random augmentations rarely yield “hard negatives,” while existing hard-negative sampling strategies ignore structural semantic disparities. To address these, this work introduces the Kolmogorov–Arnold Network (KAN) as the GCL encoder for the first time, leveraging its superior nonlinear modeling capacity and interpretable learnable coefficients. Building upon KAN’s coefficient structure, we propose a dual-path key feature identification mechanism that explicitly captures semantic distinctions across graphs, enabling semantics-driven hard negative generation. Evaluated on multiple benchmark datasets, our method achieves state-of-the-art performance in both graph classification and cross-domain transfer tasks, with substantial gains in generalization capability and model interpretability.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Computer Vision: Generative Adversarial Networks (GANs) for Vision

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Graph contrastive learning (GCL) has demonstrated great promise for learning generalizable graph representations from unlabeled data. However, conventional GCL approaches face two critical limitations: (1) the restricted expressive capacity of multilayer perceptron (MLP) based encoders, and (2) suboptimal negative samples that either from random augmentations-failing to provide effective 'hard negatives'-or generated hard negatives without addressing the semantic distinctions crucial for discriminating graph data. To this end, we propose Khan-GCL, a novel framework that integrates the Kolmogorov-Arnold Network (KAN) into the GCL encoder architecture, substantially enhancing its representational capacity. Furthermore, we exploit the rich information embedded within KAN coefficient parameters to develop two novel critical feature identification techniques that enable the generation of semantically meaningful hard negative samples for each graph representation. These strategically constructed hard negatives guide the encoder to learn more discriminative features by emphasizing critical semantic differences between graphs. Extensive experiments demonstrate that our approach achieves state-of-the-art performance compared to existing GCL methods across a variety of datasets and tasks.

Problem

• Enhancing GCL encoder capacity beyond MLP limitations
• Generating semantically meaningful hard negative samples
• Improving graph representation discriminative feature learning

Innovation

• Uses Kolmogorov-Arnold Network for GCL encoder
• Develops feature identification for hard negatives
• Enhances semantic discrimination in graph representations