NLGCL: Naturally Existing Neighbor Layers Graph Contrastive Learning for Recommendation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xu, Jinfeng, Chen, Zheyu, Yang, Shuo, Li, Jinze, Wang, Hewei, Wang, Wei, Hu, Xiping, Ngai, Edith
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912474410778624
author Xu, Jinfeng
Chen, Zheyu
Yang, Shuo
Li, Jinze
Wang, Hewei
Wang, Wei
Hu, Xiping
Ngai, Edith
author_facet Xu, Jinfeng
Chen, Zheyu
Yang, Shuo
Li, Jinze
Wang, Hewei
Wang, Wei
Hu, Xiping
Ngai, Edith
contents Graph Neural Networks (GNNs) are widely used in collaborative filtering to capture high-order user-item relationships. To address the data sparsity problem in recommendation systems, Graph Contrastive Learning (GCL) has emerged as a promising paradigm that maximizes mutual information between contrastive views. However, existing GCL methods rely on augmentation techniques that introduce semantically irrelevant noise and incur significant computational and storage costs, limiting effectiveness and efficiency. To overcome these challenges, we propose NLGCL, a novel contrastive learning framework that leverages naturally contrastive views between neighbor layers within GNNs. By treating each node and its neighbors in the next layer as positive pairs, and other nodes as negatives, NLGCL avoids augmentation-based noise while preserving semantic relevance. This paradigm eliminates costly view construction and storage, making it computationally efficient and practical for real-world scenarios. Extensive experiments on four public datasets demonstrate that NLGCL outperforms state-of-the-art baselines in effectiveness and efficiency.
format Preprint
id arxiv_https___arxiv_org_abs_2507_07522
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle NLGCL: Naturally Existing Neighbor Layers Graph Contrastive Learning for Recommendation
Xu, Jinfeng
Chen, Zheyu
Yang, Shuo
Li, Jinze
Wang, Hewei
Wang, Wei
Hu, Xiping
Ngai, Edith
Information Retrieval
Graph Neural Networks (GNNs) are widely used in collaborative filtering to capture high-order user-item relationships. To address the data sparsity problem in recommendation systems, Graph Contrastive Learning (GCL) has emerged as a promising paradigm that maximizes mutual information between contrastive views. However, existing GCL methods rely on augmentation techniques that introduce semantically irrelevant noise and incur significant computational and storage costs, limiting effectiveness and efficiency. To overcome these challenges, we propose NLGCL, a novel contrastive learning framework that leverages naturally contrastive views between neighbor layers within GNNs. By treating each node and its neighbors in the next layer as positive pairs, and other nodes as negatives, NLGCL avoids augmentation-based noise while preserving semantic relevance. This paradigm eliminates costly view construction and storage, making it computationally efficient and practical for real-world scenarios. Extensive experiments on four public datasets demonstrate that NLGCL outperforms state-of-the-art baselines in effectiveness and efficiency.
title NLGCL: Naturally Existing Neighbor Layers Graph Contrastive Learning for Recommendation
topic Information Retrieval
url https://arxiv.org/abs/2507.07522