arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2402.08583cs.LG
收藏

图上链接预测器的混合模型

Mixture of Link Predictors on Graphs

Shanghai Jiao Tong University(上海交通大学);Michigan State University(密歇根州立大学)

Li Ma, Haoyu Han, Juanhui Li, Harry Shomer, Hui Liu, Xiaofeng Gao, Jiliang Tang

AI总结:

本文提出专家混合模型 Link-MoE,通过为不同节点对动态选择最合适的 GNN 专家和成对信息,显著提升链接预测性能。

摘要:

链接预测旨在预测图中未见连接,是图机器学习中的一项基础任务。启发式方法利用一系列不同的成对度量,如共同邻居和最短路径,其性能常常可与普通图神经网络(GNN)相媲美。因此,近期面向链接预测的图神经网络(GNN4LP)进展主要集中在整合一种或几种类型的成对信息上。在本工作中,我们揭示了同一数据集中不同的节点对需要不同的成对信息才能实现准确预测,而仅统一应用相同成对信息的模型可能只能达到次优性能。为此,我们提出了一种简单的专家混合模型 Link-MoE 用于链接预测。Link-MoE 利用多种 GNN 作为专家,并根据各种类型的成对信息为每个节点对策略性地选择合适的专家。在多个真实世界数据集上的实验结果表明,Link-MoE 带来了显著的性能提升。值得注意的是,与最佳基线相比,Link-MoE 在 Pubmed 数据集上的 MRR 指标相对提升了 18.71%,在 ogbl-ppa 数据集上的 Hits@100 指标相对提升了 9.59%。

英文摘要:

Link prediction, which aims to forecast unseen connections in graphs, is a fundamental task in graph machine learning. Heuristic methods, leveraging a range of different pairwise measures such as common neighbors and shortest paths, often rival the performance of vanilla Graph Neural Networks (GNNs). Therefore, recent advancements in GNNs for link prediction (GNN4LP) have primarily focused on integrating one or a few types of pairwise information. In this work, we reveal that different node pairs within the same dataset necessitate varied pairwise information for accurate prediction and models that only apply the same pairwise information uniformly could achieve suboptimal performance. As a result, we propose a simple mixture of experts model Link-MoE for link prediction. Link-MoE utilizes various GNNs as experts and strategically selects the appropriate expert for each node pair based on various types of pairwise information. Experimental results across diverse real-world datasets demonstrate substantial performance improvement from Link-MoE. Notably, Link-MoE achieves a relative improvement of 18.71\% on the MRR metric for the Pubmed dataset and 9.59\% on the Hits@100 metric for the ogbl-ppa dataset, compared to the best baselines.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑