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 2402.09711cs.LGcs.SI
收藏

节点复制改善冷启动链路预测

Node Duplication Improves Cold-start Link Prediction

University of Washington(华盛顿大学);Snap Inc.(Snap公司);University of Notre Dame(诺特丹大学);University of California, Riverside(加州大学河滨分校)

Zhichun Guo, Tong Zhao, Yozen Liu, Kaiwen Dong, William Shiao, Mingxuan Ju, Neil Shah, Nitesh V. Chawla

AI总结:

针对图神经网络在链路预测中低度节点(冷启动场景)性能不佳的问题,提出即插即用的NodeDup节点复制增强技术,通过多视图视角提升性能,且不损害高度节点表现,实验验证了显著效果。

摘要:

图神经网络(GNNs)是图机器学习领域的主流方法,在链路预测(LP)任务中展现出当前最优的性能。然而,近期研究表明,尽管GNNs整体表现强劲,但在处理低度节点时难以取得良好效果。在链路预测的实际应用(如推荐系统)中,提升低度节点上的性能至关重要,因为这相当于解决冷启动问题——即改善仅有少量观测交互的用户的体验。本文研究在保留GNNs在高度节点上性能的同时,提升其在低度节点上的链路预测性能,并提出了一种简单却效果惊人的增强技术,名为NodeDup。具体而言,NodeDup会复制低度节点,并在遵循标准有监督链路预测训练流程之前,在节点与其自身的副本之间创建连接。通过为低度节点引入“多视图”视角,NodeDup在不损害高度节点性能的前提下,显著提升了低度节点上的链路预测性能。此外,作为一个即插即用的增强模块,NodeDup可轻松应用于现有GNNs,且计算成本极低。大量实验表明,与GNNs及现有最优冷启动方法相比,NodeDup在所有数据集上的孤立节点、低度节点和暖启动节点上的性能平均分别提升了38.49%、13.34%和6.76%。

英文摘要:

Graph Neural Networks (GNNs) are prominent in graph machine learning and have shown state-of-the-art performance in Link Prediction (LP) tasks. Nonetheless, recent studies show that GNNs struggle to produce good results on low-degree nodes despite their overall strong performance. In practical applications of LP, like recommendation systems, improving performance on low-degree nodes is critical, as it amounts to tackling the cold-start problem of improving the experiences of users with few observed interactions. In this paper, we investigate improving GNNs' LP performance on low-degree nodes while preserving their performance on high-degree nodes and propose a simple yet surprisingly effective augmentation technique called NodeDup. Specifically, NodeDup duplicates low-degree nodes and creates links between nodes and their own duplicates before following the standard supervised LP training scheme. By leveraging a ''multi-view'' perspective for low-degree nodes, NodeDup shows significant LP performance improvements on low-degree nodes without compromising any performance on high-degree nodes. Additionally, as a plug-and-play augmentation module, NodeDup can be easily applied to existing GNNs with very light computational cost. Extensive experiments show that NodeDup achieves 38.49%, 13.34%, and 6.76% improvements on isolated, low-degree, and warm nodes, respectively, on average across all datasets compared to GNNs and state-of-the-art cold-start methods.

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。

↑