Abstract
Graph neural networks (GNNs) have achieved remarkable performance in a variety of graph-related tasks. Recent evidence in the GNN community shows that such good performance can be attributed to the homophily prior; i.e., connected nodes tend to have similar features and labels. However, in heterophilic settings where the features of connected nodes may vary significantly, GNN models exhibit notable performance deterioration. In this work, we formulate this problem as prior-data conflict and propose a model called the mixture-prior graph neural network (MPGNN). First, to address the mismatch of homophily prior on heterophilic graphs, we introduce the non-informative prior, which makes no assumptions about the relationship between connected nodes and learns such relationship from the data. Second, to avoid performance degradation on homophilic graphs, we implement a soft switch to balance the effects of homophily prior and non-informative prior by learnable weights. We evaluate the performance of MPGNN on both synthetic and real-world graphs. Results show that MPGNN can effectively capture the relationship between connected nodes, while the soft switch helps select a suitable prior according to the graph characteristics. With these two designs, MPGNN outperforms state-of-the-art methods on heterophilic graphs without sacrificing performance on homophilic graphs.
摘要
图神经网络(GNN)在各种与图相关的任务中已取得显著性能。最近GNN社区的证据表明,这种良好的性能可归因于同质性先验,即连接的节点倾向于具有相似的特征和标签。然而,在异配性设置中,连接节点的特征可能会有显著变化,导致GNN模型性能明显下降。本文将此问题定义为先验数据冲突,提出一种名为混合先验图神经网络(MPGNN)的模型。首先,为解决异配图上同质性先验不匹配的问题,引入无信息先验,它不对连接节点之间的关系做任何假设,并从数据中学习这种关系。其次,为避免同质图上性能下降,通过可学习的权重实现软开关,以平衡同质性先验和非信息先验的影响。评估了MPGNN在合成图和真实世界图上的性能。结果表明,MPGNN能够有效捕捉连接节点之间的关系,而软开关有助于根据图的特征选择合适的先验。基于这两个设计,MPGNN在异配图上优于最先进的方法,而在同质图上不会牺牲性能。
Similar content being viewed by others
Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
Arpit D, Jastrzebski S, Ballas N, et al., 2017. A closer look at memorization in deep networks. Proc 34th Int Conf on Machine Learning, p.233–242.
Chiang WL, Liu XQ, Si S, et al., 2019. Cluster-GCN: an efficient algorithm for training deep and large graph convolutional networks. Proc 25th ACM SIGKDD Int Conf on Knowledge Discovery & Data Mining, p.257–266. https://doi.org/10.1145/3292500.3330925
Chien E, Peng JH, Li P, et al., 2021. Adaptive universal generalized PageRank graph neural network. Proc 9th Int Conf on Learning Representations.
Ciotti V, Bonaventura M, Nicosia V, et al., 2016. Homophily and missing links in citation networks. EPJ Data Sci, 5(1):7. https://doi.org/10.1140/epjds/s13688-016-0068-2
Dempster AP, Laird NM, Rubin DB, 1977. Maximum likelihood from incomplete data via the EM algorithm. J R Stat Soc Ser B (Methodol), 39(1):1–22. https://doi.org/10.1111/j.2517-6161.1977.tb01600.x
Deshpande Y, Montanari A, Mossel E, et al., 2018. Contextual stochastic block models. Proc 32nd Int Conf on Neural Information Processing Systems, p.8590–8602.
Dong HD, Chen JW, Feng FL, et al., 2021. On the equivalence of decoupled graph convolution network and label propagation. Proc Web Conf, p.3651–3662. https://doi.org/10.1145/3442381.3449927
Feldman V, 2020. Does learning require memorization? A short tale about a long tail. Proc 52nd Annual ACM SIGACT Symp on Theory of Computing, p.954–959. https://doi.org/10.1145/3357713.3384290
Fey M, Lenssen JE, 2019. Fast graph representation learning with PyTorch geometric. https://arxiv.org/abs/1903.02428
Grover A, Leskovec J, 2016. node2vec: scalable feature learning for networks. Proc 22nd ACM SIGKDD Int Conf on Knowledge Discovery and Data Mining, p.855–864. https://doi.org/10.1145/2939672.2939754
Hamilton WL, Ying R, Leskovec J, 2017. Inductive representation learning on large graphs. Proc 31st Int Conf on Neural Information Processing Systems, p.1025–1035.
He MG, Wei ZW, Huang ZF, et al., 2021. BernNet: learning arbitrary graph spectral filters via Bernstein approximation. Proc 35th Int Conf on Neural Information Processing Systems, p.14239–14251.
Hochreiter S, Schmidhuber J, 1997. Long short-term memory. Neur Comput, 9(8):1735–1780. https://doi.org/10.1162/neco.1997.9.8.1735
Hu WH, Fey M, Zitnik M, et al., 2020. Open graph benchmark: datasets for machine learning on graphs. Proc 34th Int Conf on Neural Information Processing Systems, Article 1855.
Huang Q, He H, Singh A, et al., 2021. Combining label propagation and simple models outperforms graph neural networks. Proc 9th Int Conf on Learning Representations.
Jeh G, Widom J, 2003. Scaling personalized web search. Proc 12th Int Conf on World Wide Web, p.271–279. https://doi.org/10.1145/775152.775191
Jin D, Yu ZZ, Huo CY, et al., 2021. Universal graph convolutional networks. Proc 35th Int Conf on Neural Information Processing Systems, p.10654–10664.
Kipf TN, Welling M, 2017. Semi-supervised classification with graph convolutional networks. Proc 5th Int Conf on Learning Representations.
Klicpera J, Bojchevski A, Günnemann S, 2019. Predict then propagate: graph neural networks meet personalized PageRank. Proc 7th Int Conf on Learning Representations.
Krizhevsky A, Sutskever I, Hinton GE, 2017. ImageNet classification with deep convolutional neural networks. Commun ACM, 60(6):84–90. https://doi.org/10.1145/3065386
Leskovec J, Krevl A, 2014. SNAP Datasets: Stanford Large Network Dataset Collection. http://snap.stanford.edu/data/
Leskovec J, Kleinberg J, Faloutsos C, 2005. Graphs over time: densification laws, shrinking diameters and possible explanations. Proc 11th ACM SIGKDD Int Conf on Knowledge Discovery in Data Mining, p.177–187. https://doi.org/10.1145/1081870.1081893
Lim D, Hohne F, Li XY, et al., 2021. Large scale learning on non-homophilous graphs: new benchmarks and strong simple methods. Proc 35th Int Conf on Neural Information Processing Systems, p.20887–20902.
Ma JX, Zhou C, Cui P, et al., 2019. Learning disentangled representations for recommendation. Proc 33rd Int Conf on Neural Information Processing Systems, Article 513.
McCallum AK, Nigam K, Rennie J, et al., 2000. Automating the construction of Internet portals with machine learning. Inform Retr, 3(2):127–163. https://doi.org/10.1023/A:1009953814988
McLachlan GJ, Krishnan T, 1997. The EM Algorithm and Extensions. John Wiley & Sons, New York, USA.
McPherson M, Smith-Lovin L, Cook JM, 2001. Birds of a feather: homophily in social networks. Ann Rev Sociol, 27:415–444. https://doi.org/10.1146/annurev.soc.27.1.415
Pei HB, Wei BZ, Chang KCC, et al., 2020. Geom-GCN: geometric graph convolutional networks. Proc 8th Int Conf on Learning Representations.
Sen P, Namata G, Bilgic M, et al., 2008. Collective classification in network data. AI Mag, 29(3):93–106. https://doi.org/10.1609/aimag.v29i3.2157
Tang J, Sun JM, Wang C, et al., 2009. Social influence analysis in large-scale networks. Proc 15th ACM SIGKDD Int Conf on Knowledge Discovery and Data Mining, p.807–816. https://doi.org/10.1145/1557019.1557108
Veličković P, Cucurull G, Casanova A, et al., 2018. Graph attention networks. Proc 6th Int Conf on Learning Representations.
Wang Z, Wang CK, Pei JS, et al., 2016. Causality based propagation history ranking in social network. Proc 25th Int Joint Conf on Artificial Intelligence, p.3917–3923.
Wu F, Souza AHJr, Zhang TY, et al., 2019. Simplifying graph convolutional networks. Proc 36th Int Conf on Machine Learning, p.6861–6871.
Wu ZH, Pan SR, Long GD, et al., 2020. Connecting the dots: multivariate time series forecasting with graph neural networks. Proc 26th ACM SIGKDD Int Conf on Knowledge Discovery & Data Mining, p.753–763. https://doi.org/10.1145/3394486.3403118
Xu K, Li CT, Tian YL, et al., 2018. Representation learning on graphs with jumping knowledge networks. Proc 35th Int Conf on Machine Learning, p.5449–5458.
Yang TM, Wang YJ, Yue ZH, et al., 2022. Graph pointer neural networks. Proc 36th AAAI Conf on Artificial Intelligence, p.8832–8839. https://doi.org/10.1609/aaai.v36i8.20864
Ying R, He RN, Chen KF, et al., 2018. Graph convolutional neural networks for web-scale recommender systems. Proc 24th ACM SIGKDD Int Conf on Knowledge Discovery & Data Mining, p.974–983. https://doi.org/10.1145/3219819.3219890
Zhang CY, Bengio S, Hardt M, et al., 2017. Understanding deep learning requires rethinking generalization. Proc 5th Int Conf on Learning Representations.
Zhang MH, Chen YX, 2018. Link prediction based on graph neural networks. Proc 32nd Int Conf on Neural Information Processing Systems, p.5171–5181.
Zhang ZW, Cui P, Zhu WW, 2022. Deep learning on graphs: asurvey. IEEE Trans Knowl Data Eng, 34(1):249–270. https://doi.org/10.1109/TKDE.2020.2981333
Zhao JL, Dong YX, Ding M, et al., 2021. Adaptive diffusion in graph neural networks. Proc 35th Int Conf on Neural Information Processing Systems, p.23321–23333.
Zhu J, Yan YJ, Zhao LX, et al., 2020. Beyond homophily in graph neural networks: current limitations and effective designs. Proc 34th Int Conf on Neural Information Processing Systems, Article 653.
Zhu J, Rossi RA, Rao A, et al., 2021. Graph neural networks with heterophily. Proc 35th AAAI Conf on Artificial Intelligence, p.11168–11176. https://doi.org/10.1609/aaai.v35i12.17332
Author information
Authors and Affiliations
Contributions
Xugang WU designed the research. Kai LU, Huijun WU, Ruibo WANG, and Xu ZHOU improved the design. Xugang WU implemented the method and drafted the paper. Huijun WU and Kai LU helped organize the paper. Xugang WU revised and finalized the paper.
Corresponding author
Ethics declarations
All the authors declare that they have no conflict of interest.
Additional information
Project supported by the National University of Defense Technology Foundation (Nos. ZK20-09 and ZK21-17), the Natural Science Foundation of Hunan Province, China (No. 2021JJ40692), and the National Key R&D Program of China (No. 2021YFB0300101)
Rights and permissions
About this article
Cite this article
Wu, X., Wu, H., Wang, R. et al. Towards adaptive graph neural networks via solving prior-data conflicts. Front Inform Technol Electron Eng 25, 369–383 (2024). https://doi.org/10.1631/FITEE.2300194
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1631/FITEE.2300194