Graph Neural Networks for Relational Learning and Recommender Systems: Architectures and Emerging Applications
Keywords:
Graph neural network, graph representation learning, relational learningAbstract
Graphs provide a natural representation for systems in which entities interact through complex relationships, including social networks, citation networks, molecular structures, knowledge graphs, communication systems, and recommender platforms. Conventional deep-learning methods were largely developed for regularly structured Euclidean data and are therefore not directly suited to graph-structured information. Graph neural networks (GNNs) address this limitation by learning representations through aggregation and propagation of information across nodes and edges. This review examines major GNN developments available through 2020, beginning with early recurrent graph models and progressing to spectral and spatial graph convolution, message-passing neural networks, graph attention networks, inductive representation learning, relational graph convolution, and graph autoencoders. The application of graph learning to recommender systems is examined through GraphSAGE, PinSage, neural graph collaborative filtering, knowledge-graph attention networks, and LightGCN. Particular attention is given to neighborhood aggregation, higher-order connectivity, graph sparsity, scalability, over-smoothing, computational complexity, cold-start behavior, and the integration of side information. The review demonstrates how graph-based learning extends conventional collaborative filtering by directly modeling structural dependencies between users, items, and auxiliary entities. Future research needs identified in the 2020 landscape include scalable graph sampling, dynamic graph learning, interpretability, heterogeneous graphs, and robust representation learning.
References
Scarselli F, Gori M, Tsoi AC, Hagenbuchner M, Monfardini G. The graph neural network model. IEEE Trans Neural Netw. 2009;20(1):61-80.
Gori M, Monfardini G, Scarselli F. A new model for learning in graph domains. In: International Joint Conference on Neural Networks. Piscataway: IEEE; 2005. p. 729-734.
Bruna J, Zaremba W, Szlam A, LeCun Y. Spectral networks and locally connected networks on graphs. In: International Conference on Learning Representations. 2014.
Defferrard M, Bresson X, Vandergheynst P. Convolutional neural networks on graphs with fast localized spectral filtering. Adv Neural Inf Process Syst. 2016;29:3844-3852.
Kipf TN, Welling M. Semi-supervised classification with graph convolutional networks. In: International Conference on Learning Representations. 2017.
Hamilton WL, Ying R, Leskovec J. Inductive representation learning on large graphs. Adv Neural Inf Process Syst. 2017;30:1024-1034.
Veličković P, Cucurull G, Casanova A, Romero A, Liò P, Bengio Y. Graph attention networks. In: International Conference on Learning Representations. 2018.
Gilmer J, Schoenholz SS, Riley PF, Vinyals O, Dahl GE. Neural message passing for quantum chemistry. Proc Mach Learn Res. 2017;70:1263-1272.
Xu K, Hu W, Leskovec J, Jegelka S. How powerful are graph neural networks? In: International Conference on Learning Representations. 2019.
Schlichtkrull M, Kipf TN, Bloem P, van den Berg R, Titov I, Welling M. Modeling relational data with graph convolutional networks. In: European Semantic Web Conference. Cham: Springer; 2018. p. 593-607.
Kipf TN, Welling M. Variational graph auto-encoders. arXiv. 2016;1611.07308.
Zhang M, Chen Y. Link prediction based on graph neural networks. Adv Neural Inf Process Syst. 2018;31:5165-5175.
Ying R, He R, Chen K, Eksombatchai P, Hamilton WL, Leskovec J. Graph convolutional neural networks for web-scale recommender systems. In: Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2018. p. 974-983.
van den Berg R, Kipf TN, Welling M. Graph convolutional matrix completion. arXiv. 2017;1706.02263.
Wang X, He X, Wang M, Feng F, Chua TS. Neural graph collaborative filtering. In: Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval. 2019. p. 165-174.
Wang X, He X, Cao Y, Liu M, Chua TS. KGAT: knowledge graph attention network for recommendation. In: Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2019. p. 950-958.
He X, Deng K, Wang X, Li Y, Zhang Y, Wang M. LightGCN: simplifying and powering graph convolution network for recommendation. In: Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval. 2020. p. 639-648.
Wu F, Souza A, Zhang T, Fifty C, Yu T, Weinberger KQ. Simplifying graph convolutional networks. Proc Mach Learn Res. 2019;97:6861-6871.
Perozzi B, Al-Rfou R, Skiena S. DeepWalk: online learning of social representations. In: Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2014. p. 701-710.
Grover A, Leskovec J. node2vec: scalable feature learning for networks. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2016. p. 855-864.
Battaglia PW, Hamrick JB, Bapst V, Sanchez-Gonzalez A, Zambaldi V, Malinowski M, et al. Relational inductive biases, deep learning, and graph networks. arXiv. 2018;1806.01261.
Bacciu D, Errica F, Micheli A, Podda M. A gentle introduction to deep learning for graphs. Neural Netw. 2020;129:203-221.
