·22· 智能系统学报 第15卷 越受到学术界和产业界的关注。本文深入分析 graph convolutional networks with variance reduction[C// 了GNN推荐及其过程，对现有GNN推荐相关研 Proceedings of the 35th International Conference on Ma- 究从无向单元图推荐、无向二元图推荐、无向多 chine Learning.Stockholm,Sweden,2018:941-949. 元图推荐3个方面进行详细探讨，对各自的优点 [11]LI Q,HAN Z,WU X M.Deeper insights into graph con- 与不足进行了分析，指出了GNN推荐面临的难 volutional networks for semi-supervised learning[Cl// 点及未来的研究方向，对基于图的深度推荐系统 Thirty-Second AAAI Conference on Artificial Intelli- 进一步研究具有一定的借鉴意义。 gence.New Orleans,USA.2018. [12]BRUNA J.ZAREMBA W.SZLAM A,et al.Spectral net- 参考文献： works and locally connected networks on graphs[C]//In- [1]ZHOU J,CUI G,ZHANG Z,et al.Graph neural networks: ternational Conference on Learning Representations a review of methods and applications[J].arXiv:1812. (ICLR2014).Banff,Canada,2014:1-14. 08434.2018 [13]MONTI F.BOSCAINI D.MASCI J.et al.Geometric [2]BASTINGS J,TITOV I,AZIZ W,et al.Graph convolu- deep learning on graphs and manifolds using mixture tional encoders for syntax-aware neural machine transla- model CNNs[C]//Proceedings of 2017 IEEE Conference tion[C]//Proceedings of the 2017 Conference on Empirical on Computer Vision and Pattern Recognition.Honolulu, Methods in Natural Language Processing.Copenhagen, USA.2017:5115-5124 Denmark.2017:1957-1967. [14]ATWOOD J,TOWSLEY D.Diffusion-convolutional [3]HENAFF M,BRUNA J,LECUN Y.Deep convolutional neural networks[Cl/Advances in Neural Information Pro- networks on graph-structured data[J].arXiv:1506.05163, cessing Systems.Barcelona,Spain,2016:1993-2001. 2015. [15]LI Y,ZEMEL R,BROCKSCHMIDT M,et al.Gated [4]ZHANG Yuhao,QI Peng,MANNING C D.Graph convo- graph sequence neural networks [J].arXiv:1511.05493, lution over pruned dependency trees improves relation ex- 2015. traction[C]//Proceedings of the 2018 Conference on Empir- [16]ZHANG Yue,LIU Qi,SONG Linfeng.Sentence-state ical Methods in Natural Language Processing.Brussels, LSTM for text representation[Cl//Proceedings of the 56th Belgium,2018:2205-2215. Annual Meeting of the Association for Computational [5]WANG Xiaolong,YE Yufei,GUPTA A.Zero-shot recog- Linguistics.Melbourne,Australia,2018:317-327. nition via semantic embeddings and knowledge [17]KAMPFFMEYER M,CHEN Yinbo,LIANG Xiaodan,et graphs[C]//Proceedings of 2018 IEEE/CVF Conference on al.Rethinking knowledge graph propagation for zero-shot Computer Vision and Pattern Recognition.Salt Lake City, learning[C]//Proceedings of 2019 IEEE/CVF Conference USA.2018:6857-6866. on Computer Vision and Pattern Recognition.Long [6]RHEE S,SEO S,KIM S.Hybrid approach of relation net- Beach,USA,2019:11487-11496. work and localized graph convolutional filtering for breast [18]黄璐，林川杰，何军，等.融合主题模型和协同过滤的多 cancer subtype classification[J].arXiv:1711.058592017. 样化移动应用推荐.软件学报，2017,28(3)：708-720. [7]KAWAMOTO T.TSUBAKI M.OBUCHI T.Mean-field HUANG Lu,LIN Chuanjie,HE Jun,et al.Diversified theory of graph neural networks in graph partitioning[C]// mobile app recommendation combining topic model and Proceedings of the 32nd International Conference on Neur- collaborative filtering[J].Journal of software,2017,28(3): al Information Processing Systems.Red Hook,USA,2018: 708-720. 4361-4371. [19]胡堰，彭启民，胡晓惠.一种基于隐语义概率模型的个 [8]HAMILTON W.YING Zhotao,LESKOVEC J.Inductive 性化Web服务推荐方法[U.计算机研究与发展，2014 representation learning on large graphs[C]//Advances in 51(8):1781-1793 Neural Information Processing Systems.Long Beach,US, HU Yan,PENG Qimin,HU Xiaohui.A personalized Web 2017:1024-1034 service recommendation method based on latent semantic [9]CHEN J,MA T,XIAO C.Fastgcn:fast learning with probabilistic model[J].Journal of computer research and graph convolutional networks via importance sampling[J]. development,.2014,51(8):1781-1793. arXiv:1801.10247,2018. [20]刘建勋，石敏，周栋，等.基于主题模型的Mashup标签 [10]CHEN Jianfei.ZHU Jun,SONG Le.Stochastic training of 推荐方法[U.计算机学报，2017,40(2)520-534.越受到学术界和产业界的关注。本文深入分析 了 GNN 推荐及其过程，对现有 GNN 推荐相关研 究从无向单元图推荐、无向二元图推荐、无向多 元图推荐 3 个方面进行详细探讨，对各自的优点 与不足进行了分析，指出了 GNN 推荐面临的难 点及未来的研究方向，对基于图的深度推荐系统 进一步研究具有一定的借鉴意义。 参考文献： ZHOU J, CUI G, ZHANG Z, et al. Graph neural networks: a review of methods and applications[J]. arXiv: 1812. 08434, 2018. [1] BASTINGS J, TITOV I, AZIZ W, et al. Graph convolu￾tional encoders for syntax-aware neural machine transla￾tion[C]//Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing. Copenhagen, Denmark, 2017: 1957−1967. [2] HENAFF M, BRUNA J, LECUN Y. Deep convolutional networks on graph-structured data[J]. arXiv:1506.05163, 2015. [3] ZHANG Yuhao, QI Peng, MANNING C D. Graph convo￾lution over pruned dependency trees improves relation ex￾traction[C]//Proceedings of the 2018 Conference on Empir￾ical Methods in Natural Language Processing. Brussels, Belgium, 2018: 2205−2215. [4] WANG Xiaolong, YE Yufei, GUPTA A. Zero-shot recog￾nition via semantic embeddings and knowledge graphs[C]//Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, USA, 2018: 6857−6866. [5] RHEE S,SEO S, KIM S.Hybrid approach of relation net￾work and localized graph convolutional filtering for breast cancer subtype classification[J].arXiv: 1711.05859,2017. [6] KAWAMOTO T, TSUBAKI M, OBUCHI T. Mean-field theory of graph neural networks in graph partitioning[C]// Proceedings of the 32nd International Conference on Neur￾al Information Processing Systems. Red Hook, USA, 2018: 4361−4371. [7] HAMILTON W, YING Zhotao, LESKOVEC J. Inductive representation learning on large graphs[C]//Advances in Neural Information Processing Systems. Long Beach, US, 2017: 1024−1034. [8] CHEN J, MA T, XIAO C. Fastgcn: fast learning with graph convolutional networks via importance sampling[J]. arXiv:1801.10247, 2018. [9] [10] CHEN Jianfei, ZHU Jun, SONG Le. Stochastic training of graph convolutional networks with variance reduction[C]// Proceedings of the 35th International Conference on Ma￾chine Learning. Stockholm, Sweden, 2018: 941−949. LI Q, HAN Z, WU X M. Deeper insights into graph con￾volutional networks for semi-supervised learning[C]// Thirty-Second AAAI Conference on Artificial Intelli￾gence. New Orleans, USA,2018. [11] BRUNA J, ZAREMBA W, SZLAM A, et al. Spectral net￾works and locally connected networks on graphs[C]//In￾ternational Conference on Learning Representations (ICLR2014). Banff, Canada, 2014: 1−14. [12] MONTI F, BOSCAINI D, MASCI J, et al. Geometric deep learning on graphs and manifolds using mixture model CNNs[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, USA, 2017: 5115−5124. [13] ATWOOD J, TOWSLEY D. Diffusion-convolutional neural networks[C]//Advances in Neural Information Pro￾cessing Systems. Barcelona, Spain, 2016: 1993−2001. [14] LI Y, ZEMEL R, BROCKSCHMIDT M, et al. Gated graph sequence neural networks [J]. arXiv:1511.05493, 2015. [15] ZHANG Yue, LIU Qi, SONG Linfeng. Sentence-state LSTM for text representation[C]//Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics. Melbourne, Australia, 2018: 317−327. [16] KAMPFFMEYER M, CHEN Yinbo, LIANG Xiaodan, et al. Rethinking knowledge graph propagation for zero-shot learning[C]//Proceedings of 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Long Beach, USA, 2019: 11487−11496. [17] 黄璐, 林川杰, 何军, 等. 融合主题模型和协同过滤的多 样化移动应用推荐 [J]. 软件学报, 2017, 28(3): 708–720. HUANG Lu, LIN Chuanjie, HE Jun, et al. Diversified mobile app recommendation combining topic model and collaborative filtering[J]. Journal of software, 2017, 28(3): 708–720. [18] 胡堰, 彭启民, 胡晓惠. 一种基于隐语义概率模型的个 性化 Web 服务推荐方法 [J]. 计算机研究与发展, 2014, 51(8): 1781–1793. HU Yan, PENG Qimin, HU Xiaohui. A personalized Web service recommendation method based on latent semantic probabilistic model[J]. Journal of computer research and development, 2014, 51(8): 1781–1793. [19] 刘建勋, 石敏, 周栋, 等. 基于主题模型的 Mashup 标签 推荐方法 [J]. 计算机学报, 2017, 40(2): 520–534. [20] ·22· 智 能 系 统 学 报 第 15 卷