Citation Recommendation via Hierarchical Attributed Network Representation Learning

doi:10.3778/j.issn.1673-9418.2006066

Abstract

Abstract:

Citation recommendation (CR) is able to intelligently generate a paper list related to a query paper, which is of great value to researches. CR problem is related to papers?? semantic and structural information. Recently, network representation learning (NRL) based CR problem has gained extensive attention. However, existing studies all use single granularity networks to model CR, which have the disadvantages of high computational complexity and large memory consumption. For overcoming this challenge, this paper proposes a citation recommendation algorithm based on hierarchical attributed network representation learning (CR-HANRSL), which can greatly improve the efficiency of NRL while taking into account semantic and structural features of papers. First, the network is coarsened into a series of smaller networks based on papers?? attributes and author relationship repeatedly, and the super-nodes are made to contain child-nodes?? attributes after coarsening for further constructing semantic links. Second, this paper uses the single granularity NRL to learn the roughened network and graph convolutional network to refine the learned network representations. Finally, a multi-modal feature representation similarity between papers is generated to get a paper recommendation list. A mass of experimental results on AAN and DBLP two datasets show that the proposed method can improve NRL??s efficiency while learning high-quality network feature representations.

Key words: citation recommendation (CR), hierarchical attributed network, semantic links, representation learning

摘要：

引文推荐（CR）聚焦于智能化地产生与查询文章相关的文献列表，对科学研究具有重要价值。引文推荐有关于文章的语义信息和结构信息，近年来，基于网络表示学习（NRL）的引文推荐获得广泛关注。但现有研究使用单粒度网络来建模引文推荐问题，存在计算复杂度高、内存消耗大的弊端。为克服这个挑战，提出一种基于多粒度属性网络表示学习的引文推荐算法（CR-HANRSL），可以大大提升网络表示学习效率并同时兼顾文章的语义和结构特征。首先，根据文章结点属性的语义关联度和作者关系反复将网络粗化成更小的网络，并在每次粗化后都让超结点融合子结点的文本属性为粗化后的网络计算语义连边。随后，利用单粒度网络表示学习方法学习粗化后的网络特征表示并通过学习图卷积神经网络，对原网络的表示进行细化。最后，融合文章间多模态特征表示相似度产生推荐列表。在AAN和DBLP两个数据集上的实验结果表明提出的方法可以在学习高质量网络特征表示的前提下大大提升网络表示学习效率。

关键词: 引文推荐, 多粒度属性网络, 语言连边, 表示学习

CHEN Jie, LIU Yang, ZHAO Shu, ZHANG Yanping. Citation Recommendation via Hierarchical Attributed Network Representation Learning[J]. Journal of Frontiers of Computer Science and Technology, 2021, 15(6): 1103-1113.

陈洁, 刘洋, 赵姝, 张燕平. 利用多粒度属性网络表示学习进行引文推荐[J]. 计算机科学与探索, 2021, 15(6): 1103-1113.

References

[1] CAI X Y, ZHENG Y, YANG L B, et al. Bibliographic network representation based personalized citation recommendation[J]. IEEE Access, 2018, 7: 457-467.
[2] MCNEE S M, ALBERT I, COSLEY D, et al. On the recom-mending of citations for research papers[C]//Proceedings of the 2002 ACM Conference on Computer Supported Coopera-tive Work, New Orleans, Nov 16-20, 2002. New York: ACM, 2002: 116-125.
[3] LIU H, KONG X, BAI X, et al. Context-based collaborative filtering for citation recommendation[J]. IEEE Access, 2015, 3: 1695-1703.
[4] BHAGAVATULA C, FELDMAN S. Content-based citation recommendation[J]. arXiv:1802.08301, 2018.
[5] TANG J, ZHANG J. A discriminative approach to topic-based citation recommendation[C]//LNCS 5476: Proceedings of the 13th Pacific-Asia Conference of Advances in Knowledge Discovery and Data Mining, Bangkok, Apr 27-30, 2009. Berlin, Heidelberg: Springer, 2009: 572-579.
[6] WANG J G, SONG D D, ZHANG Z W, et al. LDTM: a latent document type model for cumulative citation recom-mendation[C]//Proceedings of the 2015 Conference on Em-pirical Methods in Natural Language Processing, Lisbon, Sep 17-21, 2015. Stroudsburg: ACL, 2015: 561-566.
[7] CHEN X, ZHAO H J, ZHAO S, et al. Citation recommenda-tion based on citation tendency[J]. Scientometrics, 2019, 121(2): 937-956.
[8] CHEN J, LIU Y, ZHAO S, et al. Citation recommendation based on weighted heterogeneous information network con-taining semantic linking[C]//Proceedings of the 2019 IEEE International Conference on Multimedia and Expo, Shanghai, Jul 8-12, 2019. Piscataway: IEEE, 2019: 31-36.
[9] CAI X, HAN J, YANG L. Generative adversarial network based heterogeneous bibliographic network representation for personalized citation recommendation[C]//Proceedings of the 32nd AAAI Conference on Artificial Intelligence, New Or-leans, Feb 2-7, 2018. Menlo Park: AAAI, 2018: 5747-5754.
[10] YANG L, ZHANG Z, CAI X, et al. Citation recommenda-tion as edge prediction in heterogeneous bibliographic net-work: a network representation approach[J]. IEEE Access, 2019, 7: 23232-23239.
[11] LIU M, LIU J, CHEN Y, et al. AHNG: representation learn-ing on attributed heterogeneous network[J]. Information Fusion, 2019, 50: 221-230.
[12] ZHANG Y M, FAN Y J, YE Y F, et al. Key player identifi-cation in underground forums over attributed heterogeneous information network embedding framework[C]//Proceedings of the 28th ACM International Conference on Information and Knowledge Management, Beijing, Nov 3-7, 2019. New York: ACM, 2019: 549-558.
[13] JIANG Z, YIN Y, GAO L, et al. Cross-language citation re-commendation via hierarchical representation learning on heterogeneous graph[C]//Proceedings of the 41st International SIGIR Conference on Research & Development in Inform-ation Retrieval, Ann Arbor, 2018. New York: ACM, 2018: 635-644.
[14] YANG C, WEI B, WU J, et al. CARES: a ranking-oriented CADAL recommender system[C]//Proceedings of the 9th ACM/IEEE-CS Joint Conference on Digital Libraries, Austin, May 12-14, 2009. New York: ACM, 2009: 203-212.
[15] KANG Z, PENG C, CHENG Q. Top-n recommender system via matrix completion[C]//Proceedings of the 30th AAAI Conference on Artificial Intelligence, Phoenix, Feb 12-17, 2016. Menlo Park: AAAI, 2016: 179-185.
[16] BALABANOVI? M, SHOHAM Y. Fab: content-based, co-llaborative recommendation[J]. Communications of the ACM, 1997, 40(3): 66-72.
[17] CHANDRASEKARAN K, GAUCH S, LAKKARAJU P, et al. Concept-based document recommendations for citeseer authors[C]//LNCS 5149: Proceedings of the 5th International Conference on Adaptive Hypermedia and Adaptive Web-Based Systems, Hannover, Jul 29-Aug 1, 2008. Berlin, Hei-delberg: Springer, 2008: 83-92.
[18] HUANG W, WU Z, LIANG C, et al. A neural probabilistic model for context based citation recommendation[C]//Pro-ceedings of the 29th AAAI Conference on Artificial Intelli-gence, Austin, Jan 25-20, 2015. Menlo Park: AAAI, 2015: 2404-2410.
[19] STROHMAN T, CROFT W B, JENSEN D. Recommend-ing citations for academic papers[C]//Proceedings of the 30th International Conference on Research & Development in Information Retrieval, Amsterdam, Jul 23-27, 2007. New York: ACM, 2007: 705-706.
[20] MU D, GUO L, CAI X, et al. Query-focused personalized citation recommendation with mutually reinforced ranking[J]. IEEE Access, 2017, 6: 3107-3119.
[21] GUPTA S, VARMA V. Scientific article recommendation by using distributed representations of text and graph[C]//Pro-ceedings of the 26th International Conference on World Wide Web Companion, Perth, Apr 3-7, 2017. New York: ACM, 2017: 1267-1268.
[22] GONOG L, ZHOU Y. A review: generative adversarial net-works[C]//Proceedings of the 14th IEEE Conference on In-dustrial Electronics and Applications, Xi??an, Jun 19-21, 2019. Piscataway: IEEE, 2019.
[23] LIANG J, GURUKAR S, PARTHASARATHY S. Mile: a multi-level framework for scalable graph embedding[J]. arXiv: 1802.09612, 2018.