[Purpose/Significance] Accurate annotation of research entities is fundamental to constructing high-quality academic social networks. Although mainstream open academic knowledge graphs (e.g., AMiner OAG) offer broad entity coverage, they suffer from over-conflation of authors and institutions, which undermine analytical accuracy. Enhancing the granularity and accuracy of entity annotations can provide more reliable data support for applications such as analyzing collaboration patterns, tracing disciplinary evolution, and identifying research trends, thus holding significant theoretical and practical value. [Method/Process] A disambiguation framework was proposed, integrating semantic embeddings derived from large language models with academic social relation analysis. The framework consisted of three modules: semantic embedding, institution disaggregation, and author disaggregation. In the semantic embedding module, millions of institution name variants and tens of millions of research-related keywords were vectorized using a large embedding-based language model. In the institution disaggregation module, institution name variants were clustered based on semantic similarity to generate fine-grained secondary institution entities. In the author disaggregation module, a multi-relational author signature network was constructed by incorporating co-affiliation, co-author name matching, and keyword domain overlap to identify distinct author entities. Applying the proposed method to the DBLP-Citation-network dataset, 10 794 overly heterogeneous OAG institution entities were refined into 56 527 secondary-level entities such as departments and research institutes, and 2 794 014 OAG author entities were disaggregated into 3 706 896 more accurate entities. [Result/Conclusion] The proposed approach improvs the annotation granularity while preserving appropriate inter-entity heterogeneity. It also demonstrates strong robustness across different embedding models, allowing flexible model selection based on practical constraints. These findings validate the applicability of embedding-based large language models in research entity disambiguation and provide a more reliable data foundation for academic social network analysis and recommendation system development.
[1] WAN H, ZHANG Y, ZHANG J, et al. AMiner: search and mining of academic social networks[J]. Data intelligence, 2019, 1(1): 58-76.
[2] GAO M, CHEN Y, GONG Q, et al. Understanding Scholar Social Networks: Taking SCHOLAT as an Example[M]. Computer Supported Cooperative Work and Social Computing. Singapore: Springer, 2022: 326-339.
[3] PARASKEVOPOULOS P, BOLDRINI C, PASSARELLA A, et al. The academic wanderer: structure of collaboration network and relation with research performance[J]. Applied network science, 2021, 6(1): 31.
[4] AJIFERUKE I, GRÁCIO M C C, YANG S. Editorial: research collaboration and networks: characteristics, evolution and trends[J]. Frontiers in research metrics and analytics, 2021, 6: 690986.
[5] ZHANG F, CAO K, CEN Y, et al. PST-Bench: tracing and benchmarking the source of publications[EB/OL]. [2024/02/25]. https://arxiv.org/abs/2402.16009.
[6] ZHU J, HUA W. Visualizing the knowledge domain of sustainable development research between 1987 and 2015: a bibliometric analysis[J]. Scientometrics, 2017, 110(2): 893-914.
[7] WANG K, SHEN Z, HUANG C, et al. Microsoft academic graph: when experts are not enough[J]. Quantitative science studies, 2020, 1(1): 396-413.
[8] LIU M, JAISWAL A, BU Y, et al. Team formation and team impact: the balance between team freshness and repeat collaboration[J]. Journal of informetrics, 2022, 16(4): 101337.
[9] HUANG Y, LU W, LIU J, et al. Towards transdisciplinary impact of scientific publications: a longitudinal, comprehensive, and large-scale analysis on Microsoft Academic Graph[J]. Information processing and management, 2022, 59(2): 102859.
[10] KANAKIA A, EIDE D, SHEN Z, et al. A scalable hybrid research paper recommender system for Microsoft academic[C]//Proceedings of the World Wide Web conference. San Francisco: ACM, 2019: 2893-2899.
[11] ZHANG F, LIU X, TANG J, et al. OAG: linking entities across large-scale heterogeneous knowledge graphs[J]. IEEE transactions on knowledge and data engineering, 2023, 35(9): 9225-9239.
[12] TANG J, ZHANG J, YAO L, et al. ArnetMiner: extraction and mining of academic social networks[C]//Proceedings of the 14th ACM SIGKDD international conference on knowledge discovery and data mining. New York: ACM, 2008: 990-998.
[13] SANTINI C, GESESE G A, PERONI S, et al. A knowledge graph embeddings based approach for author name disambiguation using literals[J]. Scientometrics, 2022, 127(8): 4887-4912.
[14] ZHANG Z, WU C, LI Z, et al. Author name disambiguation using multiple graph attention networks[C]//Proceedings of the international joint conference on neural networks. Shenzhen: IEEE, 2021: 1-8.
[15] CHEN Y, YUAN H, LIU T, et al. Name disambiguation based on graph convolutional network[J]. Scientific programming, 2021, 2021: 5577692.
[16] ZHANG F, LIU X, TANG J, et al. OAG: toward linking large-scale heterogeneous entity graphs[C]//Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining. New York: ACM, 2019: 2585-2595.
[17] GETOOR L, MACHANAVAJJHALA A. Entity resolution: theory, practice & open challenges[J]. Proceedings of the VLDB endowment, 2012, 5(12): 2018-2019.
[18] YIN X, HAN J, YU P S. Object distinction: distinguishing objects with identical names[C]//2007 IEEE 23rd international conference on data engineering. Istanbul: IEEE, 2007: 1242-1246.
[19] LVEXING Z, XUEQIANG L, KUN L, et al. Recognition of chinese personal names based on CRFs and law of names[M]. Web Technologies and Applications. Berlin: Springer, 2012: 163-170.
[20] ZHAO X, ZENG W, TANG J. Entity alignment - concepts, recent advances and novel approaches[M]. Singapore: Springer, 2023.
[21] KHAN H, SAQIB M, KHATTAK H A, et al. Ontology alignment for accurate ontology matching: a survey[M]. Digital Health Transformation, Smart Ageing, and Managing Disability. Switzerland: Springer, 2023: 338-349.
[22] 董文佳, 孙坦, 赵瑞雪, 等. 融合规范知识库和异质信息网络嵌入的作者姓名消歧研究[J]. 数据分析与知识发现, 2024: 1-17. (DONG W, SUN T, ZHAO R, et al. Author name disambiguation with standardized knowledge bases and heterogeneous information network embedding[J]. Data analysis and knowledge discovery, 2024: 1-17.)
[23] CHEN Y, JIANG Z, GAO J, et al. A supervised and distributed framework for cold-start author disambiguation in large-scale publications[J]. Neural computing and applications, 2023, 35(18): 13093-13108.
[24] ZHANG Y, YAO P, ZHANG F, et al. Name disambiguation in Aminer: clustering, maintenance, and human in the loop[C]//Proceedings of the ACM SIGKDD international conference on knowledge discovery and data mining. London: ACM, 2018: 1002-1011.
[25] REHS A. A supervised machine learning approach to author disambiguation in the Web of Science[J]. Journal of informetrics, 2021, 15(3): 101166.
[26] HAN H, ZHA H, GILES C L. Name disambiguation in author citations using a K-way spectral clustering method[C]//Proceedings of the 5th ACM/IEEE-CS joint conference on digital libraries. Denver: ACM, 2005: 334-343.
[27] DE BONIS M, FALCHI F, MANGHI P. Graph-based methods for author name disambiguation: a survey[J]. PeerJ computer science, 2023, 9: e1536.
[28] QIAO Z, DU Y, FU Y, et al. Unsupervised author disambiguation using heterogeneous graph convolutional network embedding[C]//Proceedings of the 2019 IEEE international conference on big data. Los Angeles: IEEE, 2019: 910-919.
[29] LEY M. The DBLP Computer science bibliography: evolution, research issues, perspectives[M]. String Processing and Information Retrieval. Berlin: Springer, 2002: 1-10.
[30] XIAO S, LIU Z, ZHANG P, et al. C-Pack: packaged resources to advance general Chinese embedding[C]//Proceedings of the 47th international ACM SIGIR conference on research and development in information retrieval. Washington DC: ACM, 2023: 641-649.
[31] ROSE S, ENGEL D, CRAMER N, et al. Automatic keyword extraction from individual documents[M]. New York: Wiley, 2010: 1-20.
[32] MÜLLNER D. Fastcluster: fast hierarchical, agglomerative clustering routines for R and Python[J]. Journal of statistical software, 2013, 53(9): 1-18.
[33] WANG W, BAO H, HUANG S, et al. MINILMv2: multi-head self-attention relation distillation for compressing pretrained transformers[C]//Findings of the association for computational linguistics: ACL-IJCNLP 2021. Bangkok: ACL, 2021: 2140-2151.
[34] SUN X, KAUR J, POSSAMAI L, et al. Ambiguous author query detection using crowdsourced digital library annotations[J]. Information processing & management, 2013, 49(2): 454-464.
[35] GASPARYAN A Y, YESSIRKEPOV M, GERASIMOV A N, et al. Scientific author names: errors, corrections, and identity profiles[J]. Biochemia medica, 2016: 169-173.
[36] XU S B, HU G. Rethinking the author name ambiguity problem and beyond: the case of the Chinese context[J]. Accountability in research, 2024: 1-24.
