{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,16]],"date-time":"2026-07-16T14:48:52Z","timestamp":1784213332934,"version":"3.55.0"},"reference-count":63,"publisher":"Springer Science and Business Media LLC","issue":"5","license":[{"start":{"date-parts":[[2025,7,18]],"date-time":"2025-07-18T00:00:00Z","timestamp":1752796800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2025,7,18]],"date-time":"2025-07-18T00:00:00Z","timestamp":1752796800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/501100001774","name":"University of Sydney","doi-asserted-by":"crossref","id":[{"id":"10.13039\/501100001774","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Data Min Knowl Disc"],"published-print":{"date-parts":[[2025,9]]},"abstract":"<jats:title>Abstract<\/jats:title>\n          <jats:p>Entity alignment (EA) aims at identifying equivalent entity pairs across different knowledge graphs (KGs) that refer to the same real-world identity. It has been a compelling but challenging task that requires the integration of heterogeneous information from different KGs to expand the knowledge coverage and enhance inference abilities. To circumvent the shortage of prior seed alignments provided for training, recent EA models utilize pseudo-labeling strategies to iteratively add unaligned entity pairs predicted with high confidence to the seed alignments for model training. However, the adverse impact of confirmation bias during pseudo-labeling has been largely overlooked, thus hindering entity alignment performance. To systematically combat confirmation bias, we propose a new <jats:underline>U<\/jats:underline>nified <jats:underline>P<\/jats:underline>seudo-<jats:underline>L<\/jats:underline>abeling framework for <jats:underline>E<\/jats:underline>ntity <jats:underline>A<\/jats:underline>lignment (UPL-EA) that explicitly alleviates pseudo-labeling errors to boost the performance of entity alignment. UPL-EA achieves this goal through two key innovations: (1) Optimal Transport (OT)-based pseudo-labeling uses discrete OT modeling as an effective means to determine entity correspondences and reduce erroneous matches across two KGs. An effective criterion is derived to infer pseudo-labeled alignments that satisfy one-to-one correspondences; (2) Parallel pseudo-label ensembling refines pseudo-labeled alignments by combining predictions over multiple models independently trained in parallel. The ensembled pseudo-labeled alignments are thereafter used to augment seed alignments to reinforce subsequent model training for alignment inference. The effectiveness of UPL-EA in eliminating pseudo-labeling errors is both theoretically supported and experimentally validated. Our extensive results and in-depth analyses demonstrate the superiority of UPL-EA over 15 competitive baselines and its utility as a general pseudo-labeling framework for entity alignment.<\/jats:p>","DOI":"10.1007\/s10618-025-01128-0","type":"journal-article","created":{"date-parts":[[2025,7,18]],"date-time":"2025-07-18T06:28:10Z","timestamp":1752820090000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":5,"title":["Combating confirmation bias: a unified pseudo-labeling framework for entity alignment"],"prefix":"10.1007","volume":"39","author":[{"given":"Qijie","family":"Ding","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2063-8437","authenticated-orcid":false,"given":"Jie","family":"Yin","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1803-5768","authenticated-orcid":false,"given":"Daokun","family":"Zhang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9803-0256","authenticated-orcid":false,"given":"Junbin","family":"Gao","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2025,7,18]]},"reference":[{"key":"1128_CR1","doi-asserted-by":"crossref","unstructured":"Arazo E, Ortego D, Albert P et\u00a0al. (2020) Pseudo-labeling and confirmation bias in deep semi-supervised learning. In: IJCNN, IEEE, pp 1\u20138","DOI":"10.1109\/IJCNN48605.2020.9207304"},{"key":"1128_CR2","doi-asserted-by":"crossref","unstructured":"Auer S, Bizer C, Kobilarov G et\u00a0al. (2007) DBpedia: A nucleus for a web of open data. In: ISWC, Springer, pp 722\u2013735","DOI":"10.1007\/978-3-540-76298-0_52"},{"key":"1128_CR3","doi-asserted-by":"crossref","unstructured":"Bollacker K, Evans C, Paritosh P et\u00a0al. (2008) Freebase: A collaboratively created graph database for structuring human knowledge. In: SIGMOD, pp 1247\u20131250","DOI":"10.1145\/1376616.1376746"},{"key":"1128_CR4","unstructured":"Bordes A, Usunier N, Garcia-Dur\u00e1n A et\u00a0al. (2013) Translating embeddings for modeling multi-relational data. In: NeurIPS, pp 2787\u20132795"},{"key":"1128_CR5","doi-asserted-by":"crossref","unstructured":"Cascante-Bonilla P, Tan F, Qi Y et\u00a0al. (2021) Curriculum labeling: Revisiting pseudo-labeling for semi-supervised learning. In: AAAI, pp 6912\u20136920","DOI":"10.1609\/aaai.v35i8.16852"},{"key":"1128_CR6","unstructured":"Chen L, Gan Z, Cheng Y et\u00a0al. (2020) Graph optimal transport for cross-domain alignment. In: ICML, pp 1542\u20131553"},{"key":"1128_CR7","doi-asserted-by":"crossref","unstructured":"Chen M, Tian Y, Yang M et\u00a0al. (2017) Multilingual knowledge graph embeddings for cross-lingual knowledge alignment. In: IJCAI, pp 1511\u20131517","DOI":"10.24963\/ijcai.2017\/209"},{"key":"1128_CR8","unstructured":"Cuturi M (2013) Sinkhorn distances: Lightspeed computation of optimal transport. In: NeurIPS, pp 2292\u20132300"},{"key":"1128_CR9","unstructured":"Devlin J, Chang M, Lee K et\u00a0al. (2019) BERT: Pre-training of deep bidirectional transformers for language understanding. In: NAACL-HLT, ACL, pp 4171\u20134186"},{"key":"1128_CR10","doi-asserted-by":"crossref","unstructured":"Ding Q, Zhang D, Yin J (2022) Conflict-aware pseudo labeling via optimal transport for entity alignment. In: ICDM, IEEE, pp 915\u2013920","DOI":"10.1109\/ICDM54844.2022.00107"},{"key":"1128_CR11","unstructured":"Guo L, Sun Z, Hu W (2019) Learning to exploit long-term relational dependencies in knowledge graphs. In: ICML, pp 2505\u20132514"},{"issue":"8","key":"1128_CR12","doi-asserted-by":"publisher","first-page":"3549","DOI":"10.1109\/TKDE.2020.3028705","volume":"34","author":"Q Guo","year":"2022","unstructured":"Guo Q, Zhuang F, Qin C et al. (2022) A survey on knowledge graph-based recommender systems. IEEE Trans Knowl Data Eng 34(8):3549\u20133568","journal-title":"IEEE Trans Knowl Data Eng"},{"key":"1128_CR13","doi-asserted-by":"crossref","unstructured":"Jiang X, Yang Z, Wen P et\u00a0al. (2022) A sparse-motif ensemble graph convolutional network against over-smoothing. In: IJCAI, pp 2094\u20132100","DOI":"10.24963\/ijcai.2022\/291"},{"key":"1128_CR14","doi-asserted-by":"crossref","unstructured":"Ke Z, Wang D, Yan Q et\u00a0al. (2019) Dual student: Breaking the limits of the teacher in semi-supervised learning. In: ICCV, pp 6728\u20136736","DOI":"10.1109\/ICCV.2019.00683"},{"key":"1128_CR15","unstructured":"Laine S, Aila T (2017) Temporal ensembling for semi-supervised learning. In: ICLR"},{"key":"1128_CR16","unstructured":"Lee DH (2013) Pseudo-label: The simple and efficient semi-supervised learning method for deep neural networks. In: ICML Workshop: Challenges in Representation Learning, p 896"},{"key":"1128_CR17","doi-asserted-by":"crossref","unstructured":"Leone M, Huber S, Arora A et al. (2022) A critical re-evaluation of neural methods for entity alignment. In: Proceedings of the VLDB Endowment 15(8):1712\u20131725","DOI":"10.14778\/3529337.3529355"},{"key":"1128_CR18","doi-asserted-by":"crossref","unstructured":"Li Q, Han Z, Wu X (2018) Deeper insights into graph convolutional networks for semi-supervised learning. In: AAAI, pp 3538\u20133545","DOI":"10.1609\/aaai.v32i1.11604"},{"key":"1128_CR19","doi-asserted-by":"publisher","first-page":"228","DOI":"10.1007\/s10618-022-00879-4","volume":"37","author":"Y Li","year":"2023","unstructured":"Li Y, Yin J, Chen L (2023) Informative pseudo-labeling for graph neural networks with few labels. Data Min Knowl Disc 37:228\u2013254","journal-title":"Data Min Knowl Disc"},{"key":"1128_CR20","doi-asserted-by":"crossref","unstructured":"Liu X, Hong H, Wang X et\u00a0al. (2022) SelfKG: Self-supervised entity alignment in knowledge graphs. In: WWW, pp 860\u2013870","DOI":"10.1145\/3485447.3511945"},{"key":"1128_CR21","doi-asserted-by":"crossref","unstructured":"Mao X, Wang W, Xu H et\u00a0al. (2020) MRAEA: An efficient and robust entity alignment approach for cross-lingual knowledge graph. In: WSDM, pp 420\u2013428","DOI":"10.1145\/3336191.3371804"},{"key":"1128_CR22","doi-asserted-by":"crossref","unstructured":"Mao X, Wang W, Wu Y et\u00a0al. (2022) LightEA: A scalable, robust, and interpretable entity alignment framework via three-view label propagation. In: EMNLP, pp 825\u2013838","DOI":"10.18653\/v1\/2022.emnlp-main.52"},{"key":"1128_CR23","unstructured":"Maretic HP, Gheche ME, Chierchia G et\u00a0al. (2019) GOT: An optimal transport framework for graph comparison. In: NeurIPS, pp 13876\u201313887"},{"key":"1128_CR24","unstructured":"Min Y, Wenkel F, Wolf G (2020) Scattering GCN: Overcoming oversmoothness in graph convolutional networks. In: NeurIPS, pp 14498\u201314508"},{"key":"1128_CR25","doi-asserted-by":"crossref","unstructured":"Motschenbacher H (2022) Linguistic barriers in foreign language education. In: Research Questions in Language Education and Applied Linguistics: A Reference Guide. Springer, pp 711\u2013716","DOI":"10.1007\/978-3-030-79143-8_124"},{"key":"1128_CR26","doi-asserted-by":"publisher","first-page":"109","DOI":"10.1007\/BF02614365","volume":"78","author":"JB Orlin","year":"1997","unstructured":"Orlin JB (1997) A polynomial time primal network simplex algorithm for minimum cost flows. Math Program 78:109\u2013129","journal-title":"Math Program"},{"issue":"3","key":"1128_CR27","first-page":"489","volume":"8","author":"H Paulheim","year":"2017","unstructured":"Paulheim H (2017) Knowledge graph refinement: A survey of approaches and evaluation methods. Semantic Web 8(3):489\u2013508","journal-title":"Semantic Web"},{"key":"1128_CR28","doi-asserted-by":"crossref","unstructured":"Pei S, Yu L, Zhang X (2019) Improving cross-lingual entity alignment via optimal transport. In: IJCAI, pp 3231\u20133237","DOI":"10.24963\/ijcai.2019\/448"},{"key":"1128_CR29","doi-asserted-by":"crossref","unstructured":"Pennington J, Socher R, Manning CD (2014) GloVe: Global vectors for word representation. In: EMNLP, pp 1532\u20131543","DOI":"10.3115\/v1\/D14-1162"},{"key":"1128_CR30","doi-asserted-by":"crossref","unstructured":"Pham H, Xie Q, Dai Z et\u00a0al. (2021) Meta pseudo labels. In: CVPR, pp 11557\u201311568","DOI":"10.1109\/CVPR46437.2021.01139"},{"key":"1128_CR31","doi-asserted-by":"crossref","unstructured":"Raunak V, Gupta V, Metze F (2019) Effective dimensionality reduction for word embeddings. In: The 4th Workshop on Representation Learning for NLP, pp 235\u2013243","DOI":"10.18653\/v1\/W19-4328"},{"key":"1128_CR32","unstructured":"Rizve MN, Duarte K, Rawat YS et\u00a0al. (2021) In defense of pseudo-labeling: An uncertainty-aware pseudo-label selection framework for semi-supervised learning. In: ICLR"},{"key":"1128_CR33","doi-asserted-by":"crossref","unstructured":"Sajjadi M, Javanmardi M, Tasdizen T (2016) Mutual exclusivity loss for semi-supervised deep learning. In: ICIP, pp 1908\u20131912","DOI":"10.1109\/ICIP.2016.7532690"},{"key":"1128_CR34","doi-asserted-by":"crossref","unstructured":"Shi W, Gong Y, Ding C et al. (2018) Transductive semi-supervised deep learning using min-max features. In: ECCV, pp 311\u2013327","DOI":"10.1007\/978-3-030-01228-1_19"},{"key":"1128_CR35","unstructured":"Srivastava RK, Greff K, Schmidhuber J (2015) Highway networks. In: ICML Workshop: Deep Learning"},{"key":"1128_CR36","doi-asserted-by":"crossref","unstructured":"Suchanek FM, Kasneci G, Weikum G (2007) YAGO: A core of semantic knowledge. In: WWW, pp 697\u2013706","DOI":"10.1145\/1242572.1242667"},{"key":"1128_CR37","doi-asserted-by":"crossref","unstructured":"Suchanek FM, Abiteboul S, Senellart P (2011) PARIS: Probabilistic alignment of relations, instances, and schema. In: Proceedings of the VLDB Endowment 5(3):157\u2013168","DOI":"10.14778\/2078331.2078332"},{"key":"1128_CR38","doi-asserted-by":"crossref","unstructured":"Sun K, Zhu Z, Lin Z (2020a) Multi-stage self-supervised learning for graph convolutional networks. In: AAAI, pp 5892\u20135899","DOI":"10.1609\/aaai.v34i04.6048"},{"key":"1128_CR39","doi-asserted-by":"crossref","unstructured":"Sun Z, Hu W, Li C (2017) Cross-lingual entity alignment via joint attribute-preserving embedding. In: ISWC, pp 628\u2013644","DOI":"10.1007\/978-3-319-68288-4_37"},{"key":"1128_CR40","doi-asserted-by":"crossref","unstructured":"Sun Z, Hu W, Zhang Q et\u00a0al. (2018) Bootstrapping entity alignment with knowledge graph embedding. In: IJCAI, pp 4396\u20134402","DOI":"10.24963\/ijcai.2018\/611"},{"key":"1128_CR41","doi-asserted-by":"crossref","unstructured":"Sun Z, Huang J, Hu W et\u00a0al. (2019) TransEdge: Translating relation-contextualized embeddings for knowledge graphs. In: ISWC, Springer, pp 612\u2013629","DOI":"10.1007\/978-3-030-30793-6_35"},{"key":"1128_CR42","doi-asserted-by":"crossref","unstructured":"Sun Z, Zhang Q, Hu W et\u00a0al. (2020b) A benchmarking study of embedding-based entity alignment for knowledge graphs. In: Proceedings of the VLDB Endowment 13(12)","DOI":"10.14778\/3407790.3407828"},{"key":"1128_CR43","doi-asserted-by":"crossref","unstructured":"Tang J, Zhang W, Li J et\u00a0al. (2023) Robust attributed graph alignment via joint structure learning and optimal transport. In: arXiv preprint arXiv:2301.12721","DOI":"10.1109\/ICDE55515.2023.00129"},{"key":"1128_CR44","doi-asserted-by":"crossref","unstructured":"Tang X, Zhang J, Chen B et\u00a0al. (2021) BERT-INT: A BERT-based interaction model for knowledge graph alignment. In: IJCAI, pp 3174\u20133180","DOI":"10.24963\/ijcai.2020\/439"},{"key":"1128_CR45","unstructured":"Tarvainen A, Valpola H (2017) Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results. In: NeurIPS, pp 1195\u20131204"},{"key":"1128_CR46","unstructured":"Titouan V, Courty N, Tavenard R et\u00a0al. (2019) Optimal transport for structured data with application on graphs. In: ICML, pp 6275\u20136284"},{"key":"1128_CR47","unstructured":"Torres LC, Pereira LM, Amini MH (2021) A survey on optimal transport for machine learning: Theory and applications. arXiv preprint arXiv:2106.01963"},{"key":"1128_CR48","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-540-71050-9","volume-title":"Optimal transport: old and new,","author":"C Villani","year":"2009","unstructured":"Villani C (2009) Optimal transport: old and new, vol 338. Springer, Berlin"},{"issue":"10","key":"1128_CR49","doi-asserted-by":"publisher","first-page":"78","DOI":"10.1145\/2629489","volume":"57","author":"D Vrande\u010di\u0107","year":"2014","unstructured":"Vrande\u010di\u0107 D, Kr\u00f6tzsch M (2014) Wikidata: A free collaborative knowledge base. Commun ACM 57(10):78\u201385","journal-title":"Commun ACM"},{"key":"1128_CR50","doi-asserted-by":"publisher","first-page":"25","DOI":"10.1007\/s10107-004-0559-y","volume":"106","author":"A W\u00e4chter","year":"2006","unstructured":"W\u00e4chter A, Biegler LT (2006) On the implementation of an interior-point filter line-search algorithm for large-scale nonlinear programming. Math Program 106:25\u201357","journal-title":"Math Program"},{"key":"1128_CR51","doi-asserted-by":"crossref","unstructured":"Wang Z, Zhang J, Feng J et\u00a0al. (2014) Knowledge graph embedding by translating on hyperplanes. In: AAAI, pp 1112\u20131119","DOI":"10.1609\/aaai.v28i1.8870"},{"key":"1128_CR52","doi-asserted-by":"crossref","unstructured":"Wang Z, Lv Q, Lan X et\u00a0al. (2018) Cross-lingual knowledge graph alignment via graph convolutional networks. In: EMNLP, pp 349\u2013357","DOI":"10.18653\/v1\/D18-1032"},{"key":"1128_CR53","doi-asserted-by":"crossref","unstructured":"Wu Y, Liu X, Feng Y et\u00a0al. (2019a) Relation-aware entity alignment for heterogeneous knowledge graphs. In: IJCAI, pp 5278\u20135284","DOI":"10.24963\/ijcai.2019\/733"},{"key":"1128_CR54","doi-asserted-by":"crossref","unstructured":"Wu Y, Liu X, Feng Y et\u00a0al. (2019b) Jointly learning entity and relation representations for entity alignment. In: EMNLP\/IJCNLP, pp 240\u2013249","DOI":"10.18653\/v1\/D19-1023"},{"key":"1128_CR55","unstructured":"Xu H, Luo D, Carin L (2019a) Scalable gromov-wasserstein learning for graph partitioning and matching. In: NeurIPS, pp 3046\u20133056"},{"key":"1128_CR56","unstructured":"Xu H, Luo D, Zha H et\u00a0al. (2019b) Gromov-wasserstein learning for graph matching and node embedding. In: ICML, pp 6932\u20136941"},{"key":"1128_CR57","doi-asserted-by":"crossref","unstructured":"Yang H, Zou Y, Shi P et\u00a0al. (2019) Aligning cross-lingual entities with multi-aspect information. In: EMNLP\/IJCNLP, pp 4430\u20134440","DOI":"10.18653\/v1\/D19-1451"},{"key":"1128_CR58","doi-asserted-by":"crossref","unstructured":"Yang Z, Qi P, Zhang S et\u00a0al. (2018) HotpotQA: A dataset for diverse, explainable multi-hop question answering. In: EMNLP, pp 2369\u20132380","DOI":"10.18653\/v1\/D18-1259"},{"key":"1128_CR59","doi-asserted-by":"crossref","unstructured":"Zeng W, Zhao X, Tang J et\u00a0al. (2020) Collective entity alignment via adaptive features. In: ICDE, IEEE, pp 1870\u20131873","DOI":"10.1109\/ICDE48307.2020.00191"},{"key":"1128_CR60","unstructured":"Zhang B, Wang Y, Hou W et\u00a0al. (2021) FlexMatch: Boosting semi-supervised learning with curriculum pseudo labeling. In: NeurIPS, pp 18408\u201318419"},{"issue":"6","key":"1128_CR61","first-page":"2610","volume":"34","author":"X Zhao","year":"2020","unstructured":"Zhao X, Zeng W, Tang J et al. (2020) An experimental study of state-of-the-art entity alignment approaches. IEEE Trans Knowl Data Eng 34(6):2610\u20132625","journal-title":"IEEE Trans Knowl Data Eng"},{"key":"1128_CR62","doi-asserted-by":"crossref","unstructured":"Zhu H, Xie R, Liu Z et\u00a0al. (2017) Iterative entity alignment via knowledge embeddings. In: IJCAI, pp 4258\u20134264","DOI":"10.24963\/ijcai.2017\/595"},{"key":"1128_CR63","doi-asserted-by":"crossref","unstructured":"Zhu Y, Liu H, Wu Z et\u00a0al. (2021) Relation-aware neighborhood matching model for entity alignment. In: AAAI, pp 4749\u20134756","DOI":"10.1609\/aaai.v35i5.16606"}],"container-title":["Data Mining and Knowledge Discovery"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10618-025-01128-0.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s10618-025-01128-0\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s10618-025-01128-0.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,9,12]],"date-time":"2025-09-12T10:30:28Z","timestamp":1757673028000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s10618-025-01128-0"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,7,18]]},"references-count":63,"journal-issue":{"issue":"5","published-print":{"date-parts":[[2025,9]]}},"alternative-id":["1128"],"URL":"https:\/\/doi.org\/10.1007\/s10618-025-01128-0","relation":{},"ISSN":["1384-5810","1573-756X"],"issn-type":[{"value":"1384-5810","type":"print"},{"value":"1573-756X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,7,18]]},"assertion":[{"value":"7 July 2023","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"13 June 2025","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"18 July 2025","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors have no conflict of interest or conflict of interest to declare that are relevant to the content of this article.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Conflict of interest"}}],"article-number":"51"}}