{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,10]],"date-time":"2026-09-10T05:09:45Z","timestamp":1789016985338,"version":"build-2803163510"},"reference-count":36,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,10]],"date-time":"2026-09-10T00:00:00Z","timestamp":1788998400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Big Data"],"abstract":"<jats:sec>\n                    <jats:title>Introduction<\/jats:title>\n                    <jats:p>Matching short, specialized skill expressions across English, Russian, and Kazakh is challenging because general-domain multilingual encoders underperform on terse, code-mixed, domain-specific phrases, particularly in the low-resource Kazakh setting.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>We fine-tuned a multilingual Sentence Transformer using a staged Multiple Negatives Ranking objective on a trilingual paraphrase corpus, including Russian augmentation pairs. We evaluated skill matching and semantic similarity across languages and assessed the resulting embeddings through downstream skill-taxonomy clustering.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>Fine-tuning preserved English performance while improving Russian and Kazakh similarity quality. Russian cosine Pearson correlation increased from 0.8125 to 0.8221, while Kazakh cosine Pearson increased from 0.5989 to 0.6050 and Kazakh dot-product Pearson from 0.4487 to 0.4912. Agglomerative clustering improved mean silhouette from 0.27126 to 0.2851 and reduced erroneous clusters from 19.74% to 13.97%.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Discussion<\/jats:title>\n                    <jats:p>The results provide evidence consistent with cross-lingual transfer as an important mechanism of improvement for Kazakh. They also motivate language-specific threshold calibration and demonstrate that intrinsic similarity improvements translate into a cleaner downstream skill taxonomy.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/fdata.2026.1900836","type":"journal-article","created":{"date-parts":[[2026,9,10]],"date-time":"2026-09-10T04:29:26Z","timestamp":1789014566000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Fine-tuning multilingual sentence transformers for low-resource skill-concept matching across Kazakh, Russian, and English"],"prefix":"10.3389","volume":"9","author":[{"given":"Sandugash","family":"Serikbayeva","sequence":"first","affiliation":[{"name":"Department of Information Systems, L.N. Gumilyov Eurasian National University","place":["Astana, Kazakhstan"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Madina","family":"Sambetbayeva","sequence":"additional","affiliation":[{"name":"Department of Information Systems, L.N. Gumilyov Eurasian National University","place":["Astana, Kazakhstan"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Valiya","family":"Ramazanova","sequence":"additional","affiliation":[{"name":"Department of Information Systems, L.N. Gumilyov Eurasian National University","place":["Astana, Kazakhstan"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Aigerim","family":"Yerimbetova","sequence":"additional","affiliation":[{"name":"Institute of Information and Computational Technologies, CS MSHE RK","place":["Almaty, Kazakhstan"]},{"name":"School of Engineering and Information Technology, META University","place":["Almaty, Kazakhstan"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhanar","family":"Lamasheva","sequence":"additional","affiliation":[{"name":"Department of Information Systems, L.N. Gumilyov Eurasian National University","place":["Astana, Kazakhstan"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhanna","family":"Sadirmekova","sequence":"additional","affiliation":[{"name":"School of Software Engineering, Astana IT University","place":["Astana, Kazakhstan"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ardak","family":"Batyrkhanov","sequence":"additional","affiliation":[{"name":"Department of Software Engineering, Atyrau University named after Kh. Dosmukhamedov","place":["Atyrau, Kazakhstan"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yersaiyn","family":"Mailybayev","sequence":"additional","affiliation":[{"name":"Department of Computer Technology and Telecommunications, International University of Transportation and Humanities","place":["Almaty, Kazakhstan"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,10]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"497","DOI":"10.18653\/v1\/S16-1081","article-title":"\u201cSemEval-2016 Task 1: semantic textual similarity, monolingual and cross-lingual evaluation,\u201d","author":"Agirre","year":"2016","journal-title":"Proceedings of SemEval-2016"},{"key":"B2","doi-asserted-by":"publisher","first-page":"597","DOI":"10.1162\/tacl_a_00288","article-title":"Massively multilingual sentence embeddings for zero-shot cross-lingual transfer and beyond","volume":"7","author":"Artetxe","year":"2019","journal-title":"Trans. Assoc. Comput. Linguist."},{"key":"B3","doi-asserted-by":"publisher","first-page":"5832","DOI":"10.18653\/v1\/2020.coling-main.513","article-title":"\u201cRetrieving skills from job descriptions: a language model based extreme multi-label classification framework,\u201d","author":"Bhola","year":"2020","journal-title":"Proceedings of COLING 2020"},{"key":"B4","doi-asserted-by":"crossref","first-page":"160","DOI":"10.1007\/978-3-642-37456-2_14","article-title":"\u201cDensity-based clustering based on hierarchical density estimates,\u201d","volume-title":"Pacific-Asia Conference on Knowledge Discovery and Data Mining","author":"Campello","year":"2013"},{"key":"B5","doi-asserted-by":"publisher","first-page":"1","DOI":"10.18653\/v1\/S17-2001","article-title":"\u201cSemEval-2017 Task 1: semantic textual similarity multilingual and crosslingual focused evaluation,\u201d","author":"Cer","year":"2017","journal-title":"Proceedings of SemEval-2017"},{"key":"B6","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-acl.137","article-title":"BGE M3-embedding: multi-lingual, multi-functionality, multi-granularity text embeddings through self-knowledge distillation","author":"Chen","year":"2024","journal-title":"arXiv"},{"key":"B7","doi-asserted-by":"publisher","first-page":"8440","DOI":"10.18653\/v1\/2020.acl-main.747","article-title":"\u201cUnsupervised cross-lingual representation learning at scale,\u201d","author":"Conneau","year":"2020","journal-title":"Proceedings of ACL 2020"},{"key":"B8","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2109.09605","article-title":"JobBERT: understanding job titles through skills","author":"Decorte","year":"2021","journal-title":"arXiv"},{"key":"B9","doi-asserted-by":"publisher","first-page":"4171","DOI":"10.18653\/v1\/N19-1423","article-title":"\u201cBERT: pre-training of deep bidirectional transformers for language understanding,\u201d","author":"Devlin","year":"2019","journal-title":"Proceedings of NAACL-HLT 2019"},{"key":"B10","first-page":"226","article-title":"\u201cA density-based algorithm for discovering clusters in large spatial databases with noise,\u201d","volume-title":"Proceedings of KDD-96","author":"Ester","year":"1996"},{"key":"B11","doi-asserted-by":"publisher","first-page":"878","DOI":"10.18653\/v1\/2022.acl-long.62","article-title":"\u201cLanguage-agnostic BERT sentence embedding,\u201d","author":"Feng","year":"2022","journal-title":"Proceedings of ACL 2022"},{"key":"B110","doi-asserted-by":"publisher","first-page":"86","DOI":"10.1145\/3458723","article-title":"Datasheets for datasets","volume":"64","author":"Gebru","year":"2021","journal-title":"Commun. ACM"},{"key":"B12","doi-asserted-by":"publisher","first-page":"107049","DOI":"10.1016\/j.asoc.2020.107049","article-title":"Skills2Job: a recommender system that encodes job offer embeddings on graph databases","volume":"101","author":"Giabelli","year":"2021","journal-title":"Appl. Soft Comput."},{"key":"B13","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1705.00652","article-title":"Efficient natural language response suggestion for Smart Reply","author":"Henderson","year":"2017","journal-title":"arXiv"},{"key":"B14","doi-asserted-by":"publisher","first-page":"6282","DOI":"10.18653\/v1\/2020.acl-main.560","article-title":"\u201cThe state and fate of linguistic diversity and inclusion in the NLP world,\u201d","author":"Joshi","year":"2020","journal-title":"Proceedings of ACL 2020"},{"key":"B15","article-title":"\u201cCross-lingual ability of multilingual BERT: an empirical study,\u201d","volume-title":"Proceedings of ICLR 2020","author":"Karthikeyan","year":"2020"},{"key":"B16","doi-asserted-by":"publisher","first-page":"118134","DOI":"10.1109\/ACCESS.2021.3106120","article-title":"A survey on skill identification from online job ads","volume":"9","author":"Khaouja","year":"2021","journal-title":"IEEE Access"},{"key":"B17","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2308.03281","article-title":"Towards general text embeddings with multi-stage contrastive learning","author":"Li","year":"2023","journal-title":"arXiv"},{"key":"B18","doi-asserted-by":"publisher","first-page":"11427","DOI":"10.3390\/ijms241411427","article-title":"Bioreactor technologies for enhanced organoid culture","volume":"24","author":"Licata","year":"2023","journal-title":"Int. J. Mol. Sci."},{"key":"B19","article-title":"\u201cDecoupled weight decay regularization,\u201d","author":"Loshchilov","year":"2019","journal-title":"7th International Conference on Learning Representations, New Orleans"},{"key":"B20","doi-asserted-by":"publisher","first-page":"1022","DOI":"10.18653\/v1\/D13-1104","article-title":"\u201cAssembling the Kazakh language corpus,\u201d","author":"Makhambetov","year":"2013","journal-title":"Proceedings of EMNLP 2013"},{"key":"B21","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1301.3781","article-title":"Efficient estimation of word representations in vector space","author":"Mikolov","year":"2013","journal-title":"arXiv"},{"key":"B22","doi-asserted-by":"publisher","first-page":"5404","DOI":"10.63317\/2592g9pbfyft","article-title":"\u201cKazakhTTS2: extending the open-source Kazakh TTS corpus with more data, speakers, and topics,\u201d","author":"Mussakhojayeva","year":"2022","journal-title":"Proceedings of LREC 2022"},{"key":"B23","doi-asserted-by":"publisher","first-page":"604","DOI":"10.1109\/TNNLS.2020.2979670","article-title":"A survey of the usages of deep learning for natural language processing","volume":"32","author":"Otter","year":"2021","journal-title":"IEEE Trans. Neural Netw. Learn. Syst."},{"key":"B24","doi-asserted-by":"publisher","first-page":"1532","DOI":"10.3115\/v1\/D14-1162","article-title":"\u201cGloVe: global vectors for word representation,\u201d","author":"Pennington","year":"2014","journal-title":"Proceedings of EMNLP 2014"},{"key":"B25","doi-asserted-by":"publisher","first-page":"4996","DOI":"10.18653\/v1\/P19-1493","article-title":"\u201cHow multilingual is multilingual BERT?\u201d","author":"Pires","year":"2019","journal-title":"Proceedings of ACL 2019"},{"key":"B26","doi-asserted-by":"publisher","first-page":"423","DOI":"10.52209\/1609-1825_2024_2_425","volume":"2","author":"Ramazanova","year":"2024"},{"key":"B27","doi-asserted-by":"publisher","first-page":"3982","DOI":"10.18653\/v1\/D19-1410","article-title":"\u201cSentence-BERT: sentence embeddings using Siamese BERT-networks,\u201d","author":"Reimers","year":"2019","journal-title":"Proceedings of EMNLP-IJCNLP 2019"},{"key":"B28","doi-asserted-by":"publisher","first-page":"4512","DOI":"10.18653\/v1\/2020.emnlp-main.365","article-title":"\u201cMaking monolingual sentence embeddings multilingual using knowledge distillation,\u201d","author":"Reimers","year":"2020","journal-title":"Proceedings of EMNLP 2020"},{"key":"B29","doi-asserted-by":"publisher","first-page":"53","DOI":"10.1016\/0377-0427(87)90125-7","article-title":"Silhouettes: a graphical aid to the interpretation and validation of cluster analysis","volume":"20","author":"Rousseeuw","year":"1987","journal-title":"J. Comput. Appl. Math."},{"key":"B30","first-page":"16857","article-title":"\u201cMPNet: masked and permuted pre-training for language understanding,\u201d","volume-title":"Advances in Neural Information Processing Systems 33","author":"Song","year":"2020"},{"key":"B31","first-page":"5998","article-title":"\u201cAttention is all you need,\u201d","volume-title":"Advances in Neural Information Processing Systems 30","author":"Vaswani","year":"2017"},{"key":"B32","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2212.03533","article-title":"Text embeddings by weakly-supervised contrastive pre-training","author":"Wang","year":"2022","journal-title":"arXiv"},{"key":"B33","doi-asserted-by":"publisher","first-page":"833","DOI":"10.18653\/v1\/D19-1077","article-title":"\u201cBeto, Bentz, Becas: the surprising cross-lingual effectiveness of BERT,\u201d","author":"Wu","year":"2019","journal-title":"Proceedings of EMNLP-IJCNLP 2019"},{"key":"B34","doi-asserted-by":"publisher","first-page":"417","DOI":"10.63317\/5398gpuejb6t","article-title":"\u201cKazNERD: Kazakh named entity recognition dataset,\u201d","author":"Yeshpanov","year":"2022","journal-title":"Proceedings of LREC 2022"},{"key":"B35","doi-asserted-by":"publisher","first-page":"4962","DOI":"10.18653\/v1\/2022.naacl-main.366","article-title":"\u201cSkillSpan: hard and soft skill extraction from English job postings,\u201d","author":"Zhang","year":"2022","journal-title":"Proceedings of NAACL 2022"}],"container-title":["Frontiers in Big Data"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdata.2026.1900836\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,10]],"date-time":"2026-09-10T04:29:29Z","timestamp":1789014569000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdata.2026.1900836\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,10]]},"references-count":36,"alternative-id":["10.3389\/fdata.2026.1900836"],"URL":"https:\/\/doi.org\/10.3389\/fdata.2026.1900836","relation":{},"ISSN":["2624-909X"],"issn-type":[{"value":"2624-909X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,10]]},"article-number":"1900836"}}