{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,1,13]],"date-time":"2026-01-13T12:17:19Z","timestamp":1768306639548,"version":"3.49.0"},"reference-count":66,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,1,13]],"date-time":"2026-01-13T00:00:00Z","timestamp":1768262400000},"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. Artif. Intell."],"abstract":"<jats:p>\n                    Cyberbullying on social networks has emerged as a pressing global issue, yet research in low-resource languages such as Bengali remains underdeveloped due to the scarcity of high-quality datasets, linguistic resources, and targeted methodologies. Many existing approaches overlook essential language-specific preprocessing, neglect the integration of advanced transformer-based models, and do not adequately address model validation, scalability, and adaptability. To address these limitations, this study introduces three Bengali-specific preprocessing strategies to enhance feature representation. It then proposes\n                    <jats:italic>Transformer-stacking<\/jats:italic>\n                    , an effective hybrid detection framework that combines three transformer models, XLM-R-base, multilingual BERT, and Bangla-Bert-Base, via a stacking strategy with a multi-layer perceptron classifier. The framework is evaluated on a publicly available Bengali cyberbullying dataset comprising 44,001 samples across both binary (\n                    <jats:italic>Sub-task A<\/jats:italic>\n                    ) and multiclass (\n                    <jats:italic>Sub-task B<\/jats:italic>\n                    ) classification settings.\n                    <jats:italic>Transformer-stacking<\/jats:italic>\n                    achieves an F1-score of 93.61% and an accuracy of 93.62% for\n                    <jats:italic>Sub-task A<\/jats:italic>\n                    , and an F1-score and accuracy of 89.23% for\n                    <jats:italic>Sub-task B<\/jats:italic>\n                    , outperforming eight baseline transformer models, four transformer ensemble techniques, and recent state-of-the-art methods. These improvements are statistically validated using McNemar's test. Furthermore, experiments on two external Bengali datasets, focused on hate speech and abusive language, demonstrate the model's scalability and adaptability. Overall,\n                    <jats:italic>Transformer-stacking<\/jats:italic>\n                    offers an effective and generalizable solution for Bengali cyberbullying detection, establishing a new benchmark in this underexplored domain.\n                  <\/jats:p>","DOI":"10.3389\/frai.2025.1679962","type":"journal-article","created":{"date-parts":[[2026,1,13]],"date-time":"2026-01-13T06:36:03Z","timestamp":1768286163000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Advancing cyberbullying detection in low-resource languages: a transformer- stacking framework for Bengali"],"prefix":"10.3389","volume":"8","author":[{"given":"Md. Nesarul","family":"Hoque","sequence":"first","affiliation":[]},{"given":"Rudra Pratap","family":"Deb Nath","sequence":"additional","affiliation":[]},{"given":"Abu Nowshed","family":"Chy","sequence":"additional","affiliation":[]},{"given":"Debasish","family":"Ghose","sequence":"additional","affiliation":[]},{"given":"Md Hanif","family":"Seddiqui","sequence":"additional","affiliation":[]}],"member":"1965","published-online":{"date-parts":[[2026,1,13]]},"reference":[{"key":"B1","article-title":"Bangla text dataset and exploratory analysis for online harassment detection","author":"Ahmed","year":"","journal-title":"arXiv preprint arXiv:2102.02478"},{"key":"B2","article-title":"Cyberbullying detection using deep neural network from social media comments in Bangla language","author":"Ahmed","year":"","journal-title":"arXiv preprint arXiv:2106.04506"},{"key":"B3","doi-asserted-by":"publisher","first-page":"100027","DOI":"10.1016\/j.nlp.2023.100027","article-title":"A robust hybrid machine learning model for Bengali cyber bullying detection in social media","volume":"4","author":"Akhter","year":"2023","journal-title":"Nat. Lang. Proc. J"},{"key":"B4","doi-asserted-by":"publisher","first-page":"156","DOI":"10.3390\/make6010009","article-title":"An ensemble-based multi-classification machine learning classifiers approach to detect multiple classes of cyberbullying","volume":"6","author":"Alqahtani","year":"2024","journal-title":"Mach. Learn. Knowl. Extr"},{"key":"B5","doi-asserted-by":"publisher","first-page":"1","DOI":"10.32604\/cmc.2025.061674","article-title":"Leveraging transformers for detection of Arabic cyberbullying on social media: hybrid Arabic transformers","volume":"83","author":"Alsuwaylimi","year":"2025","journal-title":"Comput. Mater. Continua"},{"key":"B6","doi-asserted-by":"crossref","first-page":"35","DOI":"10.1109\/ICCITechn.2014.7073079","article-title":"\u201cIntroducing active learning on text to emotion analyzer,\u201d","volume-title":"2014 17th International Conference on Computer and Information Technology (ICCIT)","author":"Asad","year":"2014"},{"key":"B7","doi-asserted-by":"publisher","first-page":"24","DOI":"10.1007\/s13278-021-00852-x","article-title":"Abusive bangla comments detection on Facebook using transformer-based deep learning models","volume":"12","author":"Aurpa","year":"2022","journal-title":"Soc. Netw. Anal. Mining"},{"key":"B8","doi-asserted-by":"publisher","author":"Bhattacharjee","year":"2022","DOI":"10.18653\/v1\/2022.findings-naacl.98"},{"key":"B9","doi-asserted-by":"crossref","first-page":"535","DOI":"10.1145\/1150402.1150464","article-title":"\u201cModel compression,\u201d","volume-title":"Proceedings of the 12th ACM SIGKDD international conference on Knowledge discovery and Data Mining","author":"Bucilu\u01ce","year":"2006"},{"key":"B10","doi-asserted-by":"publisher","first-page":"61677","DOI":"10.1109\/ACCESS.2018.2874767","article-title":"Performance analysis of Google colaboratory as a tool for accelerating deep learning applications","volume":"6","author":"Carneiro","year":"2018","journal-title":"IEEE Access"},{"key":"B11","first-page":"1","article-title":"\u201cThreat and abusive language detection on social media in Bengali language,\u201d","volume-title":"2019 1st International Conference on Advances in Science, Engineering and Robotics Technology (ICASERT)","author":"Chakraborty","year":"2019"},{"key":"B12","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.banglalp-1.23","article-title":"\u201cVaclm at blp-2023 task 1: Leveraging bert models for violence detection in Bangla,\u201d","author":"Chatterjee","year":"2023","journal-title":"Proceedings of the First Workshop on Bangla Language Processing (BLP-2023)"},{"key":"B13","article-title":"Electra: pre-training text encoders as discriminators rather than generators","author":"Clark","year":"2020","journal-title":"arXiv preprint arXiv:2003.10555"},{"key":"B14","doi-asserted-by":"crossref","DOI":"10.18653\/v1\/2020.acl-main.747","article-title":"\u201cUnsupervised cross-lingual representation learning at scale,\u201d","volume-title":"Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics","author":"Conneau","year":"2020"},{"key":"B15","article-title":"Bert: Pre-training of deep bidirectional transformers for language understanding","author":"Devlin","year":"2018","journal-title":"arXiv preprint arXiv:1810.04805"},{"key":"B16","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1007\/3-540-45014-9_1","author":"Dietterich","year":"2000"},{"key":"B17","doi-asserted-by":"crossref","first-page":"277","DOI":"10.15439\/2019F143","article-title":"\u201cLanguages' impact on emotional classification methods,\u201d","volume-title":"2019 Federated Conference on Computer Science and Information Systems (FedCSIS)","author":"Eilertsen","year":"2019"},{"key":"B18","doi-asserted-by":"crossref","first-page":"86","DOI":"10.1007\/978-3-031-12638-3_8","article-title":"\u201cDetection of Bangla hate comments and cyberbullying in social media using NLP and transformer models,\u201d","volume-title":"Advances in Computing and Data Sciences: 6th International Conference, ICACDS 2022, Kurnool, India, April 22\u201323, 2022, Revised Selected Papers, Part I","author":"Emon","year":"2022"},{"key":"B19","doi-asserted-by":"publisher","first-page":"393","DOI":"10.1017\/nlp.2024.28","article-title":"Hate speech detection in low-resourced Indian languages: an analysis of transformer-based monolingual and multilingual models with cross-lingual experiments","volume":"31","author":"Ghosh","year":"2024","journal-title":"Nat. Lang. Proc"},{"key":"B20","doi-asserted-by":"publisher","first-page":"2381","DOI":"10.11591\/eei.v12i4.4743","article-title":"Bengali slang detection using state-of-the-art supervised models from a given text","volume":"12","author":"Hamid","year":"2023","journal-title":"Bull. Electr. Eng. Inform"},{"key":"B21","article-title":"Distilling the knowledge in a neural network","author":"Hinton","year":"2015","journal-title":"arXiv preprint arXiv:1503.02531"},{"key":"B22","first-page":"192","article-title":"The challenges and approaches during the detection of cyberbullying text for low-resource language: a literature review","volume":"17","author":"Hoque","year":"2023","journal-title":"ECTI Trans. Comput. Inf. Technol"},{"key":"B23","doi-asserted-by":"publisher","first-page":"49","DOI":"10.38032\/jea.2023.02.003","article-title":"Detecting level of depression from social media posts for the low-resource Bengali language","volume":"4","author":"Hoque","year":"2023","journal-title":"J. Eng. Advanc"},{"key":"B24","doi-asserted-by":"publisher","first-page":"100091","DOI":"10.1016\/j.nlp.2024.100091","article-title":"Exploring transformer models in the sentiment analysis task for the under-resource Bengali language","volume":"8","author":"Hoque","year":"","journal-title":"Nat. Lang. Proc. J"},{"key":"B25","doi-asserted-by":"publisher","first-page":"27","DOI":"10.38032\/jea.2024.02.001","article-title":"Depression intensity identification using transformer ensemble technique for the resource-constrained Bengali language","volume":"5","author":"Hoque","year":"","journal-title":"J. Eng. Advanc"},{"key":"B26","first-page":"1","article-title":"\u201cLeveraging transformer models in the cyberbullying text classification system for the low-resource Bengali language,\u201d","volume-title":"2023 26th International Conference on Computer and Information Technology (ICCIT)","author":"Hoque","year":"2023"},{"key":"B27","doi-asserted-by":"publisher","first-page":"358","DOI":"10.11591\/ijai.v13.i1.pp358-367","article-title":"Detecting cyberbullying text using the approaches with machine learning models for the low-resource Bengali language","volume":"13","author":"Hoque","year":"","journal-title":"IAES Int. J. Artif. Intell"},{"key":"B28","first-page":"1","article-title":"\u201cExploring transformer ensemble approach to classify cyberbullying text for the low-resource Bengali language,\u201d","volume-title":"2024 International Conference on Advances in Computing, Communication, Electrical, and Smart Systems (iCACCESS)","author":"Hoque","year":""},{"key":"B29","doi-asserted-by":"publisher","first-page":"988","DOI":"10.1016\/j.procs.2024.03.288","article-title":"An innovative novel transformer model and datasets for safeguarding religious sensitivities in online social platforms","volume":"233","author":"Islam","year":"2024","journal-title":"Procedia Comput. Sci"},{"key":"B30","doi-asserted-by":"publisher","first-page":"141","DOI":"10.1017\/S135132492000056X","article-title":"Automatic classification of participant roles in cyberbullying: can we detect victims, bullies, and bystanders in social media text?","volume":"28","author":"Jacobs","year":"2020","journal-title":"Nat. Lang. Eng"},{"key":"B31","doi-asserted-by":"publisher","first-page":"4160","DOI":"10.47852\/bonviewJCCE52024162","article-title":"Deep learning approaches for detecting cyberbullying on social media","volume":"3","author":"Jaradat","year":"2025","journal-title":"J. Comput. Cogn. Eng"},{"key":"B32","doi-asserted-by":"crossref","DOI":"10.18653\/v1\/2020.findings-emnlp.445","article-title":"\u201cIndicnlpsuite: Monolingual corpora, evaluation benchmarks and pre-trained multilingual language models for indian languages,\u201d","volume-title":"Findings of the Association for Computational Linguistics: EMNLP 2020","author":"Kakwani","year":"2020"},{"key":"B33","doi-asserted-by":"publisher","first-page":"11","DOI":"10.1002\/joe.22175","article-title":"Cyberbullying on social media under the influence of covid-19","volume":"41","author":"Kee","year":"2022","journal-title":"Global Bus. Organ. Excel"},{"key":"B34","doi-asserted-by":"publisher","first-page":"101856","DOI":"10.1016\/j.avb.2023.101856","article-title":"Anonymity and its role in digital aggression: a systematic review","volume":"72","author":"Kim","year":"2023","journal-title":"Aggress. Violent Behav"},{"key":"B35","doi-asserted-by":"crossref","DOI":"10.18653\/v1\/P18-1007","article-title":"\u201cSubword regularization: Improving neural network translation models with multiple subword candidates,\u201d","volume-title":"Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","author":"Kudo","year":"2018"},{"key":"B36","doi-asserted-by":"crossref","DOI":"10.18653\/v1\/D18-2012","article-title":"\u201cSentencepiece: a simple and language independent subword tokenizer and detokenizer for neural text processing,\u201d","volume-title":"Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing: System Demonstrations","author":"Kudo","year":"2018"},{"key":"B37","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1007\/s42979-020-00414-6","article-title":"Aggressive and offensive language identification in Hindi, Bangla, and English: a comparative study","volume":"2","author":"Kumar","year":"2021","journal-title":"SN Comput. Sci"},{"key":"B38","article-title":"Albert: A lite bert for self-supervised learning of language representations","author":"Lan","year":"2019","journal-title":"arXiv preprint arXiv:1909.11942"},{"key":"B39","article-title":"Roberta: a robustly optimized bert pretraining approach","author":"Liu","year":"2019","journal-title":"arXiv preprint arXiv:1907.11692"},{"key":"B40","doi-asserted-by":"crossref","first-page":"2750","DOI":"10.1109\/ICACCI.2014.6968484","article-title":"\u201cA rule based Bengali stemmer,\u201d","volume-title":"2014 International Conference on Advances in Computing, Communications and Informatics (ICACCI)","author":"Mahmud","year":"2014"},{"key":"B41","first-page":"1","article-title":"ktrain: a low-code library for augmented machine learning","volume":"23","author":"Maiya","year":"2022","journal-title":"J. Mach. Learn. Res"},{"key":"B42","doi-asserted-by":"publisher","first-page":"125641","DOI":"10.1016\/j.eswa.2024.125641","article-title":"Automatic detection of cyberbullying behaviour on social media using stacked BI-GRU attention with bert model","volume":"262","author":"Mali","year":"2025","journal-title":"Expert Syst. Appl"},{"key":"B43","doi-asserted-by":"publisher","first-page":"108241","DOI":"10.1016\/j.engappai.2024.108241","article-title":"Shielding against online harm: a survey on text analysis to prevent cyberbullying","volume":"133","author":"Mishra","year":"2024","journal-title":"Eng. Appl. Artif. Intell"},{"key":"B44","doi-asserted-by":"publisher","first-page":"e0308862","DOI":"10.1371\/journal.pone.0308862","article-title":"Machine learning and deep learning-based approach to categorize Bengali comments on social networks using fused dataset","volume":"19","author":"Mohi Uddin","year":"2024","journal-title":"PLoS ONE"},{"key":"B45","doi-asserted-by":"publisher","first-page":"77733","DOI":"10.1007\/s11042-023-17934-x","article-title":"Combining multiple pre-trained models for hate speech detection in Bengali, Marathi, and Hindi","volume":"83","author":"Nandi","year":"2024","journal-title":"Multimed. Tools Appl"},{"key":"B46","doi-asserted-by":"publisher","first-page":"382","DOI":"10.1504\/IJDMMM.2025.10066891","article-title":"Sentiment analysis of danish health care industries' financial text","volume":"17","author":"Nath","year":"2025","journal-title":"Int. J. Data Min. Modell. Manag"},{"key":"B47","doi-asserted-by":"publisher","first-page":"1269366","DOI":"10.3389\/frai.2024.1269366","article-title":"Protect: a hybrid deep learning model for proactive detection of cyberbullying on social media","volume":"7","author":"Nitya Harshitha","year":"2024","journal-title":"Front. Artif. Intell"},{"key":"B48","doi-asserted-by":"publisher","first-page":"97391","DOI":"10.1109\/ACCESS.2023.3313113","article-title":"Cyberbullying detection and severity determination model","volume":"11","author":"Obaid","year":"2023","journal-title":"IEEE Access"},{"key":"B49","doi-asserted-by":"publisher","first-page":"e01393","DOI":"10.1016\/j.heliyon.2019.e01393","article-title":"Cyberbullying and its influence on academic, social, and emotional development of undergraduate students","volume":"5","author":"Peled","year":"2019","journal-title":"Heliyon"},{"key":"B50","doi-asserted-by":"crossref","DOI":"10.18653\/v1\/N18-1202","article-title":"\u201cDeep contextualized word representations,\u201d","volume-title":"Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers)","author":"Peters","year":"2018"},{"key":"B51","doi-asserted-by":"crossref","DOI":"10.18653\/v1\/P19-1493","article-title":"\u201cHow multilingual is multilingual bert?,\u201d","volume-title":"Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics","author":"Pires","year":"2019"},{"key":"B52","author":"Radford","year":"2018","journal-title":"Improving language understanding by generative pre-training"},{"key":"B53","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3457610","article-title":"Multilingual offensive language identification for low-resource languages","volume":"21","author":"Ranasinghe","year":"2021","journal-title":"Trans. Asian Low-Resour. Lang. Inf. Proc"},{"key":"B54","doi-asserted-by":"publisher","first-page":"457","DOI":"10.1007\/978-981-16-0586-4_37","author":"Romim","year":"2021","journal-title":"Hate Speech Detection in the Bengali Language: A Dataset and Its Baseline Evaluation"},{"key":"B55","doi-asserted-by":"publisher","DOI":"10.3389\/fsoc.2023.1067726","article-title":"Sociological perspectives of social media, rumors, and attacks on minorities: evidence from Bangladesh","author":"Roy","year":"2023","journal-title":"Front. Sociol"},{"key":"B56","doi-asserted-by":"publisher","first-page":"124524","DOI":"10.1109\/ACCESS.2023.3329347","article-title":"Safeguarding online spaces: a powerful fusion of federated learning, word embeddings, and emotional features for cyberbullying detection","volume":"11","author":"Samee","year":"2023","journal-title":"IEEE Access"},{"key":"B57","article-title":"Distilbert, a distilled version of bert: smaller, faster, cheaper and lighter","author":"Sanh","year":"2019","journal-title":"arXiv preprint arXiv:1910.01108"},{"key":"B58","article-title":"Banglabert: Bengali mask language model for Bengali language understanding","author":"Sarker","year":"2020","journal-title":"textsIGitHub"},{"key":"B59","doi-asserted-by":"crossref","DOI":"10.18653\/v1\/P16-1162","article-title":"\u201cNeural machine translation of rare words with subword units,\u201d","volume-title":"Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","author":"Sennrich","year":"2016"},{"key":"B60","doi-asserted-by":"publisher","first-page":"100111","DOI":"10.1016\/j.nlp.2024.100111","article-title":"Recurrent neural network based multiclass cyber bullying classification","volume":"9","author":"Sifath","year":"2024","journal-title":"Nat. Lang. Proc. J"},{"key":"B61","doi-asserted-by":"publisher","first-page":"225","DOI":"10.1007\/s10994-013-5422-z","article-title":"An instance level analysis of data complexity","volume":"95","author":"Smith","year":"2014","journal-title":"Mach. Learn"},{"key":"B62","doi-asserted-by":"crossref","first-page":"998","DOI":"10.1109\/ICCIT64611.2024.11022404","article-title":"\u201cA cross-language analysis on sarcasm detection,\u201d","volume-title":"2024 27th International Conference on Computer and Information Technology (ICCIT)","author":"Tasnim","year":"2024"},{"key":"B63","doi-asserted-by":"publisher","first-page":"55533","DOI":"10.1109\/ACCESS.2023.3275130","article-title":"Cyberbullying detection in social networks: a comparison between machine learning and transfer learning approaches","volume":"11","author":"Teng","year":"2023","journal-title":"IEEE Access"},{"key":"B64","author":"Vaswani","year":"2017"},{"key":"B65","first-page":"439","article-title":"\u201cMulti-feature transformer for multiclass cyberbullying detection in bangla,\u201d","volume-title":"IFIP International Conference on Artificial Intelligence Applications and Innovations","author":"Wahid","year":"2023"},{"key":"B66","article-title":"Google's neural machine translation system: bridging the gap between human and machine translation","author":"Wu","year":"2016","journal-title":"arXiv preprint arXiv:1609.08144"}],"container-title":["Frontiers in Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2025.1679962\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,1,13]],"date-time":"2026-01-13T06:36:08Z","timestamp":1768286168000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2025.1679962\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,1,13]]},"references-count":66,"alternative-id":["10.3389\/frai.2025.1679962"],"URL":"https:\/\/doi.org\/10.3389\/frai.2025.1679962","relation":{},"ISSN":["2624-8212"],"issn-type":[{"value":"2624-8212","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,1,13]]},"article-number":"1679962"}}