--- title: "zhangpeng at SemEval-2026 Task 9: Detecting Multilingual," description: "This system paper describes a submission to SemEval-2026 Task 9, a shared task on detecting online… 9 references, cited by 1 paper here." authors: ["Zhang Peng", "Lu Gehao"] year: 2026 venue: "Proceedings of the 20th International Workshop on Semantic Evaluation (2026)" source: "https://aclanthology.org/2026.semeval-1.14/" doi: "10.18653/v1/2026.semeval-1.14" licence: "CC BY 4.0 (ACL Anthology)" licence_url: "https://creativecommons.org/licenses/by/4.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/en/papers/peng-2026-zhangpeng-semeval-2026-task-9/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/peng-2026-zhangpeng-semeval-2026-task-9/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/19/) / zhangpeng at SemEval-2026 Task 9: Detecting Multilingual, Multicultural and Multievent Online Polarization # zhangpeng at SemEval-2026 Task 9: Detecting Multilingual, Multicultural and Multievent Online Polarization **Authors:** [Zhang Peng](https://md.research.elamur.ai/free/en/authors/zhang-peng/) and [Lu Gehao](https://md.research.elamur.ai/free/en/authors/lu-gehao/) · **Year:** 2026 · **Venue:** [Proceedings of the 20th International Workshop on Semantic Evaluation (2026)](https://md.research.elamur.ai/free/en/venues/semeval/) ([SemEval 2026](https://md.research.elamur.ai/free/en/venues/semeval/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/) **Links:** [Publisher's page](https://aclanthology.org/2026.semeval-1.14/) · [PDF](https://aclanthology.org/2026.semeval-1.14.pdf) · [DOI 10.18653/v1/2026.semeval-1.14](https://doi.org/10.18653/v1/2026.semeval-1.14) · [OpenAlex](https://openalex.org/W7166846161) **Citations:** 0 citations (OpenAlex, 2026-10-14) · [1](https://md.research.elamur.ai/free/en/papers/peng-2026-zhangpeng-semeval-2026-task-9/cited-by/) in our library, of the 149,439 papers whose bibliographies we read ## Summary This system paper describes a submission to SemEval-2026 Task 9, a shared task on detecting online polarization in 22 languages. The task has three subtasks: deciding whether a post is polarized, classifying the polarization target type, and identifying how polarization is manifested. The authors work only on English and Spanish. For the binary subtask, they compare three classical document representations, each paired with a scikit-learn Random Forest: - Bag-of-Words, - averaged Word2Vec vectors, - Bag-of-Centroids, which uses k-means clusters of word embeddings. For the two multi-label subtasks, they fine-tune several pre-trained transformers with sigmoid outputs using the simpletransformers library. The models' predictions are combined by voting, with each model weighted by its normalized validation F1. Hyperparameters were tuned over two optimizers and three learning rates. Bag-of-Centroids performed best for detection. The ensemble did reasonably on Spanish type classification but was weak on English type classification and on manifestation identification. The work matters mainly as a baseline: inexpensive classical features remain competitive for polarization detection in short posts, while finer-grained labels are much harder. ## Key points - **Subtask 1 (binary detection):** Bag-of-Centroids with a 100-tree Random Forest won on the development set. - English: Macro-F1 0.7039 - Spanish: Macro-F1 0.6659 - Averaged Word2Vec performed much worse (0.4605 in English). - **Transformers used for Subtasks 2–3:** google-bert, FacebookAI-roberta, dccuchile-bert (BETO, Spanish-specific) and distilbert-multi. Hyperparameters tested were AdamW vs. Adafactor and learning rates of 2e-5, 4e-5 and 8e-5. - **Official test Macro-F1 (English / Spanish):** - Subtask 1: 0.6882 / 0.6711 - Subtask 2: 0.3752 / 0.6386 - Subtask 3: 0.3561 / 0.4366 - **Subtask 2 training data:** 2,676 texts, and roughly 75% of them contain at most 20 words. Labels are heavily imbalanced: political 996, racial/ethnic 264, other 121, religious 106, gender/sexual 67. - **Code:** available at github.com/zpp1219/SemEval-2026-Task-9. *Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.* ## Institutions - [Yunnan University](https://md.research.elamur.ai/free/en/universities/yunnan-university/) — University, Kunming, China · 17 works · 40 citations 📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works · [🤝 Social Sciences](https://md.research.elamur.ai/free/en/shelves/social-sciences/) · 1,772 works · [🤝 Sociology and Political Science](https://md.research.elamur.ai/free/en/shelves/social-sciences/sociology-and-political-science/) · 577 works 🧭 Topics (1): [🌍 Multilingual & cross-lingual NLP](https://md.research.elamur.ai/free/en/topics/multilingual-nlp/) · 828 works ## Cite Cite as: Zhang Peng and Lu Gehao. 2026. zhangpeng at SemEval-2026 Task 9: Detecting Multilingual, Multicultural and Multievent Online Polarization. Proceedings of the 20th International Workshop on Semantic Evaluation (2026). The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @inproceedings{peng2026zhangpeng, title = {{zhangpeng at SemEval-2026 Task 9: Detecting Multilingual, Multicultural and Multievent Online Polarization}}, author = {Zhang Peng and Lu Gehao}, booktitle = {Proceedings of the 20th International Workshop on Semantic Evaluation (2026)}, year = {2026}, doi = {10.18653/v1/2026.semeval-1.14}, url = {https://doi.org/10.18653/v1/2026.semeval-1.14} } ``` - [← previous zhangpeng at SemEval-2026 Task 10: PsyCoMark - Psycholinguistic Conspiracy Marker Extraction and Detection](https://md.research.elamur.ai/free/en/papers/peng-2026-zhangpeng-semeval-2026-task-10/) - [next → ZYC at SemEval-2026 Task 5: Application of BERT-based Contextual Embeddings Similarity for WSD](https://md.research.elamur.ai/free/en/papers/zhou-2026-zyc-semeval-2026-task-5/) - [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/19/) ## References 9 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. Usman Naseem, Robert Geislinger, Juan Ren, Sarah Kohail, Rudy Garrido Veliz, P Sam Sahil, Yiran Zhang, Marco Antonio Stranisci, Idris Abdulmumin, "Ozge Alacam, Cengiz Acar"urk, Aisha Jabr, Saba Anwar, Abinew Ali Ayele, Elena Tutubalina, Aung Kyaw Htet, Xintong Wang, Surendrabikram Thapa, Tanmoy Chak… [find](https://scholar.google.com/scholar?q=Usman+Naseem%2C+Robert+Geislinger%2C+Juan+Ren%2C+Sarah+Kohail%2C+Rudy+Garrido+Veliz%2C+P+Sam+Sahil%2C+Yiran+Zhang%2C+Marco+Antonio+Stranisci%2C+Idris+Abdulmumin%2C+%22Ozge+Alacam%2C+Cengiz+Acar%22urk%2C+Aisha+Jabr%2C+Saba+Anwar%2C) 2. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. [Bert: Pre-training of deep bidirectional transformers for language understanding](https://md.research.elamur.ai/free/en/papers/devlin-2019-bert-pre-training-deep-bidirectional/). Preprint, arXiv:1810.04805. *Table 4: Test set results for all subtasks. The last column reports the Macro-F1 score of the official POLAR baseline… 3. English English English Spanish Spanish Spanish 2 2 English Spanish 3 3 English Spanish Method Accuracy Precision Recall F1 Binary Subtask 1: Polarization Detection Bag-of-Words 0.7149 [0.6291 0.5441 0.5835 Word2Vec Avg](https://md.research.elamur.ai/free/en/papers/naseem-2026-semeval-2026-task-9-detecting/). 0.6398 0.5439 0.1163 0.1917 Bag-of-Centroids 0.7176 0.6300 0.5591 0.5924 Bag-of-… 4. Usman Naseem, Robert Geislinger, Juan Ren, Sarah Kohail, Rudy Garrido Veliz, P Sam Sahil, Yiran Zhang, Marco Antonio Stranisci, Idris Abdulmumin, Özge Alacam, Cengiz Acartürk, Aisha Jabr, Saba Anwar, Abinew Ali Ayele, Simona Frenda, Alessandra Teresa Cignarella, Elena Tutubalina, Oleg Rogov, Aung Ky… — [POLAR: A Benchmark for Multilingual, Multicultural, and Multi-Event Online Polarization](https://md.research.elamur.ai/free/en/papers/naseem-2026-polar-benchmark-multilingual-multicultural-multi/) 5. Use or Not yes no Omer Sagi and Lior Rokach. 2018. Ensemble learning: A survey. Wiley interdisciplinary reviews: data mining and knowledge discovery, 8(4):e1249. [find](https://scholar.google.com/scholar?q=Use+or+Not+yes+no+Omer+Sagi+and+Lior+Rokach%2E+2018%2E+Ensemble+learning%3A+A+survey%2E+Wiley+interdisciplinary+reviews%3A+data+mining+and+knowledge+discovery%2C+8%284%29%3Ae1249%2E) 6. Ankit Vaidya, Aditya Gokhale, Arnav Desai, Ishaan Shukla, and Sheetal Sonawane. 2024. CLTeam1 at SemEval-2024 task 10: Large language model based ensemble for emotion detection in Hinglish. In Proceedings of the 18th International Workshop on Semantic Evaluation (SemEval-2024), pages 365–369, Mexico… [find](https://scholar.google.com/scholar?q=Ankit+Vaidya%2C+Aditya+Gokhale%2C+Arnav+Desai%2C+Ishaan+Shukla%2C+and+Sheetal+Sonawane%2E+2024%2E+CLTeam1+at+SemEval%2D2024+task+10%3A+Large+language+model+based+ensemble+for+emotion+detection+in+Hinglish%2E+In+Proceed) 7. Yuxia Wang, Jonibek Mansurov, Petar Ivanov, Jinyan Su, Artem Shelmanov, Akim Tsvigun, Osama Mohammed Afzal, Tarek Mahmoud, Giovanni Puccetti, and Thomas Arnold. 2024. SemEval-2024 task 8: Multidomain, multimodel and multilingual machine-generated text detection. In Proceedings of the 18th Internatio… [find](https://scholar.google.com/scholar?q=Yuxia+Wang%2C+Jonibek+Mansurov%2C+Petar+Ivanov%2C+Jinyan+Su%2C+Artem+Shelmanov%2C+Akim+Tsvigun%2C+Osama+Mohammed+Afzal%2C+Tarek+Mahmoud%2C+Giovanni+Puccetti%2C+and+Thomas+Arnold%2E+2024%2E+SemEval%2D2024+task+8%3A+Multidomain%2C) 8. Jeffrey Pennington, Richard Socher, and Christopher D Manning. 2014. [Glove: Global vectors for word representation](https://md.research.elamur.ai/free/en/papers/pennington-2014-glove-global-vectors-word-representation/). In Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP), pages 1532–1543. 9. Nils Reimers and Iryna Gurevych. 2019. [Sentence-bert: Sentence embeddings using siamese bert-networks](https://md.research.elamur.ai/free/en/papers/reimers-2019-sentence-bert-sentence-embeddings-siamese/). In Proceedings of the 2019 conference on empirical methods in natural language processing and the 9th international joint conference on natural language processing (EMNLP-IJCNLP), pages 3982–3992. *Updated 31 October 2026*