2026· SemEval@ACL· pp. 2810-2816· 0 citations· 7 references
Computer Science
Abstract
This paper presents the PingAn-NLP system developed for SemEval-2026 Task 9, focusing on multilingual online polarization identification across 18 languages. We pro-pose a multi-stage optimization framework combining Supervised Fine-Tuning (SFT) and Group Relative Policy Optimization (GRPO). To overcome label imbalance and linguistic nuances, we utilized synthetic reasoning chain augmentation via a high-capacity teacher model (Qwen3-235B) and developed a Smart-Tradeoff reward mechanism to balance precision and recall during reinforcement learning. A language-aware tiered ensemble voting strategy was further implemented to optimize inference performance across diverse linguistic tracks. Our 8B-GRPO-Vote configuration achieved the highest Macro-F1 scores among our experimental variants in 7 out of 18 languages. Officially, our system secured second place in the Bengali, English, Odia, and Turkish tracks.
Alation studies reveal a strong synergy between POS features and CRF decoding, with the combined approach yielding the best single-model performance, and ensembling models trained with different objectives improves both overall F1 score and discontinuous MWE scores, demonstrating the importance of training diversity fo...
Yunus Karatepe, Mert Sülük, Zeynep Tu˘gçe Kırımlı et al.· 0 citations
This paper uses a fine-tuned TwHIN-BERT model, a multilingual encoder-based transformer pre-trained on social media data, to classify text across five polarization categories, suggesting that performance in such a multi-label polarization detection task is more strongly driven by local distributional properties than by...
The results show that hybrid ensemble methods can improve token-level accuracy in low-resource POS tagging, while also revealing a trade-off between frequent-tag accuracy and rare-tag robustness.
Abdelouahed Moussaoui, Nor-Eddine Azalmad, Said Bahassine et al.· Information· 0 citations
UIT-Polar system achieves Top-5 results on the English test set, demonstrating the effectiveness of integrating encoder-based screening with LLM-based refinement.
Doan K. Nguyen, H. Tran, T. Nguyen et al.· 0 citations
Error analysis shows that a non-trivial fraction of failures are placeholder strings caused by API errors rather than incorrect generations, and that surface-level mismatches (verbosity, ortho-graphic variation) account for many of the remaining errors.
A morphological barrier is shown: while pivot-based adaptation facilitates transfer for distant languages through script alignment, it induces negative transfer in morphologically complex agglutinative languages like Gondi.
R. Mundotiya· Annual International ACM SIG...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
