针对22种语言的立场检测，提出自适应选择通用、专用或集成模型的策略。
MKJ at SemEval-2026 Task 9: A Comparative Study of Generalist, Specialist, and Ensemble Strategies for Multilingual Polarization
- 根据语言特性动态切换通用、专用或混合模型
- 整体宏平均F1达0.796，平均准确率0.826
- 适合多语言立场分析任务的研究者与实践者
我们针对SemEval-2026任务9（子任务1）中22种语言的多语言立场检测，系统比较了通用模型、语言专用模型及混合集成策略。当分词器与目标文本匹配时，标准通用模型如XLM-RoBERTa表现良好；但在使用不同书写系统的语言（如高棉语、奥里亚语）上，专用模型显著提升性能。为此，我们采用基于开发集表现的自适应框架，在通用模型、专用模型与混合模型间动态切换。此外，通过NLLB-200进行跨语言增强效果参差不齐，常不如架构选择有效，且在形态丰富的语言上表现下降。最终系统在全部22个语言赛道上取得0.796的宏平均F1和0.826的平均准确率。代码与测试预测已公开于：https://github.com/Maziarkiani/SemEval2026-Task9-Subtask1-Polarization。
原文摘要 · Abstract (English)
We present a systematic study of multilingual polarization detection across 22 languages for SemEval-2026 Task 9 (Subtask 1), contrasting multilingual generalists with language-specific specialists and hybrid ensembles. While a standard generalist like XLM-RoBERTa suffices when its tokenizer aligns with the target text, it may struggle with distinct scripts (e.g., Khmer, Odia) where monolingual specialists yield significant gains. Rather than enforcing a single universal architecture, we adopt a language-adaptive framework that switches between multilingual generalists, language-specific specialists, and hybrid ensembles based on development performance. Additionally, cross-lingual augmentation via NLLB-200 yielded mixed results, often underperforming native architecture selection and degrading morphologically rich tracks. Our final system achieves an overall macro-averaged F1 score of 0.796 and an average accuracy of 0.826 across all 22 tracks. Code and final test predictions are publicly available at: https://github.com/Maziarkiani/SemEval2026-Task9-Subtask1-Polarization.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。