arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2604.14907cs.CLcs.LG
收藏

多语言文本嵌入技术在仇恨言论检测任务中的比较

Comparison of Modern Multilingual Text Embedding Techniques for Hate Speech Detection Task

Department of Information Systems, Faculty of Informatics, Kaunas University of Technology, Kaunas, Lithuania(信息系统系,信息学院,凯纳斯技术大学,凯纳斯,立陶宛);Centre of Information Systems Design Technologies, Faculty of Informatics, Kaunas University of Technology, Kaunas, Lithuania(信息系统设计技术中心,信息学院,凯纳斯技术大学,凯纳斯,立陶宛)

Evaldas Vaiciukynas, Paulius Danenas, Linas Ablonskis, Algirdas Sukys, Edgaras Dambrauskas, Voldemaras Zitkus, Rita Butkiene, Rimantas Butleris

AI总结:

本文比较了现代多语言句子嵌入模型在立陶宛、俄语和英语中的仇恨言论检测性能,探讨了下游建模选择和特征维度对性能的影响,引入了新的立陶宛仇恨言论语料库并评估了六种多语言编码器。

摘要:

在线仇恨言论和攻击性语言对内容审核构成日益增长的挑战,尤其是在多语言环境和低资源语言如立陶宛语中。本文研究了现代多语言句子嵌入模型在立陶宛、俄语和英语中支持准确仇恨言论检测的程度,以及其性能如何依赖于下游建模选择和特征维度。我们介绍了LtHate,一个从新闻门户和社会网络衍生的新立陶宛仇恨言论语料库,并在LtHate、RuToxic和EnSuperset上使用统一的Python管道评估了六个现代多语言编码器(potion、gemma、bge、snow、jina、e5)。对于每种嵌入,我们训练了单类HBOS异常检测器和双类CatBoost分类器,并在不使用主成分分析(PCA)压缩到64维特征向量的情况下进行训练。在所有数据集中,双类监督模型始终显著优于单类异常检测,最佳配置在立陶宛(jina)达到80.96%的准确率和0.887的AUC ROC,在俄语(e5)达到92.19%的准确率和0.978的AUC ROC,在英语(e5加PCA)达到77.21%的准确率和0.859的AUC ROC。PCA压缩在监督设置中几乎保留了所有判别能力,但在无监督异常检测情况下显示出一些负面影响。这些结果展示了现代多语言句子嵌入结合梯度提升决策树如何为多语言仇恨言论检测应用提供稳健的软计算解决方案。

英文摘要:

Online hate speech and abusive language pose a growing challenge for content moderation, especially in multilingual settings and for low-resource languages such as Lithuanian. This paper investigates to what extent modern multilingual sentence embedding models can support accurate hate speech detection in Lithuanian, Russian, and English, and how their performance depends on downstream modeling choices and feature dimensionality. We introduce LtHate, a new Lithuanian hate speech corpus derived from news portals and social networks, and benchmark six modern multilingual encoders (potion, gemma, bge, snow, jina, e5) on LtHate, RuToxic, and EnSuperset using a unified Python pipeline. For each embedding, we train both a one class HBOS anomaly detector and a two class CatBoost classifier, with and without principal component analysis (PCA) compression to 64-dimensional feature vectors. Across all datasets, two class supervised models consistently and substantially outperform one class anomaly detection, with the best configurations achieving up to 80.96% accuracy and AUC ROC of 0.887 in Lithuanian (jina), 92.19% accuracy and AUC ROC of 0.978 in Russian (e5), and 77.21% accuracy and AUC ROC of 0.859 in English (e5 with PCA). PCA compression preserves almost all discriminative power in the supervised setting, while showing some negative impact for the unsupervised anomaly detection case. These results demonstrate how modern multilingual sentence embeddings combined with gradient boosted decision trees provide robust soft-computing solutions for multilingual hate speech detection applications.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑