BriefGPT.xyz
Jan, 2025
BBPOS：基于BERT的乌兹别克语词性标注
BBPOS: BERT-based Part-of-Speech Tagging for Uzbek
HTML
PDF
Latofat Bobojonova, Arofat Akhundjanova, Phil Ostheimer, Sophie Fellenz
TL;DR
本研究针对乌兹别克语这一低资源语言的自然语言处理，评估了两种之前未测试的单语乌兹别克BERT模型在词性标注任务上的表现，并引入了首个公开可用的乌兹别克语UPOS标注基准数据集。经微调的模型平均准确率达到91%，超越了基线的多语言BERT和基于规则的标注器，显示出相比现有规则标注器更强的上下文敏感性和词缀处理能力。
Abstract
This paper advances NLP research for the low-resource
Uzbek Language
by evaluating two previously untested monolingual Uzbek BERT models on the part-of-speech (POS) tagging task and introducing the first publicly available UPOS-tagged
→