BriefGPT.xyz
Sep, 2022
基于领域分类的源特定术语罚款方法，用于仇恨言语检测中的领域适应
Domain Classification-based Source-specific Term Penalization for Domain Adaptation in Hate-speech Detection
HTML
PDF
Tulika Bose, Nikolaos Aletras, Irina Illina, Dominique Fohr
TL;DR
提出了一种域自适应方法，通过自动提取和惩罚源特定术语来适应不同领域的仇恨言论检测，以实现跨领域评估的一致提高。
Abstract
State-of-the-art approaches for
hate-speech detection
usually exhibit poor performance in
out-of-domain settings
. This occurs, typically, due to classifiers overemphasizing source-specific information that negati
→