“This paper outlines aiXplain Inc.’s ensemble based approach to detecting offensive speech in the Arabic language based on OSACT5’s shared sub-task A. Additionally, this paper highlights multiple challenges that may hinder progress on detecting abusive speech and provides potential avenues and techniques that may lead to significant progress.”https://aclanthology.org/2022.osact-1.28/Share this:FacebookXLike this:Like Loading... Post navigation Improving Generalization of Hate Speech Detection Systems to Novel Target Groups via Domain Adaptation (ACL Anthology) Combating Hate Speech and Disinformation Against Social Polarization and How to be a Fact-Checking Influencer (Friedrich Naumann Foundation for Freedom)