Reproduce “AfriHate: A Multilingual Collection of Hate Speech and Abusive Language Datasets for African Languages” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Shamsuddeen Hassan Muhammad, Idris Abdulmumin, Abinew Ali Ayele, David Ifeoluwa Adelani, Ibrahim Said Ahmad, Saminu Mohammad Aliyu, Nelson Odhiambo Onyango, Lilian D. A. Wanzare and 19 more
AfriHate presents 90,437 tweets across 15 African-language datasets, annotated by native speakers for hate, abusive or offensive, and neutral speech. The authors use the collection to compare monolingual and multilingual model training, with multilingual training pooling data across languages. AfroXLMR-76L reaches average macro-F1 78.16, versus 76.45 monolingually, on AfriHate test sets; this does not establish real-world moderation performance.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.