MasakhaNEWS: News Topic Classification for African languages
Abstract
African languages are severely under-represented in NLP research due to lack of datasets covering several NLP tasks. While there are individual language specific datasets that are being expanded to different tasks, only a handful of NLP tasks (e.g. named entity recognition and machine translation) have standardized benchmark datasets covering several geographical and typologically-diverse African languages. In this paper, we develop MasakhaNEWS -- a new benchmark dataset for news topic classification covering 16 languages widely spoken in Africa. We provide an evaluation of baseline models by training classical machine learning models and fine-tuning several language models. Furthermore, we explore several alternatives to full fine-tuning of language models that are better suited for zero-shot and few-shot learning such as cross-lingual parameter-efficient fine-tuning (like MAD-X), pattern exploiting training (PET), prompting language models (like ChatGPT), and prompt-free sentence transformer fine-tuning (SetFit and Cohere Embedding API). Our evaluation in zero-shot setting shows the potential of prompting ChatGPT for news topic classification in low-resource African languages, achieving an average performance of 70 F1 points without leveraging additional supervision like MAD-X. In few-shot setting, we show that with as little as 10 examples per label, we achieved more than 90\% (i.e. 86.0 F1 points) of the performance of full supervised training (92.6 F1 points) leveraging the PET approach.
Keywords
Cite
@article{arxiv.2304.09972,
title = {MasakhaNEWS: News Topic Classification for African languages},
author = {David Ifeoluwa Adelani and Marek Masiak and Israel Abebe Azime and Jesujoba Alabi and Atnafu Lambebo Tonja and Christine Mwase and Odunayo Ogundepo and Bonaventure F. P. Dossou and Akintunde Oladipo and Doreen Nixdorf and Chris Chinenye Emezue and sana al-azzawi and Blessing Sibanda and Davis David and Lolwethu Ndolela and Jonathan Mukiibi and Tunde Ajayi and Tatiana Moteu and Brian Odhiambo and Abraham Owodunni and Nnaemeka Obiefuna and Muhidin Mohamed and Shamsuddeen Hassan Muhammad and Teshome Mulugeta Ababu and Saheed Abdullahi Salahudeen and Mesay Gemeda Yigezu and Tajuddeen Gwadabe and Idris Abdulmumin and Mahlet Taye and Oluwabusayo Awoyomi and Iyanuoluwa Shode and Tolulope Adelani and Habiba Abdulganiyu and Abdul-Hakeem Omotayo and Adetola Adeeko and Abeeb Afolabi and Anuoluwapo Aremu and Olanrewaju Samuel and Clemencia Siro and Wangari Kimotho and Onyekachi Ogbu and Chinedu Mbonu and Chiamaka Chukwuneke and Samuel Fanijo and Jessica Ojo and Oyinkansola Awosan and Tadesse Kebede and Toadoum Sari Sakayo and Pamela Nyatsine and Freedmore Sidume and Oreen Yousuf and Mardiyyah Oduwole and Tshinu Tshinu and Ussen Kimanuka and Thina Diko and Siyanda Nxakama and Sinodos Nigusse and Abdulmejid Johar and Shafie Mohamed and Fuad Mire Hassan and Moges Ahmed Mehamed and Evrard Ngabire and Jules Jules and Ivan Ssenkungu and Pontus Stenetorp},
journal= {arXiv preprint arXiv:2304.09972},
year = {2023}
}
Comments
Accepted to IJCNLP-AACL 2023 (main conference)