%0 Generic
%A Bamba Dione, CM
%A Adelani, DI
%A Nabende, P
%A Alabi, JO
%A Sindane, T
%A Buzaaba, H
%A Muhammad, SH
%A Emezue, CC
%A Ogayo, P
%A Aremu, A
%A Gitau, C
%A Mbaye, D
%A Mukiibi, J
%A Sibanda, B
%A Dossou, BFP
%A Bukula, A
%A Mabuya, R
%A Tapo, AA
%A Munkoh-Buabeng, E
%A Koagne, VM
%A Kabore, FO
%A Taylor, A
%A Kalipe, G
%A Macucwa, T
%A Marivate, V
%A Gwadabe, T
%A Mboning, ET
%A Onyenwe, I
%A Atindogbe, G
%A Adelani, TA
%A Akinade, I
%A Samuel, O
%A Nahimana, M
%A Musabeyezu, T
%A Niyomutabazi, E
%A Chimhenga, E
%A Gotosa, K
%A Mizha, P
%A Agbolo, A
%A Traore, S
%A Uchechukwu, C
%A Yusuf, A
%A Abdullahi, M
%A Klakow, D
%C Toronto, Canada
%D 2023
%E Rogers, Anna
%E Boyd-Graber, Jordan
%E Okazaki, Naoaki
%F discovery:10181863
%I Association for Computational Linguistics
%P 10883-10900
%T MasakhaPOS: Part-of-Speech Tagging for Typologically Diverse African languages
%U https://discovery.ucl.ac.uk/id/eprint/10181863/
%V 1
%X In this paper, we present AfricaPOS, the largest part-of-speech (POS) dataset for 20 typologically diverse African languages. We discuss the challenges in annotating POS for these languages using the universal dependencies (UD) guidelines. We conducted extensive POS baseline experiments using both conditional random field and several multilingual pre-trained language models. We applied various cross-lingual transfer models trained with data available in the UD. Evaluating on the AfricaPOS dataset, we show that choosing the best transfer language(s) in both single-source and multi-source setups greatly improves the POS tagging performance of the target languages, in particular when combined with parameter-fine-tuning methods. Crucially, transferring knowledge from a language that matches the language family and morphosyntactic properties seems to be more effective for POS tagging in unseen languages.
%Z ACL materials are Copyright © 1963–2023 ACL; other materials are copyrighted by their respective copyright holders. Materials prior to 2016 here are licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 International License. Permission is granted to make copies for the purposes of teaching and research. Materials published in or after 2016 are licensed on a Creative Commons Attribution 4.0 International License, https://creativecommons.org/licenses/by/4.0/.