MasakhaNER 2.0: Africa-centric Transfer Learning for Named Entity Recognition
Domain:
natural language processingRecord type:
paperdatasetmodel
Creator:
AdeNeuRudRij
Host:
African languages are spoken by over a billion people, but are underrepresented in NLP research and development. The challenges impeding progress include the limited availability of annotated datasets, as well as a lack of understanding of the settings where current methods are effective. In this paper, we make progress towards solutions for these challenges, focusing on the task of named entity recognition (NER). We create the largest human-annotated NER dataset for 20 African languages, and we study the behavior of state-of-the-art cross-lingual transfer methods in an Africa-centric setting, demonstrating that the choice of source language significantly affects performance. We show that choosing the best transfer language improves zero-shot F1 scores by an average of 14 points across 20 languages compared to using English. Our results highlight the need for benchmark datasets and models that cover typologically-diverse African languages. Accepted to EMNLP 2022 (updated Github link)
Similar
MasakhaNER 2.0: Africa-centric Transfer Learning for Named Entity Recognition MasakhaNER 2.0: تعلم التحويل المتمحور حول أفريقيا للتعرف على الكيان المسمى MasakhaNER 2.0 : Apprentissage par transfert centré sur l'Afrique pour la reconnaissance des entités nommées MasakhaNER 2.0: Aprendizaje de transferencia centrado en África para el reconocimiento de entidades nombradasMasakhaNER: Named Entity Recognition for African Languages MasakhaNER: Named Entity Recognition Dataset for 20 African languages.Transfer Learning for Named Entity Recognition in Setswana Language Using CNN-BiLSTM Modelarmin-pousti/Language-Transfer-in-Named-Entity-RecognitionNamed entity recognition for Amharic using deep learning