Low-Resource Cross-Lingual Adaptive Training for Nigerian Pidgin

Domain:

natural language processing

Record type:

paperdataset

Creator:

LinSaeChaSch

Host:

Developing effective spoken language processing systems for low-resource languages poses several challenges due to the lack of parallel data and limited resources for fine-tuning models. In this work, we target on improving upon both text classification and translation of Nigerian Pidgin (Naija) by collecting a large-scale parallel English-Pidgin corpus and further propose a framework of cross-lingual adaptive training that includes both continual and task adaptive training so as to adapt a base pre-trained model to low-resource languages. Our studies show that English pre-trained language models serve as a stronger prior than multilingual language models on English-Pidgin tasks with up to 2.38 BLEU improvements; and demonstrate that augmenting orthographic data and using task adaptive training with back-translation can have a significant impact on model performance. To appear in INTERSPEECH 2023

Visit

arxiv.org

Tasks

machine translation text classification transfer learning

Low Resource Cross Lingual Adaptive Training for Nigerien Pidgin

Low Resource Cross Lingual Adaptive Training for Nigerien Pidgin

Poster presented at the Deep Learning Indaba 2023 by Muhammed Saeed

Multilingual Intermediate-Task Training for Low-Resource Cross-Lingual Transfer

Intermediate-task training---fine-tuning a pretrained model on an intermediate task before fine-tuni

Deep Persian sentiment analysis: Cross-lingual training for low-resource languages

With the advent of deep neural models in natural language processing tasks, having a large amount of

Adversarial Training for Robust Cross-Lingual NER in Low-Resource Settings

Cross-lingual Named Entity Recognition (NER) leverages knowledge transfer between languages to ident

Hybrid Training Strategies for Low-Resource Cross-Lingual NER Sample Efficiency