Skip to main content
MuCoT: Multilingual Contrastive Training for Question-Answering in Low-resource Languages
Proceedings of the Second Workshop on Speech and Language Technologies for Dravidian Languages
  • Gokul Karthik Kumar, Mohamed bin Zayed University of Artificial Intelligence
  • Abhishek Singh Gehlot, Mohamed bin Zayed University of Artificial Intelligence
  • Sahal Shaji Mullappilly, Mohamed bin Zayed University of Artificial Intelligence
  • Karthik Nandakumar, Mohamed bin Zayed University of Artificial Intelligence
Document Type
Conference Proceeding

Accuracy of English-language Question Answering (QA) systems has improved significantly in recent years with the advent of Transformer-based models (e.g., BERT). These models are pre-trained in a self-supervised fashion with a large English text corpus and further fine-tuned with a massive English QA dataset (e.g., SQuAD). However, QA datasets on such a scale are not available for most of the other languages. Multi-lingual BERT-based models (mBERT) are often used to transfer knowledge from high-resource languages to low-resource languages. Since these models are pre-trained with huge text corpora containing multiple languages, they typically learn language-agnostic embeddings for tokens from different languages. However, directly training an mBERT-based QA system for low-resource languages is challenging due to the paucity of training data. In this work, we augment the QA samples of the target language using translation and transliteration into other languages and use the augmented data to fine-tune an mBERT-based QA model, which is already pre-trained in English. Experiments on the Google ChAII dataset show that fine-tuning the mBERT model with translations from the same language family boosts the question-answering performance, whereas the performance degrades in the case of cross-language families. We further show that introducing a contrastive loss between the translated question-context feature pairs during the fine-tuning process, prevents such degradation with cross-lingual family translations and leads to marginal improvement. The code for this work is available at Copyright © 2022, The Authors. All rights reserved.

Publication Date
  • Natural language processing systems,
  • Translation (languages),
  • Artificial Intelligence (cs.AI),
  • Computation and Language (cs.CL),
  • Machine Learning (cs.LG),
  • English languages,
  • Fine tuning,
  • Learn+,
  • Low resource languages,
  • Model-based OPC,
  • Multiple languages,
  • Performance,
  • Question Answering,
  • Question answering systems,
  • Text corpora

Archived with thanks to ACL Anthology

License: CC by 4.0

Uploaded 05 October 2022

Citation Information
G.K. Kumar, A.S. Gehlot, S.S. Mullappilly, and K. Nandakumar, "MuCoT: Multilingual Contrastive Training for Question-Answering in Low-resource Languages", in Proc. of the 2nd Workshop on Speech and Language Tech. for Dravidian Languages, , Dublin, Ireland, 2022, pp. 15-24, doi: 10.18653/v1/2022.dravidianlangtech-1.3