A Survey on Transfer Learning in Natural Language Processing

TL;DR

This survey reviews recent advances in transfer learning for NLP, highlighting methods to address data scarcity and resource demands, and provides a taxonomy of different approaches.

Contribution

It offers a comprehensive overview and categorization of transfer learning techniques in NLP, aiding understanding of recent developments.

Findings

Various transfer learning methods improve NLP performance.

Taxonomy clarifies different transfer learning approaches.

Highlights challenges and future directions in NLP transfer learning.

Abstract

Deep learning models usually require a huge amount of data. However, these large datasets are not always attainable. This is common in many challenging NLP tasks. Consider Neural Machine Translation, for instance, where curating such large datasets may not be possible specially for low resource languages. Another limitation of deep learning models is the demand for huge computing resources. These obstacles motivate research to question the possibility of knowledge transfer using large trained models. The demand for transfer learning is increasing as many large models are emerging. In this survey, we feature the recent transfer learning advances in the field of NLP. We also provide a taxonomy for categorizing different transfer learning approaches from the literature.