MELM: Data Augmentation with Masked Entity Language Modeling for Low-Resource NER

TL;DR

This paper introduces MELM, a novel data augmentation method for low-resource NER that explicitly incorporates entity labels into language modeling, improving the quality of augmented data and boosting NER performance across multiple languages.

Contribution

MELM is the first approach to explicitly inject NER labels into masked language modeling for data augmentation, addressing token-label misalignment in low-resource NER tasks.

Findings

MELM significantly improves NER performance in low-resource settings.

MELM outperforms baseline data augmentation methods.

Combining MELM with code-mixing further enhances results.

Abstract

Data augmentation is an effective solution to data scarcity in low-resource scenarios. However, when applied to token-level tasks such as NER, data augmentation methods often suffer from token-label misalignment, which leads to unsatsifactory performance. In this work, we propose Masked Entity Language Modeling (MELM) as a novel data augmentation framework for low-resource NER. To alleviate the token-label misalignment issue, we explicitly inject NER labels into sentence context, and thus the fine-tuned MELM is able to predict masked entity tokens by explicitly conditioning on their labels. Thereby, MELM generates high-quality augmented data with novel entities, which provides rich entity regularity knowledge and boosts NER performance. When training data from multiple languages are available, we also integrate MELM with code-mixing for further improvement. We demonstrate the effectiveness of MELM on monolingual, cross-lingual and multilingual NER across various low-resource levels. Experimental results show that our MELM presents substantial improvement over the baseline methods.