Enhanced Meta-Learning for Cross-lingual Named Entity Recognition with Minimal Resources
This addresses the problem of named entity recognition for low-resource languages, offering an incremental improvement over existing transfer methods.
The paper tackles cross-lingual named entity recognition for languages with no annotated resources by proposing a meta-learning algorithm that fine-tunes models with similar examples per test case, achieving significant outperformance over state-of-the-art methods across five target languages.
For languages with no annotated resources, transferring knowledge from rich-resource languages is an effective solution for named entity recognition (NER). While all existing methods directly transfer from source-learned model to a target language, in this paper, we propose to fine-tune the learned model with a few similar examples given a test case, which could benefit the prediction by leveraging the structural and semantic information conveyed in such similar examples. To this end, we present a meta-learning algorithm to find a good model parameter initialization that could fast adapt to the given test case and propose to construct multiple pseudo-NER tasks for meta-training by computing sentence similarities. To further improve the model's generalization ability across different languages, we introduce a masking scheme and augment the loss function with an additional maximum term during meta-training. We conduct extensive experiments on cross-lingual named entity recognition with minimal resources over five target languages. The results show that our approach significantly outperforms existing state-of-the-art methods across the board.