Mix of Experts Language Model for Named Entity Recognition

TL;DR

This paper introduces BOND-MoE, a robust NER model utilizing a mixture of experts and EM framework to mitigate noise from distant supervision, achieving state-of-the-art results.

Contribution

The paper proposes a novel MoE-based NER model with a fair assignment module and EM training to improve robustness against noisy annotations.

Findings

Achieves state-of-the-art performance on real-world datasets.

Effectively alleviates noise from distant supervision.

Demonstrates robustness through extensive experiments.

Abstract

Named Entity Recognition (NER) is an essential steppingstone in the field of natural language processing. Although promising performance has been achieved by various distantly supervised models, we argue that distant supervision inevitably introduces incomplete and noisy annotations, which may mislead the model training process. To address this issue, we propose a robust NER model named BOND-MoE based on Mixture of Experts (MoE). Instead of relying on a single model for NER prediction, multiple models are trained and ensembled under the Expectation-Maximization (EM) framework, so that noisy supervision can be dramatically alleviated. In addition, we introduce a fair assignment module to balance the document-model assignment process. Extensive experiments on real-world datasets show that the proposed method achieves state-of-the-art performance compared with other distantly supervised NER.