Are Data Augmentation Methods in Named Entity Recognition Applicable for Uncertainty Estimation?
This addresses the problem of miscalibration in DNNs for NER in safety-critical fields like healthcare and finance, but it is incremental as it builds on existing data augmentation methods.
The study examined how data augmentation affects confidence calibration and uncertainty estimation in Named Entity Recognition (NER), finding that it improves calibration and uncertainty, particularly in in-domain settings, with lower perplexity and larger augmentation sizes enhancing effectiveness.
This work investigates the impact of data augmentation on confidence calibration and uncertainty estimation in Named Entity Recognition (NER) tasks. For the future advance of NER in safety-critical fields like healthcare and finance, it is essential to achieve accurate predictions with calibrated confidence when applying Deep Neural Networks (DNNs), including Pre-trained Language Models (PLMs), as a real-world application. However, DNNs are prone to miscalibration, which limits their applicability. Moreover, existing methods for calibration and uncertainty estimation are computational expensive. Our investigation in NER found that data augmentation improves calibration and uncertainty in cross-genre and cross-lingual setting, especially in-domain setting. Furthermore, we showed that the calibration for NER tends to be more effective when the perplexity of the sentences generated by data augmentation is lower, and that increasing the size of the augmentation further improves calibration and uncertainty.