Annotation Errors and NER: A Study with OntoNotes 5.0

TL;DR

This study identifies and corrects annotation errors in the OntoNotes 5.0 NER dataset using simple techniques, leading to improved model performance and demonstrating the importance of dataset quality in NLP tasks.

Contribution

The paper introduces three simple, largely language-agnostic techniques for detecting annotation errors in large NER datasets, significantly improving data quality and model accuracy.

Findings

Corrected ~10% of sentences in OntoNotes 5.0

Improved NER model F-scores by 1.23% on average

Large improvements (>10%) for some entity types

Abstract

Named Entity Recognition (NER) is a well-studied problem in NLP. However, there is much less focus on studying NER datasets, compared to developing new NER models. In this paper, we employed three simple techniques to detect annotation errors in the OntoNotes 5.0 corpus for English NER, which is the largest available NER corpus for English. Our techniques corrected ~10% of the sentences in train/dev/test data. In terms of entity mentions, we corrected the span and/or type of ~8% of mentions in the dataset, while adding/deleting/splitting/merging a few more. These are large numbers of changes, considering the size of OntoNotes. We used three NER libraries to train, evaluate and compare the models trained with the original and the re-annotated datasets, which showed an average improvement of 1.23% in overall F-scores, with large (>10%) improvements for some of the entity types. While our annotation error detection methods are not exhaustive and there is some manual annotation effort involved, they are largely language agnostic and can be employed with other NER datasets, and other sequence labelling tasks.