Empirical Analysis of Unlabeled Entity Problem in Named Entity Recognition
This work addresses a practical data annotation challenge for anyone developing NER models, particularly in scenarios with incomplete annotations, offering a robust solution.
This paper investigates the 'unlabeled entity problem' in Named Entity Recognition (NER), where entities are not fully annotated. They identify that treating unlabeled entities as negative instances is the primary cause of performance degradation, and propose a negative sampling approach to mitigate this, resulting in a model robust to this problem and competitive with SOTA on well-annotated datasets.
In many scenarios, named entity recognition (NER) models severely suffer from unlabeled entity problem, where the entities of a sentence may not be fully annotated. Through empirical studies performed on synthetic datasets, we find two causes of performance degradation. One is the reduction of annotated entities and the other is treating unlabeled entities as negative instances. The first cause has less impact than the second one and can be mitigated by adopting pretraining language models. The second cause seriously misguides a model in training and greatly affects its performances. Based on the above observations, we propose a general approach, which can almost eliminate the misguidance brought by unlabeled entities. The key idea is to use negative sampling that, to a large extent, avoids training NER models with unlabeled entities. Experiments on synthetic datasets and real-world datasets show that our model is robust to unlabeled entity problem and surpasses prior baselines. On well-annotated datasets, our model is competitive with the state-of-the-art method.