Improving Model Generalization: A Chinese Named Entity Recognition Case Study

Abstract

Generalization is an important ability that helps to ensure that a machine learning model can perform well on unseen data. In this paper, we study the effect of data bias on model generalization, using Chinese Named Entity Recognition (NER) as a case study. Specifically, we analyzed five benchmarking datasets for Chinese NER, and observed the following two types of data bias that may compromise model generalization ability. Firstly, the test sets of all the five datasets contain a significant proportion of entities that have been seen in the training sets. These test data are therefore not suitable for evaluating how well a model can handle unseen data. Secondly, all datasets are dominated by a few fat-head entities, i.e., entities appearing with particularly high frequency. As a result, a model might be able to produce high prediction accuracy simply by keyword memorization. To address these data biases, we first refine each test set by excluding seen entities from it, so as to better evaluate a model's generalization ability. Then, we propose a simple yet effective entity rebalancing method to make entities within the same category distributed equally, encouraging a model to leverage both name and context knowledge in the training process. Experimental results demonstrate that the proposed entity resampling method significantly improves a model's ability in detecting unseen entities, especially for company, organization and position categories.