Using Error Decay Prediction to Overcome Practical Issues of Deep Active Learning for Named Entity Recognition
This work addresses practical deployment challenges for researchers and practitioners using active learning in NLP, though it is incremental as it builds on existing methods to improve robustness and transparency.
The paper tackled practical issues in deep active learning for named entity recognition, such as incompatibility with black-box models and lack of robustness to noise, by proposing a transparent batch sampling framework based on error decay prediction, which significantly outperformed diversification-based methods on four NER tasks.
Existing deep active learning algorithms achieve impressive sampling efficiency on natural language processing tasks. However, they exhibit several weaknesses in practice, including (a) inability to use uncertainty sampling with black-box models, (b) lack of robustness to labeling noise, and (c) lack of transparency. In response, we propose a transparent batch active sampling framework by estimating the error decay curves of multiple feature-defined subsets of the data. Experiments on four named entity recognition (NER) tasks demonstrate that the proposed methods significantly outperform diversification-based methods for black-box NER taggers, and can make the sampling process more robust to labeling noise when combined with uncertainty-based methods. Furthermore, the analysis of experimental results sheds light on the weaknesses of different active sampling strategies, and when traditional uncertainty-based or diversification-based methods can be expected to work well.