Deep Active Learning for Sequence Labeling Based on Diversity and Uncertainty in Gradient
This work provides an incremental improvement in active learning strategies for sequence labeling, benefiting researchers and practitioners seeking to reduce annotation costs in NLP.
This paper addresses the challenge of data dependency in sequence labeling tasks by proposing an active learning approach that combines uncertainty and diversity in gradient embeddings for query selection. The method consistently outperforms traditional uncertainty-based and diversity-based sampling across various tasks, datasets, and models, demonstrating a reduction in the required amount of labeled training data.
Recently, several studies have investigated active learning (AL) for natural language processing tasks to alleviate data dependency. However, for query selection, most of these studies mainly rely on uncertainty-based sampling, which generally does not exploit the structural information of the unlabeled data. This leads to a sampling bias in the batch active learning setting, which selects several samples at once. In this work, we demonstrate that the amount of labeled training data can be reduced using active learning when it incorporates both uncertainty and diversity in the sequence labeling task. We examined the effects of our sequence-based approach by selecting weighted diverse in the gradient embedding approach across multiple tasks, datasets, models, and consistently outperform classic uncertainty-based sampling and diversity-based sampling.