Active Discriminative Word Embedding Learning

We propose a new active learning (AL) method for text classification based on convolutional neural networks (CNNs). In AL, one selects the instances to be manually labeled with the aim of maximizing model performance with minimal effort. Neural models capitalize on word embeddings as features, tuning these to the task at hand. We argue that AL strategies for neural text classification should focus on selecting instances that most affect the embedding space (i.e., induce discriminative word representations). This is in contrast to traditional AL approaches (e.g.,uncertainty sampling), which specify higher level objectives. We propose a simple approach that selects instances containing words whose embeddings are likely to be updated with the greatest magnitude, thereby rapidly learning discriminative, task-specific embeddings. Empirical results show that our method outperforms baseline AL approaches.