关键词: 人工智能, 文本分类, 小样本学习, 互学习, 原型网络

Abstract: Few-shot prototype networks are regarded as one of the effective methods to solve few-shot text classification problems. However, existing methods usually rely only on a single prototype for training and inference, which is susceptible to noise and other factors, resulting in insufficient generalization ability. To address this problem, a Mutual Learn-ing-Prototype Network(MLProtoNet) for small-sample text classification is proposed. On the basis of retaining the ex-isting algorithm to compute the prototype directly by text embedding features, thie paper introduces the BERT network, which inputs the text embedding features into BERT to generate a new prototype. Then, using the mutual learning algorithm, the two prototypes are mutually constrained and knowledge is exchanged to filter out the inaccurate semantic information. This process aims to enhance the feature extraction capability of the model and improve the classification accuracy by joint decision making of the two prototypes. Experimental results on two few-shot text classification da-tasets confirm the effectiveness of our proposed approach. Specifically, on the FewRel dataset, our method improves the accuracy by 2.97% in the 5-way 1-shot experiment compared to the current optimal method, and by 1.99% in the 5-way 5-shot experiment.

Key words: artificial intelligence, text classification, few-shot learning, mutual learning, prototype network