Knowledge transfer based hierarchical few-shot learning via tree-structured knowledge graph

Few-shot learning poses a great challenge for obtaining a classifier that recognizes new classes from a few labeled examples. Existing solutions perform well by leveraging meta-learning models driven by data information. However, these models only utilize the flat data information and ignore the existing hierarchical knowledge structure among classes. In this paper, we propose a knowledge transfer based hierarchical few-shot learning model, which takes advantage of a tree-structured knowledge graph to facilitate the classification results. First, we consider a tree-structured class hierarchy according to the semantic information among classes as a knowledge graph to alleviate the low-data problem. Second, we divide the tree structure into class structure and data, and build a multi-layer classifier to obtain classification results in the two parts. Finally, we consider the tradeoff between structure loss and data loss for hierarchical few-shot learning, which takes class structure information to assist learning. Experimental results on benchmark datasets show that our model outperforms several state-of-the-art models.