APPN: An Attention-based Pseudo-label Propagation Network for few-shot learning with noisy labels

Jiaqi Chen, Shizhuo Deng, Da Teng, Dongyue Chen, Tong Jia, Hao Wang

Abstract: Highlights•A novel attention-based pseudo-label propagation network improves the accuracy of few-shot classification.•Optimized approach for traditional label propagation is proposed to enhance the accuracy of label propagation.•Our model achieves better robustness and superiority in few-shot learning with noisy labels.