A Self-Supervised Deep Learning Framework for Unsupervised Few-Shot Learning and Clustering

Hongjing Zhang, Tianyang Zhan, Ian Davidson

Abstract: Highlights•Unsupervised representation learning for few-shot learning and clustering.•Novel self-supervised pre-task generation.•Method improves the clustering accuracy of recent deep clustering methods.•Method improves the accuracy of unsupervised few-shot learning baselines.•Ablation study demonstrates the contributions of different technical components.