1 paper
Xianzhong Long, Chen Peng, Yun Li
In contrastive self-supervised learning, positive samples are typically drawn from the same image but in different augmented views, resulting in a relatively limited source of posi…