发帖
雷达卡
Understanding Self-Supervised Learning Dynamics without Contrastive Pairs
Yuandong Tian 1 Xinlei Chen 1 Surya Ganguli 1 2
Abstract quiring expensive target labels (Devlin et al., 2018). Many
state-of-the-art SSL methods in computer vision employ
While contrastive approaches of self-supervised the principle of contrastive learning (Oord et al., 2018; Tian
learning (SSL) learn representations by minimiz- ...
京公网安备 11010802022788号
