1 paper · 1 filter
Qing Su, Shihao Ji
Distillation-based self-supervised learning typically leads to more compressed representations due to its radical clustering process and the implementation of a sharper target dist…