N1H111SM's Miniverse

Contrastive Learning from the Perspective of Manifold Learning

字数统计: 77阅读时长: 1 min
2020/06/15 Share

In Isomap method, the target of representation space is to keep the Geodesic Distance between arbitrary two points:

In contrastive learning, the target of representation space is that there exists a critic that can distinguish positive/negative sample pairs. The procedure of optimizing the contrastive loss is

CATALOG