Self-Supervised Transfer

Embedding Space

Low-dimensional vector space where data is projected to capture their semantic and structural relationships. In self-supervised learning, the goal is to learn a discriminative embedding space.

← Indietro