Research Frontiers in Contrastive Learning for Self-Supervised Source Representation
Application of contrastive learning techniques to learn generalizable source representations from unlabeled multi-wavelength astronomical data without ground truth labels.