Controlled benchmark studying whether DANN domain adaptation improves retrieval or damages embedding neighborhood structure.
machine-learning computer-vision retrieval pytorch representation-learning model-evaluation domain-adaptation synthetic-data dann cxr medical-ai contrastive-learning gradient-reversal embedding-analysis domain-adversarial-training
-
Updated
Jun 7, 2026 - Python