Self-Supervised Setting

AbstractIn this paper, we develop a method for grounding medical text into a physically meaningful and interpretable space corresponding to a human atlas. We build on text embedding architectures such …...