DISCOVER visual interpretability paper is published!

DISCOVER is a generative model for visual interpretability of image-based classification models. DISCOVER’s disentanglement module enables visual interpretation of one semantic classification-driving feature at a time. We applied DISCOVER to interpret classification of the morphology quality of in vitro fertilization embryos.

Project led by Oded Rotem in collaboration with AIVF.

Paper: https://www.nature.com/articles/s41467-024-51136-9

Brief Twitter summary: https://x.com/AssafZaritsky/status/1829016695299571958