Shortcuts and Identifiability in Concept-based Models from a Neuro-Symbolic Lens

Open in new window