Neurosymbolic Object-Centric Learning with Distant Supervision

ArXi:2506.16129v2 Announce Type: replace Neurosymbolic learning can use symbolic rules to provide supervision for latent concepts from weak labels, but it commonly assumes that the entities referenced by these rules are already specified. Object-centric models decompose images into slot-like representations; however, such slots are not necessarily aligned with the predicates required for symbolic reasoning.