Author name: Tim R. Davidson, Benoit Seguin, Enrico Bacis, Cesar Ilharco, Hamza Harkous

arXiv:2603.29791v1 Announce Type: cross
Abstract: Although many AI applications of interest require specialized multi-modal models, relevant data to train such models is inherently scarce or inaccessible. Filling these gaps with human annotators is pr…