AI RESEARCH

COMET: Concept Space Dissection of the Modality Gap in Audio-Text Multimodal Contrastive Embeddings

arXiv CS.AI

ArXi:2605.29628v1 Announce Type: cross Contrastive Language-Audio Pre