Multimodal Archive Indices

Foundation models that embed heterogeneous media—books, manuscripts, scans, oral histories, and films—into a shared vector space. Researchers search conceptually rather than by format, enabling cross-media discovery like finding photographs that match a described historical event.