Modal Alignment

Joint Representation Learning

Process of simultaneously learning shared features between multiple modalities to create a unified representation. Captures inter-modal correlations and complementarities in a single vector.

← Zurück