Illuminating Visual Identity in Universal Multimodal Embeddings

Jiawei Cao, Junyi Feng, Jiashen Hua +5

Universal Multimodal Embeddings (UMEs) aim to unify various modalities and tasks into a shared representation space. In recent years, this field has witnessed substantial progress…