Multi-modal Alignment using Representation Codebook

Jiali Duan, Liqun Chen, Son Tran +4

Aligning signals from different modalities is an important step in vision-language representation learning as it affects the performance of later stages such as cross-modality fusi…