Unlocking the Potential of Multimodal Unified Discrete Representation through Training-Free Codebook Optimization and Hierarchical Alignment.
Hai HuangYan XiaShengpeng JiShulei WangHanting WangJieming ZhuZhenhua DongZhou ZhaoPublished in: CoRR (2024)
Keyphrases
- image representation
- training set
- hierarchical representation
- vector quantization
- translation invariant
- hierarchical decomposition
- morse theory
- optimization problems
- optimization algorithm
- optimization process
- continuous domains
- binary partition tree
- continuous optimization
- multimodal interaction
- bag of words
- multi modal
- hierarchical structure
- training examples
- evolutionary algorithm
- pairwise
- multimodal function optimization
- multiscale