Low-Rank HOCA: Efficient High-Order Cross-Modal Attention for Video Captioning.
Tao JinSiyu HuangYingming LiZhongfei ZhangPublished in: EMNLP/IJCNLP (1) (2019)
Keyphrases
- high order
- low rank
- cross modal
- higher order
- multi modal
- tensor decomposition
- pairwise
- video data
- multimedia
- multimedia retrieval
- visual data
- singular value decomposition
- sparse representation
- matrix factorization
- video content
- video frames
- linear combination
- denoising
- semi supervised
- image data
- pattern recognition
- multiscale