HERO: Hierarchical Encoder for Video+Language Omni-representation Pre-training.
Linjie LiYen-Chun ChenYu ChengZhe GanLicheng YuJingjing LiuPublished in: CoRR (2020)
Keyphrases
- language learning
- hierarchical representation
- representation language
- video sequences
- natural language
- video data
- video analysis
- multimedia
- training set
- temporal structure
- highly expressive
- real time
- training process
- video clips
- low complexity
- hierarchical structure
- training examples
- video content
- rate control
- compressed video
- mpeg standard
- hierarchical tree
- programming language