Attend What You Need: Motion-Appearance Synergistic Networks for Video Question Answering.
Ahjeong SeoGi-Cheon KangJoonhan ParkByoung-Tak ZhangPublished in: ACL/IJCNLP (1) (2021)
Keyphrases
- question answering
- object motion
- dynamic textures
- space time
- image sequences
- key frames
- video sequences
- information extraction
- question classification
- natural language processing
- information retrieval
- named entities
- qa clef
- video data
- visual data
- video content
- syntactic information
- human motion
- cross language
- natural language questions
- relation extraction
- passage retrieval
- open domain question answering
- multimedia
- question answering systems
- natural language
- candidate answers
- answer validation
- artificial intelligence
- video shots
- knowledge representation
- answer extraction