Video Question Answering via Gradually Refined Attention over Appearance and Motion.
Dejing XuZhou ZhaoJun XiaoFei WuHanwang ZhangXiangnan HeYueting ZhuangPublished in: ACM Multimedia (2017)
Keyphrases
- question answering
- object motion
- dynamic textures
- space time
- information retrieval
- visual data
- key frames
- image sequences
- natural language processing
- natural language
- cross language
- video sequences
- question answering systems
- qa clef
- video content
- named entities
- video data
- information extraction
- question classification
- answer validation
- human motion
- natural language questions
- passage retrieval
- video frames
- multimedia
- relation extraction
- human actions
- candidate answers
- textual entailment recognition
- sentence retrieval
- answering questions
- open domain question answering
- syntactic information
- semantic roles
- relational databases
- knowledge base