Dynamic Multistep Reasoning based on Video Scene Graph for Video Question Answering.
Jianguo MaoWenbin JiangXiangdong WangZhifan FengYajuan LyuHong LiuYong ZhuPublished in: NAACL-HLT (2022)
Keyphrases
- question answering
- video scene
- video data
- scene analysis
- video analysis
- natural language processing
- information retrieval
- video sequences
- information extraction
- dynamic scenes
- answering questions
- video shots
- natural language
- named entities
- multimedia
- video content
- video clips
- question answering systems
- video streams
- feature selection
- answer extraction
- qa systems
- event detection
- video search
- multiple features
- foreground objects
- text mining
- knowledge base
- artificial intelligence