OmniTab: Pretraining with Natural and Synthetic Data for Few-shot Table-based Question Answering.
Zhengbao JiangYi MaoPengcheng HeGraham NeubigWeizhu ChenPublished in: NAACL-HLT (2022)
Keyphrases
- synthetic data
- question answering
- real world
- data sets
- passage retrieval
- natural language processing
- natural language
- qa clef
- real image data
- named entities
- natural language questions
- database
- information extraction
- relation extraction
- question answering systems
- information retrieval
- cross language
- question classification
- syntactic information
- video sequences
- open domain question answering
- sentence retrieval
- video content
- document retrieval
- answering questions
- bag of words
- video data
- textual entailment recognition
- xml documents