OmniTab: Pretraining with Natural and Synthetic Data for Few-shot Table-based Question Answering.
Zhengbao JiangYi MaoPengcheng HeGraham NeubigWeizhu ChenPublished in: CoRR (2022)
Keyphrases
- synthetic data
- question answering
- real world
- information retrieval
- data sets
- information extraction
- natural language processing
- passage retrieval
- qa clef
- real image data
- cross language
- natural language
- relation extraction
- database
- syntactic information
- question classification
- question answering systems
- natural language questions
- video sequences
- video data
- visual features
- open domain question answering
- answer extraction
- named entities
- qa systems
- semantic roles
- document retrieval
- sentence retrieval
- answering questions
- structured data
- language model