End-to-End Synthetic Data Generation for Domain Adaptation of Question Answering Systems.
Siamak ShakeriCícero Nogueira dos SantosHenghui ZhuPatrick NgFeng NanZhiguo WangRamesh NallapatiBing XiangPublished in: EMNLP (1) (2020)
Keyphrases
- end to end
- domain adaptation
- data generation
- question answering systems
- co training
- question answering
- semi supervised
- active learning
- labeled data
- natural language
- semi supervised learning
- multiple sources
- document retrieval
- unlabeled data
- data streams
- information retrieval systems
- streaming data
- cross domain
- multi view
- sentiment classification
- transfer learning
- high throughput
- information extraction
- named entities
- information retrieval
- learning algorithm
- single view
- training examples
- data mining
- text classification
- supervised learning
- feature space
- high dimensional
- training data
- search engine
- prior knowledge