DocPrompt: Large-scale continue pretrain for zero-shot and few-shot document question answering.
Sijin WuDan ZhangTeng HuShikun FengPublished in: CoRR (2023)
Keyphrases
- question answering
- passage retrieval
- information retrieval
- text summarization
- candidate answers
- natural language
- natural language processing
- document collections
- cross language
- document retrieval
- information extraction
- question classification
- named entities
- information retrieval systems
- retrieval systems
- document clustering
- qa clef
- question answering systems
- text documents
- natural language questions
- video content
- document classification
- machine learning
- video shots
- video sequences
- relevant documents
- keywords
- web documents
- sentence retrieval
- text retrieval
- vector space model
- relational databases
- syntactic information
- text mining
- answer extraction
- language model