MUTANT: A Training Paradigm for Out-of-Distribution Generalization in Visual Question Answering.
Tejas GokhalePratyay BanerjeeChitta BaralYezhou YangPublished in: EMNLP (1) (2020)
Keyphrases
- question answering
- information extraction
- natural language processing
- cross language
- information retrieval
- natural language
- visual features
- question classification
- named entities
- qa clef
- passage retrieval
- syntactic information
- natural language questions
- training set
- question answering systems
- open domain question answering
- answer validation
- low level
- semantic roles
- candidate answers
- multi modal
- visual information
- test set
- machine learning