Weakly-Supervised Visual-Retriever-Reader for Knowledge-based Question Answering
31500Knowledge Triples
| Source Entity | Relationship | Target Entity |
|---|---|---|
|
OK VQA
Material |
comparesWith |
ConceptNet
Material |
|
OK VQA
Material |
comparesWith |
Wikipedia
Material |
|
Visual Retriever Reader
Method |
includes |
Classification Reader
Method |
|
Visual Retriever Reader
Method |
includes |
Extraction Reader
Method |