WeaQA: Weak Supervision via Captions for Visual Question Answering

33882

Knowledge Triples

Source Entity	Relationship	Target Entity
Spatial Pyramid Image Patches Method	improves	Visual Question Answering Task
Synthetic Q A pairs Method	appliesTo	Visual Question Answering Task
Visual Question Answering Task	requires	Image Question Answer triplet Material
VQA CP Challenge Task	tests	Visual Question Answering Task