Visual Question Answering Dataset for Bilingual Image Understanding: A Study of Cross-Lingual Transfer Using Attention Maps

5354

Knowledge Triples

Source Entity	Relationship	Target Entity
Attention Maps Material	improves	Visual Question Answering Task
Attention Mechanism Method	creates	Attention Maps Material
Cross Lingual Method Method	enhances	Visual Question Answering Task
Cross Lingual Method Method	isBasedOn	Attention Mechanism Method
Japanese VQA Dataset Material	isBasedOn	Visual Genome Dataset Material
Visual Question Answering Task	uses	Japanese VQA Dataset Material