1 paper
Rubèn Tito, Khanh Nguyen, Marlon Tobaben +11
Document Visual Question Answering (DocVQA) has quickly grown into a central task of document understanding. But despite the fact that documents contain sensitive or copyrighted in…