1 paper · 1 filter
Suraj Jyothi Unni, Raha Moraffah, Huan Liu
Visual question answering (VQA) models are designed to demonstrate visual-textual reasoning capabilities. However, their real-world applicability is hindered by a lack of comprehen…