Transparency by Design: Closing the Gap Between Performance and Interpretability in Visual Reasoning

• Published in: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition pp. 4942 - 4950
• Main Authors: Mascharka, David, Tran, Philip, Soklaski, Ryan, Majumdar, Arjun
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.06.2018
• Subjects: Automobiles; Cognition; Image color analysis; Knowledge discovery; Neural networks; Task analysis; Visualization

Visual question answering requires high-order reasoning about an image, which is a fundamental capability needed by machine systems to follow complex directives. Recently, modular networks have been shown to be an effective framework for performing visual reasoning tasks. While modular networks were...