Modulated Self-attention Convolutional Network for VQA
This is an incremental contribution aimed at improving visual processing for VQA tasks.
The paper tackles the problem of visual feature extraction for visual question answering (VQA) by proposing a modulated self-attention convolutional network, showing encouraging relative improvements.
As new data-sets for real-world visual reasoning and compositional question answering are emerging, it might be needed to use the visual feature extraction as a end-to-end process during training. This small contribution aims to suggest new ideas to improve the visual processing of traditional convolutional network for visual question answering (VQA). In this paper, we propose to modulate by a linguistic input a CNN augmented with self-attention. We show encouraging relative improvements for future research in this direction.