Deep Modular Co-Attention Networks for Visual Question Answering

TL;DR

This paper introduces a deep Modular Co-Attention Network (MCAN) for Visual Question Answering that effectively models question-image interactions through cascaded attention layers, significantly improving accuracy over previous models.

Contribution

The paper proposes a novel deep co-attention architecture with modular layers that enhance interaction modeling in VQA, outperforming shallow models and previous state-of-the-art methods.

Findings

MCAN achieves 70.63% accuracy on VQA-v2 test-dev.

Deep modular co-attention layers outperform shallow models.

Extensive ablation studies validate the effectiveness of the proposed architecture.

Abstract

Visual Question Answering (VQA) requires a fine-grained and simultaneous understanding of both the visual content of images and the textual content of questions. Therefore, designing an effective `co-attention' model to associate key words in questions with key objects in images is central to VQA performance. So far, most successful attempts at co-attention learning have been achieved by using shallow models, and deep co-attention models show little improvement over their shallow counterparts. In this paper, we propose a deep Modular Co-Attention Network (MCAN) that consists of Modular Co-Attention (MCA) layers cascaded in depth. Each MCA layer models the self-attention of questions and images, as well as the guided-attention of images jointly using a modular composition of two basic attention units. We quantitatively and qualitatively evaluate MCAN on the benchmark VQA-v2 dataset and conduct extensive ablation studies to explore the reasons behind MCAN's effectiveness. Experimental results demonstrate that MCAN significantly outperforms the previous state-of-the-art. Our best single model delivers 70.63$\%$ overall accuracy on the test-dev set. Code is available at https://github.com/MILVLG/mcan-vqa.