In this paper, we study the problem of geometric reasoning in the context of question-answering. We introduce Dynamic Spatial Memory Network (DSMN), a new deep network architecture designed for answering questions that admit latent visual representations. DSMN learns to generate and reason over such representations. Further, we propose two synthetic benchmarks, FloorPlanQA and ShapeIntersection, to evaluate the geometric reasoning capability of QA systems. Experimental results validate the effectiveness of our proposed DSMN for visual thinking tasks.
@article{arxiv.1805.11025,
title = {Think Visually: Question Answering through Virtual Imagery},
author = {Ankit Goyal and Jian Wang and Jia Deng},
journal= {arXiv preprint arXiv:1805.11025},
year = {2018}
}