Proceedings

Learning Unsupervised Visual Grounding Through Semantic Self-Supervision

Syed Ashar Javed, Shreyas Saxena, Vineet Gandhi

Download BibTeX Cite

Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence. Main track. Pages 796-802.

https://doi.org/10.24963/ijcai.2019/112

Keywords

  • Computer Vision: Language and Vision
  • Machine Learning: Unsupervised Learning
  • Computer Vision: Recognition: Detection, Categorization, Indexing, Matching, Retrieval, Semantic Interpretation

Abstract

Localizing natural language phrases in images is a challenging problem that requires joint understanding of both the textual and visual modalities. In the unsupervised setting, lack of supervisory signals exacerbate this difficulty. In this paper, we propose a novel framework for unsupervised visual grounding which uses concept learning as a proxy task to obtain self-supervision. The intuition behind this idea is to encourage the model to localize to regions which can explain some semantic property in the data, in our case, the property being the presence of a concept in a set of images. We present thorough quantitative and qualitative experiments to demonstrate the efficacy of our approach and show a 5.6% improvement over the current state of the art on Visual Genome dataset, a 5.8% improvement on the ReferItGame dataset and comparable to state-of-art performance on the Flickr30k dataset.

Syed Ashar Javed, Shreyas Saxena, and Vineet Gandhi. 2019. Learning Unsupervised Visual Grounding Through Semantic Self-Supervision. In Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence, IJCAI-19. IJCAI, 796–802. https://doi.org/10.24963/ijcai.2019/112
Welcome Future Conferences AI Journal Division IJCAI Proceedings About IJCAI IJCAI Awards