G3raphGround: Graph-Based Language Grounding

In this paper we present an end-to-end framework for grounding of phrases in images. In contrast to previous works, our model, which we call G 3 RAPHGROUND, uses graphs to formulate more complex, non-sequential dependencies among proposal image regions and phrases. We capture intra-modal dependencies using a separate graph neural network for each modality (visual and lingual), and then use conditional message-passing in another graph neural network to fuse their outputs and capture crossmodal relationships. This final representation results in grounding decisions. The framework supports many-tomany matching and is able to ground single phrase to multiple image regions and vice versa. We validate our design choices through a series of ablation studies and illustrate state-of-the-art performance on Flickr30k and ReferIt Game benchmark datasets.