Graph Edit Distance Reward: Learning to Edit Scene Graph
This addresses a novel cross-modality task for scene graph editing in applications like VQA, though it appears incremental in combining existing techniques.
The paper tackles the problem of editing scene graphs based on user instructions, which hasn't been explored before, and proposes a Graph Edit Distance Reward method that combines Policy Gradient and Graph Matching to optimize a neural symbolic model, achieving validation on CSS and a new synthetic CRIR dataset.
Scene Graph, as a vital tool to bridge the gap between language domain and image domain, has been widely adopted in the cross-modality task like VQA. In this paper, we propose a new method to edit the scene graph according to the user instructions, which has never been explored. To be specific, in order to learn editing scene graphs as the semantics given by texts, we propose a Graph Edit Distance Reward, which is based on the Policy Gradient and Graph Matching algorithm, to optimize neural symbolic model. In the context of text-editing image retrieval, we validate the effectiveness of our method in CSS and CRIR dataset. Besides, CRIR is a new synthetic dataset generated by us, which we will publish it soon for future use.