Sparse and Structured Visual Attention

13 Feb 2020  ·  Pedro Henrique Martins, Vlad Niculae, Zita Marinho, André Martins ·

Visual attention mechanisms are widely used in multimodal tasks, as visual question answering (VQA). One drawback of softmax-based attention mechanisms is that they assign some probability mass to all image regions, regardless of their adjacency structure and of their relevance to the text. In this paper, to better link the image structure with the text, we replace the traditional softmax attention mechanism with two alternative sparsity-promoting transformations: sparsemax, which is able to select only the relevant regions (assigning zero weight to the rest), and a newly proposed Total-Variation Sparse Attention (TVmax), which further encourages the joint selection of adjacent spatial locations. Experiments in VQA show gains in accuracy as well as higher similarity to human attention, which suggests better interpretability.

PDF Abstract


Results from the Paper

  Submit results from this paper to get state-of-the-art GitHub badges and help the community compare results to other papers.