Dynamic Fusion with Intra- and Inter- Modality Attention Flow for Visual Question Answering
This addresses the challenge of effectively integrating visual and language features for visual question answering, representing an incremental improvement in a domain-specific task.
The paper tackles the problem of multi-modality feature fusion in visual question answering by proposing a dynamic fusion method with intra- and inter-modality attention flow, which achieves state-of-the-art performance on the VQA 2.0 dataset.
Learning effective fusion of multi-modality features is at the heart of visual question answering. We propose a novel method of dynamically fusing multi-modal features with intra- and inter-modality information flow, which alternatively pass dynamic information between and across the visual and language modalities. It can robustly capture the high-level interactions between language and vision domains, thus significantly improves the performance of visual question answering. We also show that the proposed dynamic intra-modality attention flow conditioned on the other modality can dynamically modulate the intra-modality attention of the target modality, which is vital for multimodality feature fusion. Experimental evaluations on the VQA 2.0 dataset show that the proposed method achieves state-of-the-art VQA performance. Extensive ablation studies are carried out for the comprehensive analysis of the proposed method.