Learning with Difference Attention for Visually Grounded Self-supervised Representations
This addresses the limitation of SSL models in handling complex images for tasks like segmentation, but it is incremental as it builds on existing SSL frameworks.
The paper tackled the problem of self-supervised learning methods performing poorly on multi-object images due to weak visual grounding, and proposed a Differentiable Difference Attention loss that substantially improved grounding to salient regions.
Recent works in self-supervised learning have shown impressive results on single-object images, but they struggle to perform well on complex multi-object images as evidenced by their poor visual grounding. To demonstrate this concretely, we propose visual difference attention (VDA) to compute visual attention maps in an unsupervised fashion by comparing an image with its salient-regions-masked-out version. We use VDA to derive attention maps for state-of-the art SSL methods and show they do not highlight all salient regions in an image accurately, suggesting their inability to learn strong representations for downstream tasks like segmentation. Motivated by these limitations, we cast VDA as a differentiable operation and propose a new learning objective, Differentiable Difference Attention (DiDA) loss, which leads to substantial improvements in an SSL model's visually grounding to an image's salient regions.