Visual Explanations of Image-Text Representations via Multi-Modal Information Bottleneck Attribution

Open in new window