Research map / Vision-language and multimodal / Human and embodied modeling
Generating Visual Explanations
Lisa Anne Hendricks, Zeynep Akata, Marcus Rohrbach, Jeff Donahue, Bernt Schiele, Trevor Darrell
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add discriminative class loss (with label conditioning) (ablation) | Explanation model · CUB-200 bird explanations | CIDEr: 44.7 -> 56.7 | Improved |
| add discriminative class loss (with label conditioning) (ablation) | Explanation model · CUB-200 bird explanations | class similarity: 40.86 -> 52.25 | Improved |
| add discriminative class loss (with label conditioning) (ablation) | Explanation model · CUB-200 bird explanations | class rank (lower better): 17.69 -> 13.12 | Improved |
| add label conditioning (with discriminative loss) (ablation) | Explanation model · CUB-200 bird explanations | CIDEr: 51.9 -> 56.7 | Improved |
| add label conditioning (with discriminative loss) (ablation) | Explanation model · CUB-200 bird explanations | class rank (lower better): 19.80 -> 13.12 | Improved |
| add label conditioning (with discriminative loss) (ablation) | Explanation model · CUB-200 bird explanations | bird expert rank (lower better): 3.22 -> 2.78 | Improved |
| add discriminative loss to label-only model (ablation) | Explanation-Label · CUB-200 bird explanations | METEOR: 28.1 -> 29.2 | Improved |
| add class label conditioning to description model | Description baseline · CUB-200 bird explanations | CIDEr: 42.0 -> 44.7 | Improved |
| add discriminative loss to Description-style model without label | Description baseline · CUB-200 bird explanations | CIDEr: 42.0 -> 51.9 | Improved |
| condition on class label (definition vs description) | Description baseline · CUB-200 bird explanations | class rank (lower better): 24.43 -> 15.82 | Improved |
| use explanation model instead of description for sentence-classifier accuracy | Description baseline · CUB-200 bird explanations | sentence classifier accuracy (%): 22.32% -> 59.13% | Improved |
Closest papers by meaning
- ProAPO: Progressively Automatic Prompt Optimization for Visual Classification (CVPR 2025 · similarity 0.87)
- Language Model as Visual Explainer (NeurIPS 2024 · similarity 0.87)
- Explainable Concept Generation through Vision-Language Preference Learning for Understanding Neural Networks' Internal Representations (ICML 2025 · similarity 0.87)
- Explaining in Diffusion: Explaining a Classifier with Diffusion Semantics (CVPR 2025 · similarity 0.87)
- Modeling Collaborator: Enabling Subjective Vision Classification With Minimal Human Effort via LLM Tool-Use (CVPR 2024 · similarity 0.86)
- Follow-Up Differential Descriptions: Language Models Resolve Ambiguities for Image Classification (ICLR 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
