Research map / Image generation / Autoregressive and latent generation
Diffusion Visual Counterfactual Explanations
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| use DVCE instead of l1.5-SVCE for generating VCEs | MNR-RN50 · ImageNet crossover evaluation | Avg. FID: 25.6 -> 17.6 | Helped |
| use DVCE instead of l1.5-SVCE | MNR-RN50 · ImageNet crossover evaluation | closeness (LPIPS-Alex): 0.20 -> 0.35 | Hurt |
| use DVCE instead of Blended Diffusion | MNR-RN50 · ImageNet crossover evaluation | Mean Conf.: 0.825 -> 0.932 | Helped |
| use DVCE instead of Blended Diffusion | MNR-RN50 · User study | % rated realistic: 52.2% -> 34.7% | Hurt |
| use DVCE instead of l1.5-SVCE | MNR-RN50 · User study | % rated meaningful: 48.4% -> 62.0% | Helped |
| add cone projection with robust model gradient (ablation) | DVCE on non-robust Swin-TF/ConvNeXt · ImageNet qualitative | semantic meaningfulness of changes | Helped |
| cone angle at or below 15 degrees vs at least 30 degrees (ablation) | DVCE cone projection · ConvNeXt with MNR-RN50, qualitative | class-specific features | Helped |
| start diffusion at T/2 instead of from noise (ablation) | DVCE · qualitative | closeness vs realism trade-off | Helped |
| vary l1 distance weight C_d from 0.05 to 0.25 (ablation) | DVCE with Swin-TF · qualitative | closeness vs realism | No clear effect |
| adaptive parameterization (normalized gradients) vs fixed coefficients as in Blended Diffusion (ablation) | DVCE · qualitative | hyperparameter transfer across images | Helped |
Closest papers by meaning
- DiG-IN: Diffusion Guidance for Investigating Networks - Uncovering Classifier Differences Neuron Visualisations and Visual Counterfactual Explanations (CVPR 2024 · similarity 0.91)
- Rethinking Visual Counterfactual Explanations Through Region Constraint (ICLR 2025 · similarity 0.90)
- V-CECE: Visual Counterfactual Explanations via Conceptual Edits (NeurIPS 2025 · similarity 0.88)
- Explaining in Diffusion: Explaining a Classifier with Diffusion Semantics (CVPR 2025 · similarity 0.86)
- Towards Fine-Grained Interpretability: Counterfactual Explanations for Misclassification with Saliency Partition (CVPR 2025 · similarity 0.86)
- Counterfactual Explanations for Misclassified Images: How Human and Machine Explanations Differ (Abstract Reprint) (AAAI 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.