Research map / Adversarial robustness and security / Certified and empirical robustness
DANCE: Enhancing saliency maps using decoys
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 |
|---|---|---|---|
| Add DANCE decoys and range aggregation to the saliency method | Vanilla gradient, integrated gradient, and SmoothGrad · ImageNet | fidelity | Helped |
| Add DANCE decoys and range aggregation to the saliency method | Vanilla gradient, integrated gradient, and SmoothGrad · SST | fidelity | Helped |
| Add DANCE decoys and range aggregation to the saliency method | Saliency methods · ImageNet adversarial attacks: top-k, target, and mass-center | sensitivity | Helped |
| Replace optimized decoy generation with constant perturbation (ablation) | DANCE · ImageNet | fidelity | Hurt |
| Replace optimized decoy generation with Gaussian noise perturbation (ablation) | DANCE · ImageNet | fidelity | Hurt |
| Replace range aggregation with mean aggregation (ablation) | DANCE · ImageNet | fidelity | Hurt |
| Replace optimized decoy generation with constant perturbation (ablation) | DANCE · SST | fidelity | Hurt |
| Replace optimized decoy generation with noise perturbation (ablation) | DANCE · SST | fidelity | Hurt |
| Replace range aggregation with mean aggregation (ablation) | DANCE · SST | fidelity | Hurt |
| Vary the number of decoys (ablation) | DANCE · ImageNet | fidelity | No clear effect |
| Vary patch size, target network layer, or initial Lagrange multiplier (ablation) | DANCE · ImageNet | fidelity | No clear effect |
Closest papers by meaning
- CAMERAS: Enhanced Resolution and Sanity Preserving Class Activation Mapping for Image Saliency (CVPR 2021 · similarity 0.88)
- There and Back Again: Revisiting Backpropagation Saliency Methods (CVPR 2020 · similarity 0.88)
- Sanity Checks for Saliency Maps (NeurIPS 2018 · similarity 0.87)
- A Novel Visual Interpretability for Deep Neural Networks by Optimizing Activation Maps with Perturbation (AAAI 2021 · similarity 0.87)
- Where do Models go Wrong? Parameter-Space Saliency Maps for Explainability (NeurIPS 2022 · similarity 0.87)
- Great Minds Think Alike: The Universal Convergence Trend of Input Salience (NeurIPS 2024 · similarity 0.85)
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.