Research map / Adversarial robustness and security / Attacks on text and graphs
Connecting the Dots: Detecting Adversarial Perturbations Using Context Inconsistency
Shasha Li, Shitong Zhu, Sudipta Paul, Amit Roy-Chowdhury, Chengyu Song, Srikanth Krishnamurthy, Ananthram Swami, Kevin S Chan
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 edge (GRU gate) features to the context profile (ablation) | SCEME · PASCAL VOC digital miscategorization | ROC-AUC: 0.866 -> 0.938 | Improved |
| add edge features to the context profile (ablation) | SCEME · PASCAL VOC digital appearing | ROC-AUC: 0.828 -> 0.869 | Improved |
| add edge features to the context profile (ablation) | SCEME · MS COCO digital miscategorization | ROC-AUC: 0.901 -> 0.959 | Improved |
| add edge features to the context profile (ablation) | SCEME · MS COCO digital hiding | ROC-AUC: 0.976 -> 0.984 | Improved |
| add edge features to the context profile (ablation) | SCEME · MS COCO physical hiding | ROC-AUC: 0.954 -> 0.968 | Improved |
| add edge features to the context profile (ablation) | SCEME · PASCAL VOC physical appearing | ROC-AUC: 0.927 -> 0.970 | Improved |
| add edge features, stop sign recall at 0.1% FPR (ablation) | SCEME · COCO stop sign (1000 perturbed) | recall: 33% -> 54% | Improved |
| add edge features, stop sign recall at 10% FPR (ablation) | SCEME · COCO stop sign (1000 perturbed) | recall: 91% -> 93% | Improved |
| use SCEME context profile instead of FeatureSqueeze | detector · COCO stop sign (1000 perturbed) | recall at 5% FPR: 3% -> 89% | Improved |
| use SCEME instead of FeatureSqueeze | detector · PASCAL VOC digital hiding | ROC-AUC: 0.620 -> 0.981 | Improved |
| use SCEME instead of Co-occurGraph | detector · MS COCO physical miscategorization | ROC-AUC: 0.546 -> 0.989 | Improved |
| more region proposals on the perturbed object (1-3 to 9-12) | SCEME · PASCAL VOC digital miscategorization | ROC-AUC: 0.830 -> 0.950 | Improved |
Closest papers by meaning
- Zero-Query Transfer Attacks on Context-Aware Object Detectors (CVPR 2022 · similarity 0.91)
- Exploiting Multi-Object Relationships for Detecting Adversarial Attacks in Complex Scenes (ICCV 2021 · similarity 0.90)
- On Detecting Adversarial Perturbations (ICLR 2017 · similarity 0.89)
- Theoretical Understanding of Learning from Adversarial Perturbations (ICLR 2024 · similarity 0.87)
- Wide Two-Layer Networks can Learn from Adversarial Perturbations (NeurIPS 2024 · similarity 0.86)
- Contrasting Adversarial Perturbations: The Space of Harmless Perturbations (AAAI 2025 · 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.
