Research map / Adversarial robustness and security / Black-box and transfer attacks
Heuristic Black-Box Adversarial Attacks on Video Recognition Models
Zhipeng Wei, Jingjing Chen, Xingxing Wei, Linxi Jiang, Tat-Seng Chua, Fengfeng Zhou, Yu-Gang Jiang
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 spatial sparsity to temporal sparsity (ablation) | Heuristic video attack · UCF-101, C3D, untargeted | MQ: 16292.0 -> 12940.0 | Improved |
| add spatial sparsity to temporal sparsity (ablation) | Heuristic video attack · UCF-101, LRCN, untargeted | MQ: 14713.5 -> 8421.5 | Improved |
| add spatial sparsity to temporal sparsity (ablation) | Heuristic video attack · HMDB-51, C3D, untargeted | MQ: 13536.5 -> 10616.0 | Improved |
| add spatial sparsity to temporal sparsity (ablation) | Heuristic video attack · HMDB-51, LRCN, untargeted | MQ: 15369.5 -> 13311.5 | Improved |
| add spatial sparsity to temporal sparsity (ablation) | Heuristic video attack · UCF-101, C3D, targeted | MQ: 313229.0 -> 167217.0 | Improved |
| add spatial sparsity to temporal sparsity (ablation) | Heuristic video attack · UCF-101, LRCN, targeted | MQ: 566719.0 -> 399655.0 | Improved |
| add spatial sparsity to temporal sparsity (ablation) | Heuristic video attack · HMDB-51, C3D, targeted | MQ: 196371.5 -> 144917.5 | Improved |
| add spatial sparsity to temporal sparsity (ablation) | Heuristic video attack · HMDB-51, LRCN, targeted | MQ: 339367.0 -> 206120.0 | Improved |
| use temporal sparsity without spatial sparsity instead of Opt-attack (ablation) | Heuristic video attack · UCF-101, C3D, targeted | MQ: 207944.5 -> 313229.0 | Hurt |
| use temporal sparsity without spatial sparsity instead of Opt-attack (ablation) | Heuristic video attack · UCF-101, LRCN, targeted | MQ: 445279.0 -> 566719.0 | Hurt |
| use temporal sparsity without spatial sparsity instead of Opt-attack (ablation) | Heuristic video attack · HMDB-51, LRCN, targeted | MQ: 224414.0 -> 339367.0 | Hurt |
| decrease salient-region area ratio to 0.2 from 1.0 (ablation) | Heuristic video attack · Untargeted attack, parameter-tuning videos | FR: 100 -> 90 | Hurt |
Closest papers by meaning
- Sparse Black-Box Video Attack with Reinforcement Learning (IJCV 2022 · similarity 0.90)
- Adversarial Attacks on Black Box Video Classifiers: Leveraging the Power of Geometric Transformations (NeurIPS 2021 · similarity 0.89)
- Blackbox Attacks via Surrogate Ensemble Search (NeurIPS 2022 · similarity 0.88)
- Simple Black-box Adversarial Attacks (ICML 2019 · similarity 0.88)
- Efficient Black-box Adversarial Attacks via Bayesian Optimization Guided by a Function Prior (ICML 2024 · similarity 0.86)
- Understanding the Robustness of Randomized Feature Defense Against Query-Based Adversarial Attacks (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.
