Research map / Adversarial robustness and security / Black-box and transfer attacks
AdvFM: Lookahead Flow-Matching Velocity-Field Attacks for Imperceptible and Transferable Adversarial Examples
Runze Liu, Zeyue Wang, Fanghui Sun, Rui Liu, Yihan Yan, Shen Wang, Zhaoyang Zhang
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 lookahead two-point loss (ablation) | AdvFM · ImageNet black-box avg (ResNet50) | ASR (%): 64.31 -> 72.05 | Improved |
| add lookahead two-point loss (ablation) | AdvFM · ImageNet white-box (ResNet50) | ASR (%): 91.72 -> 94.45 | Improved |
| attack flow-matching velocity field instead of diffusion denoise/renoise | DiffAttack/APA · ImageNet VGG19 source | average black-box ASR (%): 65.86 (APA) -> 70.35 | Improved |
| attack flow-matching velocity field instead of diffusion | APA · ImageNet ViT-B/16 source | average black-box ASR (%): 69.73 -> 68.22 | Hurt |
| use AdvFM against purification NRP | APA · NRP defense | ASR (%): 81.20 -> 94.98 | Improved |
| use AdvFM against DiffPure | APA · DiffPure defense | ASR (%): 63.30 -> 61.33 | No clear effect |
| use AdvFM against adversarially trained ViTs | AdvAD · RAT ViT-B/16 | ASR (%): 97.70 -> 98.53 | Improved |
| use AdvFM on adversarially trained ResNets | DiffAttack · Madry RN50 | ASR (%): 83.84 -> 81.89 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- DiffAttack — DiffAttack: Evasion Attacks Against Diffusion-Based Adversarial Purification (NeurIPS 2023) · better on 5 of 6 shared results
- AdvAD — AdvAD: Exploring Non-Parametric Diffusion for Imperceptible Adversarial Attacks (NeurIPS 2024) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Adversarial Flow Models (ICML 2026 · similarity 0.89)
- Continuous Adversarial Flow Models (ECCV 2026 · similarity 0.88)
- A Perturbation-Constrained Adversarial Attack for Evaluating the Robustness of Optical Flow (ECCV 2022 · similarity 0.87)
- AdvAD: Exploring Non-Parametric Diffusion for Imperceptible Adversarial Attacks (NeurIPS 2024 · similarity 0.87)
- MimicDiffusion: Purifying Adversarial Perturbation via Mimicking Clean Diffusion Model (CVPR 2024 · similarity 0.86)
- Self-Corrected Flow Distillation for Consistent One-Step and Few-Step Image Generation (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.
