Research map / Adversarial robustness and security / Attacks on text and graphs
Synthesizing Robust Adversarial Examples
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 |
|---|---|---|---|
| Optimize the input over a distribution of transformations instead of optimizing a single view | EOT adversarial images · ImageNet validation, 1,000 images with random transformations | adversariality: 96.4% | Helped |
| Optimize object textures over a distribution of 3D poses | EOT adversarial objects · 10 3D models and 20 random target classes per model | adversariality: 83.4% | Helped |
| Model printing error, lighting, and camera noise in the transformation distribution | Physical-world 3D adversarial objects · Physical-world turtle and baseball | adversarial target classification: Turtle: 82%; Baseball: 59% | Helped |
| Increase the transformation distribution to cover more kinds of variation | EOT adversarial examples · 2D and 3D transformations | required perturbation | Hurt |
| Use EOT when the transformation distribution is too large or the perturbation budget too small | EOT adversarial examples · Transformation distributions and perturbation constraints | adversariality | Hurt |
Closest papers by meaning
- Connecting the Digital and Physical World: Improving the Robustness of Adversarial Attacks (AAAI 2019 · similarity 0.91)
- Isometric 3D Adversarial Examples in the Physical World (NeurIPS 2022 · similarity 0.89)
- Image Synthesis with a Single (Robust) Classifier (NeurIPS 2019 · similarity 0.89)
- Robust Physical-World Attacks on Deep Learning Visual Classification (CVPR 2018 · similarity 0.88)
- Unadversarial Examples: Designing Objects for Robust Vision (NeurIPS 2021 · similarity 0.88)
- Adversarial Training Can Provably Improve Robustness: Theoretical Analysis of Feature Learning Process Under Structured Data (ICLR 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.