Research map / Adversarial robustness and security / Certified and empirical robustness
Adversarially Pretrained Transformers may be Universally Robust In-Context Learners
Soichiro Kumano, Hiroshi Kera, Toshihiko Yamasaki
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 |
|---|---|---|---|
| adversarially pretrain the transformer instead of standard pretraining | single-layer linear transformer · MNIST | robust accuracy (%): 4 -> 72 | Improved |
| adversarially pretrain the transformer instead of standard pretraining | single-layer linear transformer · Fashion-MNIST | robust accuracy (%): 20 -> 62 | Improved |
| adversarially pretrain the transformer instead of standard pretraining | single-layer linear transformer · CIFAR-10 | robust accuracy (%): 21 -> 34 | Improved |
| adversarially pretrain the transformer instead of standard pretraining | single-layer linear transformer · test distribution D_te | robust accuracy (%): 0 -> 95 | Improved |
| adversarially pretrain the transformer instead of standard pretraining | single-layer linear transformer · MNIST | clean accuracy (%): 94 -> 93 | Hurt |
| adversarially pretrain the transformer instead of standard pretraining | single-layer linear transformer · CIFAR-10 | clean accuracy (%): 68 -> 64 | Hurt |
| adversarially pretrain the transformer instead of standard pretraining | single-layer linear transformer · Fashion-MNIST | clean accuracy (%): 91 -> 89 | Hurt |
Closest papers by meaning
- On Transfer of Adversarial Robustness from Pretraining to Downstream Tasks (NeurIPS 2023 · similarity 0.90)
- Adversarial Robustness: From Self-Supervised Pre-Training to Fine-Tuning (CVPR 2020 · similarity 0.90)
- Robustness Tokens: Towards Adversarial Robustness of Transformers (ECCV 2024 · similarity 0.90)
- Adversarial Training Can Provably Improve Robustness: Theoretical Analysis of Feature Learning Process Under Structured Data (ICLR 2025 · similarity 0.89)
- Initialization Matters for Adversarial Transfer Learning (CVPR 2024 · similarity 0.89)
- Learning Better Certified Models from Empirically-Robust Teachers (NeurIPS 2026 · similarity 0.88)
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.
