On Adversarial Training without Perturbing all Examples

M. Losch, Mohamed Omran, David Stutz, Mario Fritz, B. Schiele
2/12/2026

No official code found. Best implementation: mlosch/SAT (4 stars, MIT, last commit 2024-07-25).

Semantic Scholar

Code Implementations(1)

mlosch/SATCommunity64%
MIT

Official repository for the paper: "On Adversarial Training without Perturbing all Examples", Accepted at ICLR 2024

40Shell, PythonAug 8, 20232 years agoMIT
adversarial-robustnessadversarial-trainingrobustness-analysis

Cite this paper

@article{losch2026adversarial,
  title  = {On Adversarial Training without Perturbing all Examples},
  author = {M. Losch and Mohamed Omran and David Stutz and Mario Fritz and B. Schiele},
  year   = {2026},
  url    = {https://api.semanticscholar.org/CorpusID:879aa6c89906adda363c06c7f9590320061685d5},
  journal = {ICLR 2024 2024}
}

Discussion