Research map / Adversarial robustness and security / Certified and empirical robustness
Robust One-Class Classification with Signed Distance Function using 1-Lipschitz Neural Networks
Louis Béthune, Paul Novello, Guillaume Coiffier, Thibaut Boissin, Mathieu Serrurier, Quentin VINCENOT, Andres Troya-Galvis
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 |
|---|---|---|---|
| Use OCSDF instead of Deep SVDD | MNIST one-versus-all · MNIST one-versus-all | mAUROC: 95.5\pm 0.4 vs 94.8\pm 0.9 | No clear effect |
| Use OCSDF instead of the strongest reported conventional baseline | CIFAR10 one-versus-all · CIFAR10 one-versus-all | mAUROC: 57.4\pm 2.1 vs 64.8\pm 8.0 | Hurt |
| Use OCSDF for Cats-versus-Dogs detection | OCSDF · Cats versus Dogs (Cats normal) | AUROC: 55.0\% | Hurt |
| Increase the certified l2 attack radius | OCSDF on MNIST · MNIST one-versus-all | certified mAUROC: 95.5\pm 0.4 -> 57.5\pm 7.5 | Hurt |
| Increase the certified l2 attack radius | OCSDF on CIFAR10 · CIFAR10 one-versus-all | certified mAUROC: 57.4\pm 2.1 -> 22.5\pm 1.4 | Hurt |
| Tune the margin separately for each CIFAR-10 class (ablation) | OCSDF on CIFAR10 · CIFAR10 one-versus-all | AUROC | Improved |
Closest papers by meaning
- Pay attention to your loss : understanding misconceptions about Lipschitz neural networks (NeurIPS 2022 · similarity 0.87)
- Robust-by-Design Classification via Unitary-Gradient Neural Networks (AAAI 2023 · similarity 0.86)
- Certified Robustness Under Bounded Levenshtein Distance (ICLR 2025 · similarity 0.85)
- A Recipe for Improved Certifiable Robustness (ICLR 2024 · similarity 0.84)
- Efficient Robust Conformal Prediction via Lipschitz-Bounded Networks (ICML 2025 · similarity 0.84)
- Enhancing Certified Robustness via Block Reflector Orthogonal Layers and Logit Annealing Loss (ICML 2025 · similarity 0.83)
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.
