Research map / Label-efficient and robust learning / Recognition and long-tail
Adversarial Label Learning
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 |
|---|---|---|---|
| Train against adversarial labels constrained by weak-signal error bounds | ALL · Fashion-MNIST (DvK) | accuracy: 0.975 -> 0.998 | Helped |
| Train against adversarial labels constrained by weak-signal error bounds | ALL · Fashion-MNIST (SvA) | accuracy: 0.719 -> 0.924 | Helped |
| Train against adversarial labels constrained by weak-signal error bounds | ALL · Breast Cancer | accuracy: 0.936 -> 0.945 | Helped |
| Train against adversarial labels constrained by weak-signal error bounds | ALL · Cardiotocography | accuracy: 0.947 -> 0.803 | Hurt |
| Train against adversarial labels constrained by weak-signal error bounds | ALL · Statlog Satellite | accuracy: 0.992 -> 0.936 | Hurt |
| Train against adversarial labels constrained by weak-signal error bounds | ALL · Wine Quality | accuracy: 0.596 -> 0.662 | Helped |
| Use fixed estimated error bounds of 0.3 instead of training-set error bounds | ALL · Fashion-MNIST (SvA) | accuracy: 0.924 -> 0.901 | No clear effect |
| Use fixed estimated error bounds of 0.3 instead of training-set error bounds | ALL · Wine Quality | accuracy: 0.623 -> 0.694 | Helped |
| Add a second weak signal to ALL (ablation) | ALL-1 · Clave Direction | accuracy: 0.646 -> 0.837 | Helped |
| Add a third weak signal to ALL (ablation) | ALL-2 · Cardiotocography | accuracy: 0.803 -> 0.803 | No clear effect |
| Add a third weak signal to ALL (ablation) | ALL-2 · Credit Card | accuracy: 0.696 -> 0.697 | No clear effect |
| Add a second weak signal to ALL (ablation) | ALL-1 · Statlog Satellite | accuracy: 0.470 -> 0.933 | Helped |
Closest papers by meaning
- Adversarial Multi Class Learning under Weak Supervision with Performance Guarantees (ICML 2021 · similarity 0.92)
- Learning from weak labelers as constraints (ICLR 2025 · similarity 0.89)
- Soften to Defend: Towards Adversarial Robustness via Self-Guided Label Refinement (CVPR 2024 · similarity 0.87)
- Fusing Conditional Submodular GAN and Programmatic Weak Supervision (AAAI 2024 · similarity 0.87)
- Weakly Supervised Contrastive Adversarial Training for Learning Robust Features from Semi-supervised Data (CVPR 2025 · similarity 0.86)
- Semantic-Aware Multi-Label Adversarial Attacks (CVPR 2024 · 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.