Research map / Label-efficient and robust learning / Recognition and long-tail
When and How Mixup Improves Calibration
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 |
|---|---|---|---|
| Add Mixup training | Fisher linear classifier · Gaussian classification, high-dimensional regime | ECE | Helped |
| Add Mixup training | Fisher linear classifier · Gaussian classification, low-dimensional regime | ECE | Hurt |
| Increase Mixup interpolation from zero | Fisher linear classifier · Gaussian classification, high-dimensional regime | ECE | Helped |
| Add Mixup training | Fisher linear classifier · Out-of-domain Gaussian data within the stated shift condition | ECE | Helped |
| Add Mixup training | Classifier with an estimated representation · Gaussian generative model, high-dimensional regime | ECE | Helped |
| Add pseudo-labels from unlabeled data (ablation) | Initial classifier · Semi-supervised learning, insufficient labeled data | ECE | Helped |
| Add pseudo-labels from unlabeled data (ablation) | Initial classifier · Semi-supervised learning, sufficient labeled data | ECE | Hurt |
| Add Mixup to the final pseudo-label training step | Pseudo-labeling algorithm · Semi-supervised Gaussian classification | ECE | Helped |
| Add Mixup augmentation | ResNet · CIFAR-10 | ECE | Mixed |
| Add Mixup augmentation | ResNet · CIFAR-100 | ECE | Mixed |
| Add Mixup augmentation | Classifier · Gaussian classification, high-dimensional regime | MCE | Helped |
| Add Mixup augmentation | Classifier · Gaussian classification, low-dimensional regime | MCE | Hurt |
Closest papers by meaning
- Tailoring Mixup to Data for Calibration (ICLR 2025 · similarity 0.90)
- Beyond One-Hot Labels: Semantic Mixing for Model Calibration (ICML 2025 · similarity 0.84)
- Reassessing How to Compare and Improve the Calibration of Machine Learning Models (ICLR 2025 · similarity 0.83)
- On the Limitations of Temperature Scaling for Distributions with Overlaps (ICLR 2024 · similarity 0.82)
- Selective Mixup Helps with Distribution Shifts, But Not (Only) because of Mixup (ICML 2024 · similarity 0.82)
- Who’s the (Multi-)Fairest of Them All: Rethinking Interpolation-Based Data Augmentation Through the Lens of Multicalibration (AAAI 2025 · similarity 0.82)
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.