Research map / Label-efficient and robust learning / Semi-supervised learning
On the Calibration of Image Semi-Supervised Learning Models
Mehrab Mustafy Rahman, Jayanth Mohan, Tiberiu Sosea, Cornelia Caragea
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 |
|---|---|---|---|
| add CalibrateMix targeted mixup to FixMatch | FixMatch · CIFAR-100 (25 labels/class) | ECE (%): 21.19 -> 12.68 | Improved |
| add CalibrateMix to SoftMatch | SoftMatch · CIFAR-100 (25 labels/class) | top-1 error (%): 26.76 -> 26.86 | No clear effect |
| add CalibrateMix to FlexMatch | FlexMatch · CIFAR-100 (25 labels/class) | ECE (%): 29.54 -> 28.23 | Mixed |
| add CalibrateMix to SoftMatch | SoftMatch · STL-10 (4 labels/class) | top-1 error (%): 20.90 -> 15.18 | Improved |
| add CalibrateMix to FixMatch at 4 labels/class | FixMatch · CIFAR-10 (4 labels/class) | ECE (%): 5.42 -> 3.24 | Improved |
| add CalibrateMix to fully supervised training | FullySupervised · CIFAR-100 | ECE (%): 6.77 -> 4.74 | Improved |
| use targeted CalibrateMix mixup instead of random mixup (ablation) | FixMatch · CIFAR-100 (4 labels/class) | ECE (%): 38.05 -> 35.43 | Improved |
| add CalibrateMix on top of label smoothing (ablation) | FixMatch + LS · CIFAR-100 (25 labels/class) | ECE (%): 17.37 -> 12.89 | Improved |
| add CalibrateMix on top of label smoothing (ablation) | SoftMatch + LS · CIFAR-10/CIFAR-100/STL-10 (error rate) | top-1 error (%): CIFAR-100 25 labels: 26.50 -> 27.41 | Mixed |
| restrict mixup to easy/hard labeled-unlabeled pairs instead of mixing all pairs (ablation) | SoftMatch + CalibrateMix · CIFAR-100 (4 labels/class) | ECE (%): 30.63 -> 23.1 | Improved |
| add warmup (ablation) | mixup-all SoftMatch · CIFAR-100 (4 labels/class) | top-1 error (%): 86.95 -> 41.5 | Improved |
| add cosine-dissimilarity sample selection (k=5%) (ablation) | SoftMatch + CalibrateMix · STL-10 (4 labels/class) | ECE (%): 16.81 -> 12.4 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- SoftMatch — SoftMatch: Addressing the Quantity-Quality Tradeoff in Semi-supervised Learning (ICLR 2023) · better on 4 of 5 shared results
- FixMatch — FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence (NeurIPS 2020) · better on 5 of 5 shared results
- FlexMatch — FlexMatch: Boosting Semi-Supervised Learning with Curriculum Pseudo Labeling (NeurIPS 2021) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Beyond One-Hot Labels: Semantic Mixing for Model Calibration (ICML 2025 · similarity 0.88)
- RegMixMatch: Optimizing Mixup Utilization in Semi-Supervised Learning (AAAI 2025 · similarity 0.88)
- Rethinking Confidence Scores and Thresholds in Pseudolabeling-based SSL (ICML 2025 · similarity 0.86)
- Reassessing How to Compare and Improve the Calibration of Machine Learning Models (ICLR 2025 · similarity 0.86)
- Divide et Calibra: Multiclass Local Calibration via Vector Quantization (NeurIPS 2026 · similarity 0.86)
- CGMatch: A Different Perspective of Semi-supervised Learning (CVPR 2025 · 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.
