Research map / Label-efficient and robust learning / Semi-supervised learning
SCAD: Super-Class-Aware Debiasing for Long-Tailed Semi-Supervised Learning
Sunguk Jang, Jinwoo Jeon, Byung-Jun Lee
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 super-class learning (auxiliary super-class classifier) (ablation) | FixMatch · CIFAR10-LT (consistent) | accuracy: 67.8 -> 69.2 | Improved |
| add super-class learning (auxiliary super-class classifier) (ablation) | FixMatch · STL10-LT (consistent) | accuracy: 56.1 -> 69.0 | Improved |
| add super-class-aware debiasing on top of LA (ablation) | FixMatch + super-class learning + LA · CIFAR10-LT (consistent) | accuracy: 76.9 -> 78.7 | Improved |
| add super-class-aware debiasing on top of LA (ablation) | FixMatch + super-class learning + LA · STL10-LT (consistent) | accuracy: 70.4 -> 71.3 | Improved |
| add SCAD to ACR | FixMatch + ACR · ImageNet-127 (32x32) | accuracy: 57.2 -> 60.1 | Improved |
| add SCAD to ACR | FixMatch + ACR · ImageNet-127 (64x64) | accuracy: 63.6 -> 66.7 | Improved |
| add SCAD to ACR | FixMatch + ACR · CIFAR100-LT (γ=10, N1=50) | accuracy: 51.3 -> 52.7 | Improved |
| add SCAD to CDMAD | FixMatch + CDMAD · CIFAR10-LT (reverse γu=1/100) | accuracy: 77.6 -> 79.4 | Improved |
| add SCAD to CoSSL | FixMatch + CoSSL · CIFAR10-LT (reverse γu=1/100) | accuracy: 85.1 -> 84.8 | No clear effect |
| add SCAD to SAW | FixMatch + SAW · CIFAR10-LT (uniform γu=1) | accuracy: 81.2 -> 81.1 | No clear effect |
| use CLIP text encoder instead of GloVe for super-classes (ablation) | FixMatch + LA + SCAD · CIFAR100-LT | accuracy: 49.7 -> 50.4 | Mixed |
| use predicted text-based super-classes instead of ground truth (ablation) | FixMatch + LA + SCAD · CIFAR100-LT | accuracy: 50.3 -> 50.5 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- CoSSL — CoSSL: Co-Learning of Representation and Classifier for Imbalanced Semi-Supervised Learning (CVPR 2022) · better on 2 of 2 shared results
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
- CDMAD: Class-Distribution-Mismatch-Aware Debiasing for Class-Imbalanced Semi-Supervised Learning (CVPR 2024 · similarity 0.88)
- Language-Assisted Debiasing and Smoothing for Foundation Model-Based Semi-Supervised Learning (CVPR 2025 · similarity 0.87)
- Three Heads Are Better than One: Complementary Experts for Long-Tailed Semi-supervised Learning (AAAI 2024 · similarity 0.87)
- LCGC: Learning from Consistency Gradient Conflicting for Class-Imbalanced Semi-Supervised Debiasing (AAAI 2025 · similarity 0.87)
- Twice Class Bias Correction for Imbalanced Semi-supervised Learning (AAAI 2024 · similarity 0.86)
- SSE-SAM: Balancing Head and Tail Classes Gradually Through Stage-Wise SAM (AAAI 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.
