Research map / Label-efficient and robust learning / Semi-supervised learning
Decouple Then Converge: Handling Unknown Unlabeled Distributions in Long-Tailed Semi-Supervised Learning
Kai Gan, Tong Wei, Min-Ling Zhang
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 alignment of the standard branch to the balanced-branch distribution (ablation) | DTC · CIFAR-10-LT reversed | accuracy: 66.8 -> 86.6 | Improved |
| add balanced pseudo-labels (logit-adjusted standard-branch outputs) (ablation) | DTC · CIFAR-10-LT consistent | accuracy: 82.2 -> 84.1 | Improved |
| add balanced softmax in the balanced branch (ablation) | DTC · CIFAR-100-LT consistent | accuracy: 51.1 -> 52.0 | Improved |
| add importance weighting of pseudo-labels (ablation) | DTC · CIFAR-100-LT uniform | accuracy: 58.3 -> 60.4 | Improved |
| add post-hoc logit adjustment (tau3=2) (ablation) | DTC · CIFAR-10-LT consistent | accuracy: 77.8 -> 84.1 | Improved |
| add post-hoc adjustment (tau3=2) to ACR (ablation) | ACR · CIFAR-10-LT consistent | accuracy: 81.6 -> 82.0 | Improved |
| add cross-branch pseudo-labeling (CBPL) to ACR, before post-hoc (ablation) | ACR · CIFAR-10-LT consistent | accuracy: 81.6 -> 78.1 | Hurt |
| add CBPL to ACR, before post-hoc (ablation) | ACR · CIFAR-10-LT uniform | accuracy: 92.1 -> 92.8 | Improved |
| add decoupled distribution alignment on top of CBPL (ablation) | DTC · CIFAR-10-LT uniform | accuracy: 92.8 -> 93.5 | Improved |
| add importance weighting on top of CBPL and DDA (ablation) | DTC · CIFAR-10-LT uniform | accuracy: 93.5 -> 93.8 | Improved |
| use DTC instead of ACR on random unlabeled distributions | FixMatch · CIFAR-100-LT random distributions | average accuracy: 43.6 -> 45.7 | Improved |
| add DTC to FixMatch (dual branch) | FixMatch · CIFAR-100-LT | accuracy and training time: 49.7 -> 57.7 | 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.
- SimPro — SimPro: A Simple Probabilistic Framework Towards Realistic Long-Tailed Semi-Supervised Learning (ICML 2024) · better on 4 of 4 shared results
- FixMatch — FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence (NeurIPS 2020) · better on 3 of 3 shared results
- DASO — DASO: Distribution-Aware Semantics-Oriented Pseudo-Label 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
- Continuous Contrastive Learning for Long-Tailed Semi-Supervised Recognition (NeurIPS 2024 · similarity 0.90)
- A Square Peg in a Square Hole: Meta-Expert for Long-Tailed Semi-Supervised Learning (ICML 2025 · similarity 0.90)
- Keep It on a Leash: Controllable Pseudo-label Generation Towards Realistic Long-Tailed Semi-Supervised Learning (NeurIPS 2025 · similarity 0.90)
- Three Heads Are Better than One: Complementary Experts for Long-Tailed Semi-supervised Learning (AAAI 2024 · similarity 0.89)
- BEM: Balanced and Entropy-based Mix for Long-Tailed Semi-Supervised Learning (CVPR 2024 · similarity 0.89)
- SimPro: A Simple Probabilistic Framework Towards Realistic Long-Tailed Semi-Supervised Learning (ICML 2024 · similarity 0.89)
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.
