Research map / Label-efficient and robust learning / Semi-supervised learning
Divide-and-Conquer for Enhancing Unlabeled Learning, Stability, and Plasticity in Semi-supervised Continual Learning
Yue Duan, Taicai Chen, Lei Qi, Yinghuan Shi
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 USP in place of DSGD | iCaRL&Fix · CIFAR10-30 | average accuracy: 77.33 -> 79.66 | Improved |
| add USP in place of DSGD | DER&Fix · CIFAR100-125 | average accuracy: 69.14 -> 71.60 | Improved |
| add USP in place of DSGD | DER&Fix · ImageNet100-13 | average accuracy: 35.73 -> 46.09 | Improved |
| add USP in place of DSGD | DER&Fix · ImageNet100-100 | average accuracy: 62.27 -> 62.29 | No clear effect |
| add USP in place of DSGD | iCaRL&Fix · CIFAR10-150 | last accuracy: 79.69 -> 78.21 | Hurt |
| add USP | UaD-CIE · CUB few-shot | average accuracy: 64.33 -> 66.43 | Improved |
| add USP | UaD-CIE · CUB novel classes | average accuracy: 44.42 -> 55.41 | Improved |
| add USP | UaD-CIE · CUB base classes | average accuracy: 72.03 -> 71.07 | Hurt |
| add low-confidence samples via DCP (ablation) | USP · CIFAR10-30 | average accuracy: 68.34 -> 81.43 | Improved |
| use CUD instead of logit distillation (ablation) | USP · CIFAR100-25 | average accuracy: 53.91 -> 54.36 | Improved |
| use CUD instead of feature distillation (ablation) | USP · CIFAR100-25 | average accuracy: 48.16 -> 54.36 | Improved |
| use equal FSR weights for unlabeled data (1.0 vs 0.5) (ablation) | USP · CIFAR10-30 | average accuracy: 79.52 -> 81.63 | Improved |
Closest papers by meaning
- Dynamic Sub-graph Distillation for Robust Semi-supervised Continual Learning (AAAI 2024 · similarity 0.91)
- Persistence Homology Distillation for Semi-supervised Continual Learning (NeurIPS 2024 · similarity 0.90)
- Continual Learning on a Diet: Learning from Sparsely Labeled Streams Under Constrained Computation (ICLR 2024 · similarity 0.89)
- CGMatch: A Different Perspective of Semi-supervised Learning (CVPR 2025 · similarity 0.87)
- InterLUDE: Interactions between Labeled and Unlabeled Data to Enhance Semi-Supervised Learning (ICML 2024 · similarity 0.87)
- Towards Macro-AUC Oriented Imbalanced Multi-Label Continual Learning (AAAI 2025 · similarity 0.87)
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.
