Research map / Label-efficient and robust learning / Semi-supervised learning
Safe Deep Semi-Supervised Learning for Unseen-Class Unlabeled Data
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 |
|---|---|---|---|
| DS3L safe weighting of unlabeled data (vs. standard deep SSL using all unlabeled data) | deep SSL on MNIST (classes 1-6) · MNIST class mismatch sweep | classification accuracy | Helped |
| DS3L safe weighting of unlabeled data | deep SSL on CIFAR-10 (6 animal classes) · CIFAR-10 class mismatch sweep | classification accuracy | Helped |
| Adding DS3L to Pi-Model, Temporal Ensembling, Mean Teacher, VAT | four deep SSL methods on MNIST and CIFAR-10 · MNIST and CIFAR-10 class mismatch sweeps | accuracy relative to supervised baseline | Helped |
| Replacing probability-estimation with DS3L weighting for unseen-class identification | unseen-class unlabeled data identification · MNIST and CIFAR-10 (1-AUC)% | (1-AUC)%: lower (1-AUC) at every ratio, e.g. 4.33 to 1.67 at ratio 0.1 on MNIST | Helped |
| Existing deep SSL methods (Pseudo-Label, Pi-Model, Temporal Ensembling, Mean Teacher, VAT) without safe weighting | deep SSL under class mismatch · MNIST and CIFAR-10 class mismatch sweep | accuracy relative to supervised baseline | Hurt |
Closest papers by meaning
- Safe-Student for Safe Deep Semi-Supervised Learning With Unseen-Class Unlabeled Data (CVPR 2022 · similarity 0.94)
- Not All Parameters Should Be Treated Equally: Deep Safe Semi-supervised Learning under Class Distribution Mismatch (AAAI 2022 · similarity 0.94)
- Realistic Evaluation of Deep Semi-Supervised Learning Algorithms (NeurIPS 2018 · similarity 0.92)
- Re-Evaluating the Impact of Unseen-Class Unlabeled Data on Semi-Supervised Learning Model (ICLR 2025 · similarity 0.89)
- Contrastive Credibility Propagation for Reliable Semi-supervised Learning (AAAI 2024 · similarity 0.86)
- Unlocking the Potential of Unlabeled Data in Semi-Supervised Domain Generalization (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.