Research map / Label-efficient and robust learning / Semi-supervised learning
Towards Safe Semi-Supervised Learning for Multivariate Performance Measures
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 |
|---|---|---|---|
| add maximin worst-case weighting of S3VM learners (UMVP vs uniform-weight UMVP-) (ablation) | UMVP · 10 datasets | avg AUC improvement over SVMperf: 0.3 -> 0.8 | Helped |
| add maximin worst-case weighting (UMVP vs UMVP-) (ablation) | UMVP · 10 datasets | win/tie/loss vs SVMperf on AUC: 7/0/3 -> 8/2/0 | Helped |
| add maximin weighting (UMVP vs UMVP-) (ablation) | UMVP · 10 datasets | avg F1 improvement: 6.0 -> 5.9 | No clear effect |
| integrate multiple SSL learners with maximin safe combination vs self-training | UMVP · 10 datasets | avg Pre@k improvement over SVMperf: -2.5 -> 1.8 | Helped |
| UMVP vs S4VM | UMVP · 10 datasets | avg AUC improvement over SVMperf: -0.5 -> 0.8 | Helped |
| use UMVP vs supervised SVMperf (safeness) | UMVP · 10 datasets | Pre@k win/tie/loss: 6/4/0 | Helped |
| increase number of semi-supervised learners b (ablation) | UMVP · 10 datasets | avg relative improvement | Helped |
| use UMVP vs SVMperf on w8a AUC with uniform weights vs learned weights (ablation) | UMVP · w8a | AUC: 79.1 -> 82.9 | Helped |
| UMVP vs Self-SVMperf training cost | UMVP · real-sim | CPU time (s): 1073.755 -> 129.196 | Helped |
Closest papers by meaning
- Learning Safe Prediction for Semi-Supervised Regression (AAAI 2017 · similarity 0.88)
- Semi-Supervised AUC Optimization Without Guessing Labels of Unlabeled Data (AAAI 2018 · similarity 0.86)
- RETRIEVE: Coreset Selection for Efficient and Robust Semi-Supervised Learning (NeurIPS 2021 · similarity 0.85)
- Erasing the Bias: Fine-Tuning Foundation Models for Semi-Supervised Learning (ICML 2024 · similarity 0.84)
- A Unified Framework for Heterogeneous Semi-supervised Learning (CVPR 2025 · similarity 0.83)
- Bi-Level Optimization for Semi-Supervised Learning with Pseudo-Labeling (AAAI 2025 · similarity 0.83)
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.