Research map / Label-efficient and robust learning / Recognition and long-tail
Beyond Myopia: Learning from Positive and Unlabeled Data through Holistic Predictive Trends
Wang Xinrui, Wenhai Wan, Chuanxing Geng, Shao-Yuan Li, Songcan Chen
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 resampling of positive examples (ablation) | trend-score PU method · Credit Fraud | recall: 88.6 -> 99.1 | Improved |
| Add resampling of positive examples (ablation) | trend-score PU method · CIFAR10-1 | accuracy: 84.1 -> 91.1 | Improved |
| Add resampling of positive examples (ablation) | trend-score PU method · Alzheimer | F1 score: 69.2 -> 74.5 | Improved |
| Use trend score instead of Mann-Kendall statistic (ablation) | trend-score PU method · Alzheimer | F1 score: 69.7 -> 74.5 | Improved |
| Use trend score instead of Mann-Kendall statistic (ablation) | trend-score PU method · CIFAR10-1 | accuracy: 90.2 -> 91.1 | Improved |
| Use trend score instead of Mann-Kendall statistic (ablation) | trend-score PU method · Credit Fraud | recall: 99.0 -> 99.1 | No clear effect |
| Simplified trend score instead of full trend score (ablation) | trend-score PU method · Alzheimer | F1 score: 74.5 -> 73.9 | Hurt |
| Simplified trend score instead of full trend score (ablation) | trend-score PU method · CIFAR10-1 | accuracy: 91.1 -> 90.7 | Hurt |
| Fisher natural break instead of k-means clustering (ablation) | trend-score PU method · Alzheimer | F1 score: 70.5 -> 74.5 | Improved |
| Fisher natural break instead of k-means clustering (ablation) | trend-score PU method · CIFAR10-1 | accuracy: 89.4 -> 91.1 | Improved |
| Fisher natural break instead of k-means clustering (ablation) | trend-score PU method · Credit Fraud | recall: 99.3 -> 99.1 | No clear effect |
| Proposed method vs best baseline Dist-PU | trend-score PU method · Credit Card unlabeled train set | recall: 76.21 -> 98.90 | Improved |
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.
- Dist-PU — Dist-PU: Positive-Unlabeled Learning From a Label Distribution Perspective (CVPR 2022) · better on 7 of 8 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
- Learning from Positive and Unlabeled Data with Arbitrary Positive Shift (NeurIPS 2020 · similarity 0.89)
- PU-BENCH: A UNIFIED BENCHMARK FOR RIGOROUS AND REPRODUCIBLE PU LEARNING (ICLR 2026 · similarity 0.87)
- Accessible, Realistic, and Fair Evaluation of Positive-Unlabeled Learning Algorithms (ICLR 2026 · similarity 0.87)
- Balancing Positive and Negative Classification Error Rates in Positive-Unlabeled Learning (NeurIPS 2025 · similarity 0.86)
- Learning from Positive and Unlabeled Data with a Selection Bias (ICLR 2019 · similarity 0.86)
- Learning from positive and unlabeled examples -Finite size sample bounds (NeurIPS 2025 · similarity 0.85)
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.
