Research map / Label-efficient and robust learning / Recognition and long-tail
Reliable Programmatic Weak Supervision With Confidence Intervals for Label Probabilities
Verónica Álvarez, Santiago Mazuelas, Steven An, Sanjoy Dasgupta
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 |
|---|---|---|---|
| use MMP instead of AMCL | weak supervision label model · SMS | Brier score: .25 -> .12 | Improved |
| use MMP instead of AMCL | weak supervision label model · OBS | Brier score: .27 -> .23 | Improved |
| use MMP instead of AMCL | weak supervision label model · Domain | Brier score: .54 -> .55 | No clear effect |
| use MMP instead of Snorkel | weak supervision label model · Cardio | log-loss: 7.01 -> .21 | Improved |
| use MMP instead of majority vote | weak supervision label model · OBS | log-loss: 8.73 -> .61 | Improved |
| use MMP instead of AMCL | weak supervision label model · IMDB | calibration error: .47 -> .21 | Improved |
| use MMP instead of Snorkel | weak supervision label model · Basket | calibration error: .14 -> .20 | Hurt |
| use MMP instead of FlyingSquid | weak supervision label model · Awa | 0-1 loss: .01 -> .01 | No clear effect |
| use MMP instead of EBCC | weak supervision label model · SMS | 0-1 loss: .08 -> .08 | No clear effect |
| add more labeling functions | WMRC confidence intervals · Awa | interval tightness | Improved |
| raise the abstention threshold on predicted probability | WMRC · Awa | Brier score | Improved |
| use standard sample errors instead of Wilson interval for lambda (ablation) | WMRC confidence intervals · Awa | coverage of actual probabilities | No clear effect |
Closest papers by meaning
- Statistical Analysis of an Adversarial Bayesian Weak Supervision Method (NeurIPS 2025 · similarity 0.90)
- Characterizing the Impacts of Semi-supervised Learning for Weak Supervision (NeurIPS 2023 · similarity 0.90)
- Learning Hyper Label Model for Programmatic Weak Supervision (ICLR 2023 · similarity 0.89)
- Weak Supervision Performance Evaluation via Partial Identification (NeurIPS 2024 · similarity 0.89)
- Label Propagation with Weak Supervision (ICLR 2023 · similarity 0.89)
- Unified Approach for Weakly Supervised Multicalibration (NeurIPS 2026 · 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.
