Research map / Label-efficient and robust learning / Semi-supervised learning
A General Formulation for Safely Exploiting Weakly Supervised 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 |
|---|---|---|---|
| SAFEW safe ensemble of base learners vs supervised LR baseline | LR baseline · 20newsgroup average | accuracy: .7876 -> .8302 | Helped |
| SAFEW vs supervised LR baseline | LR baseline · Landmine average | accuracy: .9413 -> .9458 | Helped |
| SAFEW vs 1NN baseline | 1NN · Semi-supervised regression average MSE | MSE: .035 -> .028 | Helped |
| naive combination of source and target data (Original) vs LR baseline | LR baseline · 20newsgroup average | accuracy: .7876 -> .7805 | Hurt |
| add domain adaptation methods MIDA/TrAdaBoost vs LR | LR baseline · Landmine average | accuracy: .9413 -> .9405 (MIDA), .9340 (TrAdaBoost) | Hurt |
| SAFEW vs averaging base learners | Average ensemble · Semi-supervised regression average MSE | MSE: .029 -> .028 | Helped |
| SAFEW tighter weight set M vs Safer | Safer · Semi-supervised regression average MSE | MSE: .030 -> .028 | Helped |
| self-training Self-LS vs 1NN baseline | 1NN · bodyfat MSE | MSE: .019 -> .041 | Hurt |
| SAFEW vs TrAdaBoost | TrAdaBoost · Landmine Domain 23 | accuracy: .9627 -> .9426 | Hurt |
| label-noise learning SAFEW vs compared methods | Sup-SVM · label noise datasets | accuracy | Helped |
Closest papers by meaning
- Loss factorization, weakly supervised learning and label noise robustness (ICML 2016 · similarity 0.89)
- Losses over Labels: Weakly Supervised Learning via Direct Loss Construction (AAAI 2023 · similarity 0.88)
- Consistent Multi-Class Classification from Multiple Unlabeled Datasets (ICLR 2024 · similarity 0.88)
- Label Propagation with Weak Supervision (ICLR 2023 · similarity 0.88)
- A General Framework for Learning from Weak Supervision (ICML 2024 · similarity 0.87)
- Provable weak-to-strong generalization via benign overfitting (ICLR 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.