Research map / Label-efficient and robust learning / Recognition and long-tail
Asking the Right Questions to the Right Users: Active Learning with Imperfect Oracles
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 |
|---|---|---|---|
| jointly select sample-oracle pairs via LP instead of CEAL | batch active learning · CIFAR queried-label accuracy | labeling accuracy (%): 42.20 -> 49.00 | Helped |
| jointly select sample-oracle pairs vs CEAL | batch active learning · SVHN queried-label accuracy | labeling accuracy (%): 75.92 -> 71.66 | Hurt |
| jointly select sample-oracle pairs vs CEAL | batch active learning · Sensor queried-label accuracy | labeling accuracy (%): 98.50 -> 97.91 | Hurt |
| jointly select sample-oracle pairs vs MLAL | batch active learning · Spambase queried-label accuracy | labeling accuracy (%): 79.64 -> 91.42 | Helped |
| always query best oracle with random sampling vs proposed selection | random sampling · Breast Cancer queried-label accuracy | labeling accuracy (%): 94.85 -> 96.57 | Helped |
| vary number of oracles (3, 5, 7) (ablation) | proposed method · Spambase test accuracy | accuracy | No clear effect |
| vary query budget (50 to 150) (ablation) | proposed method · Spambase test accuracy | accuracy | No clear effect |
| query worst oracle always (RW) vs random oracle (RR) | random sampling · MNIST queried-label accuracy | labeling accuracy (%): 70.83 -> 62.12 | Hurt |
Closest papers by meaning
- Search Improves Label for Active Learning (NeurIPS 2016 · similarity 0.90)
- Active Sampling for Text Classification with Subinstance Level Queries (AAAI 2022 · similarity 0.89)
- Active Learning from Imperfect Labelers (NeurIPS 2016 · similarity 0.88)
- Instance-wise Supervision-level Optimization in Active Learning (CVPR 2025 · similarity 0.88)
- Improved Algorithm for Deep Active Learning under Imbalance via Optimal Separation (ICML 2025 · similarity 0.87)
- Inconsistency-Based Data-Centric Active Open-Set Annotation (AAAI 2024 · 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.