Research map / Label-efficient and robust learning / Noisy labels
Learning from Crowds by Modeling Common Confusions
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 |
|---|---|---|---|
| model common and individual annotation confusions separately, with per-annotation noise-source selection | CoNAL · LabelMe | accuracy: 85.33 -> 87.12 | Helped |
| model common and individual annotation confusions separately, with per-annotation noise-source selection | CoNAL · Music | accuracy: 81.92 -> 84.06 | Helped |
| add regularization that separates common and annotator-specific noise layers (ablation) | CoNAL · LabelMe | accuracy: 85.68 -> 87.12 | Helped |
| set the difference regularization weight to $10^{-4}$ instead of 0 (ablation) | CoNAL · LabelMe | accuracy: 85.68 -> 86.61 | Helped |
| set the difference regularization weight to $10^{-6}$ instead of 0 (ablation) | CoNAL · LabelMe | accuracy: 85.68 -> 86.26 | Helped |
| set the difference regularization weight to $10^{-5}$ (ablation) | CoNAL · LabelMe | accuracy: 85.68 -> 87.12 | Helped |
| add regularization that separates common and annotator-specific noise layers (ablation) | CoNAL · Music | accuracy: 82.14 -> 84.06 | Helped |
| set the difference regularization weight to $10^{-4}$ instead of 0 (ablation) | CoNAL · Music | accuracy: 82.14 -> 83.52 | Helped |
| set the difference regularization weight to $10^{-6}$ instead of 0 (ablation) | CoNAL · Music | accuracy: 82.14 -> 82.98 | Helped |
| set the difference regularization weight to $10^{-5}$ (ablation) | CoNAL · Music | accuracy: 82.14 -> 84.06 | Helped |
| add another annotator-specific noise layer rather than modeling shared common confusion (ablation) | DL-CL · LabelMe | accuracy: 83.27 -> 82.34 | Hurt |
| add another annotator-specific noise layer rather than modeling shared common confusion (ablation) | DL-CL · Music | accuracy: 81.46 -> 80.47 | Hurt |
Closest papers by meaning
- Coupled Confusion Correction: Learning from Crowds with Sparse Annotations (AAAI 2024 · similarity 0.91)
- Efficient Online Crowdsourcing with Complex Annotations (AAAI 2024 · similarity 0.87)
- Noisy Label Learning with Instance-Dependent Outliers: Identifiability via Crowd Wisdom (NeurIPS 2024 · similarity 0.85)
- Evaluating LLM-contaminated Crowdsourcing Data Without Ground Truth (NeurIPS 2025 · similarity 0.85)
- Unbiased Multi-Label Learning from Crowdsourced Annotations (ICML 2024 · similarity 0.84)
- KFNN: K-Free Nearest Neighbor For Crowdsourcing (NeurIPS 2024 · similarity 0.84)
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.