Research map / Label-efficient and robust learning / Semi-supervised learning
Towards Professional Level Crowd Annotation of Expert Domain 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 |
|---|---|---|---|
| Use SSL-HF instead of the supervised expert-label baseline | CUB · CUB in-distribution | classification accuracy: 58.7 -> 68.6 (0.6) | Helped |
| Use SSL-HF instead of the supervised expert-label baseline | Fungi · Fungi in-distribution | classification accuracy: 53.8 (0.4) -> 60.0 (0.4) | Helped |
| Use SSL-HF instead of MEMORABLE | Butterflies · Butterflies | classification accuracy: 77.5 (0.5) -> 73.0 (0.7) | Hurt |
| Use SSL-HF instead of MEMORABLE | Gulls · Gulls | classification accuracy: 60.2 (1.1) -> 63.3 (0.5) | Helped |
| Add a negative support set of category names (ablation) | SSL-HF · CUB | labeling accuracy: 60.1 -> 66.2 | Helped |
| Add support-set images instead of using category names only (ablation) | SSL-HF · CUB | labeling accuracy: 66.2 -> 68.7 | Helped |
| Add positive and negative support images rather than positive images only (ablation) | SSL-HF · CUB | classification accuracy: 60.2 -> 65.9 | Helped |
| Add deliberative visual explanations (ablation) | SSL-HF with positive and negative support images · CUB | classification accuracy: 65.9 -> 68.6 | Helped |
| Select high-confidence examples instead of uncertain examples | POSER annotation · CUB | classification accuracy | Helped |
| Use a confidence threshold of 0.25 rather than a high SSL-style threshold (ablation) | SSL-HF · CUB | labeling accuracy: 0.25 | Helped |
Closest papers by meaning
- Lean Multiclass Crowdsourcing (CVPR 2018 · similarity 0.89)
- Lean Crowdsourcing: Combining Humans and Machines in an Online System (CVPR 2017 · similarity 0.87)
- Coupled Confusion Correction: Learning from Crowds with Sparse Annotations (AAAI 2024 · similarity 0.84)
- Mixture of Experts Based Multi-Task Supervise Learning from Crowds (AAAI 2025 · similarity 0.83)
- TLLC: Transfer Learning-based Label Completion for Crowdsourcing (ICML 2025 · similarity 0.83)
- Modeling Collaborator: Enabling Subjective Vision Classification With Minimal Human Effort via LLM Tool-Use (CVPR 2024 · similarity 0.83)
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.