Research map / Label-efficient and robust learning / Semi-supervised learning
Is one annotation enough? - A data-centric image classification benchmark for noisy and ambiguous label estimation
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 |
|---|---|---|---|
| collect more annotations per image (up to 10) instead of one (ablation) | supervised baseline · DCIC datasets (median) | ACC: around 55 -> 69 -> 72 | Helped |
| spend budget on more labeled images versus multiple annotations per image (ablation) | supervised baseline · DCIC | ACC: 55 -> 69 -> 72 | Mixed |
| train fixed model on teacher labels (knowledge distillation) vs raw labels (ablation) | benchmark phase 2 · DCIC | KL | Helped |
| apply state-of-the-art noisy-label/semi-supervised methods at high budget instead of the unadjusted baseline | baseline · DCIC high budget | KL | Hurt |
| use soft labels (Pseudo v2 soft) instead of hard labels | pseudo-label methods · DCIC | KL | Helped |
| use soft labels with 05-0.20 and 10-0.10 initialization (ablation) | Pseudo-Label · DCIC budget 100% | KL | Hurt |
| ImageNet pretraining | top methods · DCIC | KL | Mixed |
Closest papers by meaning
- Towards Good Practices for Efficiently Annotating Large-Scale Image Classification Datasets (CVPR 2021 · similarity 0.91)
- AQuA: A Benchmarking Tool for Label Quality Assessment (NeurIPS 2023 · similarity 0.90)
- Object Detection Benchmarks are Incomplete: The Role of Label Errors and Annotation Uncertainty (NeurIPS 2026 · similarity 0.90)
- Doomed to Re-Annotate, Forever: The ImageNet Story (NeurIPS 2026 · similarity 0.89)
- Enhanced Sample Selection with Confidence Tracking: Identifying Correctly Labeled Yet Hard-to-Learn Samples in Noisy Data (AAAI 2025 · similarity 0.88)
- SELECT: A Large-Scale Benchmark of Data Curation Strategies for Image Classification (NeurIPS 2024 · similarity 0.87)
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.