Research map / Label-efficient and robust learning / Noisy labels
Learning From Noisy Large-Scale Datasets With Minimal Supervision
Check 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 the learned label-cleaning approach instead of the noisy-label baseline | Open Images multi-label classifier · Open Images held-out validation | AP_all: 83.82 -> 87.67 | Helped |
| Use the learned label-cleaning approach instead of the noisy-label baseline | Open Images multi-label classifier · Open Images held-out validation | MAP: 61.82 -> 62.38 | Helped |
| Fine-tune using clean labels only instead of the noisy-label baseline | Open Images multi-label classifier · Open Images held-out validation | MAP: 61.82 -> 61.53 | Hurt |
| Fine-tune using clean labels only instead of the noisy-label baseline | Open Images multi-label classifier · Open Images held-out validation | AP_all: 83.82 -> 85.88 | Helped |
| Fine-tune using mixed clean and noisy labels instead of the noisy-label baseline | Open Images multi-label classifier · Open Images held-out validation | MAP: 61.82 -> 61.90 | Helped |
| Fine-tune using mixed clean and noisy labels instead of the noisy-label baseline | Open Images multi-label classifier · Open Images held-out validation | AP_all: 83.82 -> 84.80 | Helped |
| Train the cleaning network jointly from the beginning rather than pre-training it (ablation) | Our Approach · Open Images held-out validation | AP_all: 87.68 -> 87.67 | No clear effect |
| Train the cleaning network jointly from the beginning rather than pre-training it (ablation) | Our Approach · Open Images held-out validation | MAP: 62.36 -> 62.38 | No clear effect |
| Use the learned label-cleaning approach instead of direct fine-tuning with clean labels | Open Images multi-label classifier · Open Images held-out validation | MAP: 61.53 -> 62.38 | Helped |
| Use the learned label-cleaning approach instead of direct fine-tuning with clean labels | Open Images multi-label classifier · Open Images held-out validation | AP_all: 85.88 -> 87.67 | Helped |
Closest papers by meaning
- Learning with Open-world Noisy Data via Class-independent Margin in Dual Representation Space (AAAI 2025 · similarity 0.85)
- SELECT: A Large-Scale Benchmark of Data Curation Strategies for Image Classification (NeurIPS 2024 · similarity 0.85)
- Data Filtering Networks (ICLR 2024 · similarity 0.85)
- Robust Visual Recognition with Class-Imbalanced Open-World Noisy Data (AAAI 2024 · similarity 0.85)
- Low-Biased General Annotated Dataset Generation (CVPR 2025 · similarity 0.85)
- Understanding and Mitigating the Label Noise in Pre-training on Downstream Tasks (ICLR 2024 · similarity 0.85)
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.