Research map / Label-efficient and robust learning / Noisy labels
From Noisy Prediction to True Label: Noisy Prediction Calibration via Generative Model
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 |
|---|---|---|---|
| Add NPC post-processing to classifiers trained with noisy labels | Baseline classifiers on MNIST, Fashion-MNIST, and CIFAR-10 · Synthetic noisy-label settings | accuracy | Helped |
| Add NPC post-processing to a CE classifier | CE on MNIST · MNIST IDN 40% | accuracy: 66.3 -> 89.1 | Helped |
| Add NPC post-processing to a CE classifier | CE on Fashion-MNIST · Fashion-MNIST SN 20% | accuracy: 74.0 -> 84.0 | Helped |
| Add NPC post-processing to a CE classifier | CE on CIFAR-10 · CIFAR-10 SN 20% | accuracy: 73.1 -> 80.8 | Helped |
| Add NPC post-processing to a Joint classifier | Joint on Fashion-MNIST · Fashion-MNIST SN 80% | accuracy: 6.0 -> 6.0 | No clear effect |
| Add NPC post-processing to a CORES2 classifier | CORES2 on CIFAR-10 · CIFAR-10 SN 80% | accuracy: 31.2 -> 30.7 | Hurt |
| Add NPC post-processing to a REL classifier | REL on Fashion-MNIST · Fashion-MNIST clean | accuracy: 88.1 -> 86.9 | Hurt |
| Add NPC post-processing to baseline classifiers | Classifiers on Food-101 and Clothing1M · Real-world noisy-label datasets | accuracy | Mixed |
| Add NPC post-processing to CE | CE on Food-101 · Food-101 | accuracy: 78.37 -> 80.21 | Helped |
| Add NPC post-processing to DualT | DualT on Clothing1M · Clothing1M | accuracy: 70.18 -> 69.99 | Hurt |
| Estimate the transition matrix through NPC | NPC on CIFAR-10 · CIFAR-10 IDN 40% | MSE: 0.0049 -> 0.0018 | Helped |
Closest papers by meaning
- Error-Bounded Correction of Noisy Labels (ICML 2020 · similarity 0.89)
- Dual T: Reducing Estimation Error for Transition Matrix in Label-noise Learning (NeurIPS 2020 · similarity 0.89)
- Safeguarded Dynamic Label Regression for Noisy Supervision (AAAI 2019 · similarity 0.89)
- Estimating Noisy Class Posterior with Part-level Labels for Noisy Label Learning (CVPR 2024 · similarity 0.88)
- Learning the Latent Causal Structure for Modeling Label Noise (NeurIPS 2024 · similarity 0.88)
- Learning Causal Transition Matrix for Instance-dependent Label Noise (AAAI 2025 · 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.