Research map / Label-efficient and robust learning / Domain adaptation and generalization
Label-Noise Robust Domain Adaptation
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 |
|---|---|---|---|
| replace CIC/DAN-style alignment with denoising MMD (reweighting by corrected noisy-label weights) | domain adaptation with noisy source labels · MNIST/USPS LeNet, rho=0.4 and 0.2 | accuracy: e.g. 69.94 vs 65.37 (mnist->usps, rho=0.4) vs CIC+Forward Q | Helped |
| use denoising MMD (DCIC) instead of CIC for invariant component extraction | VLCS with AlexNet, rho=0.4 · VLCS | accuracy: 91.60 vs 86.83 on VLS2C; 52.47 vs 49.30 on VCS2L vs CIC | Helped |
| adding forward loss correction (true Q) to CIC | CIC on MNIST/USPS · MNIST/USPS | accuracy: 65.37 vs 65.23 (mnist->usps rho=0.4); 58.09 to 63.35 (usps->mnist rho=0.4) | Mixed |
| estimated transition matrix instead of true Q in forward correction (ablation) | DCIC on MNIST/USPS · MNIST/USPS | accuracy: e.g. 68.50 vs 69.94 (mnist->usps rho=0.4); 69.29 vs 72.33 (rho=0.2) | Hurt |
| denoising MMD (DCIC) vs CIC for class ratio estimation under target shift with noisy labels (ablation) | linear model on synthetic Gaussian data · synthetic mixture of Gaussians | class ratio estimation error | Helped |
| adding denoising MMD over CIC on invariant component extraction at increasing noise (ablation) | synthetic GeTarS linear model · synthetic mixture of Gaussians | classification accuracy | Helped |
Closest papers by meaning
- Domain Adaptation with Conditional Distribution Matching and Generalized Label Shift (NeurIPS 2020 · similarity 0.91)
- Confusing Pair Correction Based on Category Prototype for Domain Adaptation under Noisy Environments (AAAI 2024 · similarity 0.89)
- Label Noise: Ignorance Is Bliss (NeurIPS 2024 · similarity 0.89)
- A General Representation-Based Approach to Multi-Source Domain Adaptation (ICML 2025 · similarity 0.88)
- Can We Evaluate Domain Adaptation Models Without Target-Domain Labels? (ICLR 2024 · similarity 0.88)
- Test-Time Domain Adaptation by Learning Domain-Aware Batch Normalization (AAAI 2024 · similarity 0.88)
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.