Research map / Label-efficient and robust learning / Domain adaptation and generalization
A Difference Standardization Method for Mutual Transfer Learning
Haoqing Xu, Meng Wang, Beilun Wang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| Standardize domain differences (divide by Sigma_i+Sigma_j)^-1/2 before l2 distance (ablation) | DiffS · Synthetic (M=50, n=100, p=q=10, S=5) | learnability structure NMI: 0.998 vs 0.902 | Improved |
| Use learnability-structure-aware DiffS instead of k-means with BIC tuning | DiffS · Synthetic large-scale settings | wall-clock time: about 50 to 1000x faster than k-means, 2 to 20x faster than MeTaG | Improved |
| Use DiffS (tuning-free) instead of k-means with BIC for subgroup recovery | Synthetic high-dimensional settings (p=q=30, 50) | learnability structure NMI: 0.997 vs 0.047 (p=q=30); 0.954 vs 0.000 (p=q=50) | Improved |
| Use DiffS instead of k-means on NOAA nClimDiv | monthly temperature prediction · Real: NOAA nClimDiv | MSE: 82.44 vs 88.72 | Improved |
| Use DiffS instead of MeTaG on NOAA nClimDiv | monthly temperature prediction · Real: NOAA nClimDiv | MSE and time: MeTaG MSE explodes and needs >10200 s, DiffS 82.44 in 33.79 s | Improved |
| Use DiffS instead of k-means/MeTaG on Microarray data | gene expression prediction · Real: Microarray Data | MSE: 0.93 vs 0.96 (k-means) and 0.90 (MeTaG) | Improved |
| Use DiffS instead of MeTaG on Microarray data | gene expression prediction · Real: Microarray Data | MSE: MeTaG 0.90 is lower than DiffS 0.93 within noise | Mixed |
Closest papers by meaning
- Transfer Learning via Minimizing the Performance Gap Between Domains (NeurIPS 2019 · similarity 0.89)
- Transferable Calibration with Lower Bias and Variance in Domain Adaptation (NeurIPS 2020 · similarity 0.87)
- On Learning Domain-Invariant Representations for Transfer Learning with Multiple Sources (NeurIPS 2021 · similarity 0.87)
- Wasserstein Transfer Learning (NeurIPS 2025 · similarity 0.87)
- Theoretical Performance Guarantees for Partial Domain Adaptation via Partial Optimal Transport (ICML 2025 · similarity 0.85)
- Function Encoders: A Principled Approach to Transfer Learning in Hilbert Spaces (ICML 2025 · 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.
