Research map / Label-efficient and robust learning / Domain adaptation and generalization
Recursively Conditional Gaussian for Ordinal Unsupervised Domain Adaptation
Xiaofeng Liu, Site Li, Yubin Ge, Pengyi Ye, Jane You, Jun Lu
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 |
|---|---|---|---|
| add adversarial loss (ablation) | RCGUDA 2sigma · DR UDA (KDR to IDRiD) | accuracy (%): 61.5 -> 61.8 | Improved |
| add adversarial loss (ablation) | RCGUDA 3sigma · CHD UDA (echocardiogram) | accuracy (%): 74.5 -> 75.0 | Improved |
| add adversarial loss (ablation) | RCGUDA 2sigma · Age UDA (MORPH to IMDB-WIKI) | accuracy (%): 67.7 -> 68.1 | Improved |
| add adversarial loss (ablation) | RCGUDA 2sigma · DR UDA (KDR to IDRiD) | convergence speed: about 1.2 times | Improved |
| use Wasserstein loss instead of cross-entropy | RCGUDA 2sigma · DR UDA (KDR to IDRiD) | accuracy (%): 61.8 -> 62.3 | Improved |
| use Wasserstein loss instead of cross-entropy | RCGUDA 3sigma · CHD UDA (echocardiogram) | accuracy (%): 75.0 -> 75.6 | Improved |
| use Wasserstein loss instead of cross-entropy | RCGUDA 2sigma · Age UDA (MORPH to IMDB-WIKI) | accuracy (%): 68.1 -> 68.3 | Improved |
| use two-sigma instead of three-sigma RCG constraint | RCGUDA · DR UDA (KDR to IDRiD) | accuracy (%): 61.6 -> 61.8 | Improved |
| use two-sigma instead of three-sigma RCG constraint | RCGUDA · Age UDA (MORPH to IMDB-WIKI) | accuracy (%): 67.8 -> 68.1 | Improved |
| use three-sigma instead of two-sigma RCG constraint | RCGUDA · CHD UDA (echocardiogram) | accuracy (%): 74.6 -> 75.0 | Improved |
| replace non-ordinal UDA (DG-Net++ + Wasserstein) with RCG prior | UDA with disentanglement · DR UDA (KDR to IDRiD) | MAE: 0.48 -> 0.41 | Improved |
| vary sigma factor between 1.8 and 2.4 (ablation) | RCGUDA · Age UDA (MORPH to IMDB-WIKI) | QWK: same QWK from 1.8sigma to 2.4sigma | No clear effect |
Closest papers by meaning
- Ordinal Unsupervised Domain Adaptation With Recursively Conditional Gaussian Imposed Variational Disentanglement (TPAMI 2022 · similarity 0.98)
- Distributionally Robust Classification for Multi-source Unsupervised Domain Adaptation (ICLR 2026 · similarity 0.86)
- CASUAL: Conditional Support Alignment for Domain Adaptation with Label Shift (AAAI 2025 · similarity 0.85)
- Open-Set Heterogeneous Domain Adaptation: Theoretical Analysis and Algorithm (AAAI 2025 · similarity 0.84)
- Robust Domain Generalization under Divergent Marginal and Conditional Distributions (NeurIPS 2026 · similarity 0.84)
- Rethinking Multi-domain Generalization with A General Learning Objective (CVPR 2024 · similarity 0.84)
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.
