Research map / Label-efficient and robust learning / Domain adaptation and generalization
Prior Knowledge Guided Unsupervised Domain Adaptation
Tao Sun, Cheng Lu, Haibin Ling
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 unary-bound prior knowledge (sigma=0) rectification | SHOT · Office-Home RS-UT | accuracy: 63.2 -> 66.6 | Improved |
| add unary-bound prior knowledge (sigma=0) rectification | SHOT · DomainNet | accuracy: 79.1 -> 81.0 | Improved |
| add unary-bound prior knowledge (sigma=0) | SHOT · VisDA-2017 | accuracy: 82.9 -> 86.1 | Improved |
| add binary-relationship prior knowledge | SHOT · Office-Home RS-UT | accuracy: 63.2 -> 66.2 | Improved |
| add unary-bound prior knowledge (sigma=0) | SHOT · Office-Home PDA | accuracy: 79.3 -> 86.8 | Improved |
| loosen bound width sigma from 0.0 to 2.0 (ablation) | kSHOT · Office-Home RS-UT | accuracy: 66.6 -> 63.3 | Hurt |
| add unary-bound prior knowledge (sigma=0) | DINE · Office-Home | accuracy: 69.7 -> 71.4 | Improved |
| replace adaptive label smoothing with vanilla smoothing | DINE · Office-Home | accuracy: 69.7 -> 68.4 | Hurt |
| use smoothed labels instead of one-hot labels (ablation) | kDINE · Office-Home | accuracy: 70.7 -> 71.2 | Improved |
| add smooth regularization (ablation) | kSHOT UB sigma=0.0 · Office-Home RS-UT | accuracy: 66.3 -> 66.6 | Improved |
| estimate class prior from 0.5% of target data instead of full data (ablation) | kSHOT · VisDA-2017 | accuracy: 86.13 -> 84.06 | Hurt |
| combine UB sigma=0.5 with BR (ablation) | kSHOT · VisDA-2017 | accuracy: 83.9 -> 84.4 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DINE — DINE: Domain Adaptation From Single and Multiple Black-Box Predictors (CVPR 2022) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Guiding Pseudo-Labels With Uncertainty Estimation for Source-Free Unsupervised Domain Adaptation (CVPR 2023 · similarity 0.93)
- Improving Unsupervised Domain Adaptation: A Pseudo-Candidate Set Approach (ECCV 2024 · similarity 0.93)
- Can We Evaluate Domain Adaptation Models Without Target-Domain Labels? (ICLR 2024 · similarity 0.92)
- Towards Reliable Model Selection for Unsupervised Domain Adaptation: An Empirical Study and A Certified Baseline (NeurIPS 2024 · similarity 0.91)
- Pseudo-Calibration: Improving Predictive Uncertainty Estimation in Unsupervised Domain Adaptation (ICML 2024 · similarity 0.91)
- Universal Semi-Supervised Domain Adaptation by Mitigating Common-Class Bias (CVPR 2024 · similarity 0.90)
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.
