Research map / Label-efficient and robust learning / Domain adaptation and generalization
Gradual Domain Adaptation: Theory and Algorithms
Yifei He, Haoxiang Wang, Bo Li, Han Zhao
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 generated intermediate domains (4 per pair) to GST | gradual self-training · Rotated MNIST (2 given domains) | accuracy (%): 61.6 -> 70.3 | Improved |
| add generated intermediate domains (4 per pair) to GST | gradual self-training · Color-Shift MNIST (2 given domains) | accuracy (%): 67.6 -> 90.3 | Improved |
| add generated intermediate domains (3 per pair) to GST | gradual self-training · Portraits (2 given domains) | accuracy (%): 77.0 -> 79.9 | Improved |
| add generated intermediate domains (4 per pair) to GST | gradual self-training · Cover Type (2 given domains) | accuracy (%): 66.9 -> 69.8 | Improved |
| add generated intermediate domains (1 per pair) with no given domains | GST · Rotated MNIST (0 given domains) | accuracy (%): 50.3 -> 48.5 | Hurt |
| add generated domains (4 per pair) with no given domains | GST · Color-Shift MNIST (0 given domains) | accuracy (%): 40.5 -> 79.1 | Improved |
| use GOAT instead of one-off UDA (DANN) | DANN · Rotated MNIST | accuracy (%): 44.6 -> 70.3 | Improved |
| use GOAT instead of DeepJDOT | DeepJDOT · Color-Shift MNIST | accuracy (%): 65.8 -> 90.3 | Improved |
| use GOAT instead of CoVi | CoVi · Rotated MNIST (2 given domains) | accuracy (%): 64.2 -> 70.3 | Improved |
| use GOAT instead of CoVi | CoVi · Color-Shift MNIST (2 given domains) | accuracy (%): 77.6 -> 90.3 | Improved |
| use GOAT instead of CoVi | CoVi · Portraits (2 given domains) | accuracy (%): 79.8 -> 79.9 | No clear effect |
| add more given intermediate domains to GOAT (0 -> 4) | GOAT · Rotated MNIST | accuracy (%): 47.5 -> 82.7 | 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.
- DeepJDOT — DeepJDOT: Deep Joint Distribution Optimal Transport for Unsupervised Domain Adaptation (ECCV 2018) · better on 4 of 4 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
- Understanding Gradual Domain Adaptation: Improved Analysis, Optimal Path and Beyond (ICML 2022 · similarity 0.96)
- Gradual Domain Adaptation without Indexed Intermediate Domains (NeurIPS 2021 · similarity 0.94)
- Gradual Domain Adaptation via Gradient Flow (ICLR 2024 · similarity 0.94)
- Self-Training with Dynamic Weighting for Robust Gradual Domain Adaptation (NeurIPS 2025 · similarity 0.90)
- Open-Set Heterogeneous Domain Adaptation: Theoretical Analysis and Algorithm (AAAI 2025 · similarity 0.89)
- Universal Semi-Supervised Domain Adaptation by Mitigating Common-Class Bias (CVPR 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.
