Research map / Label-efficient and robust learning / Domain adaptation and generalization
MixStyle Neural Networks for Domain Generalization and Adaptation
Kaiyang Zhou, Yongxin Yang, Yu Qiao, Tao Xiang
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 MixStyle (random) | ResNet-18 · PACS (DG, avg) | accuracy: 79.5 -> 82.8 | Improved |
| add MixStyle (random) to Office-Home | ResNet-18 · Office-Home (DG, avg) | accuracy: 64.7 -> 65.5 | No clear effect |
| add MixStyle under DomainBed | ERM · Office-Home (DomainBed) | accuracy: 67.6 -> 60.4 | Hurt |
| add MixStyle under DomainBed | ERM · PACS (DomainBed) | accuracy: 84.2 -> 85.2 | Improved |
| add MixStyle (random) | ResNet-50 · Market1501->Duke re-ID | mAP: 19.3 -> 23.8 | Improved |
| add MixStyle to FixMatch | FixMatch · VisDA-17 (UDA) | accuracy: 77.1 -> 80.0 | Improved |
| add MixStyle to FixMatch | FixMatch · PACS (multi-source UDA, avg) | accuracy: 90.30 -> 92.55 | Improved |
| apply MixStyle after res1-3 instead of res1 (ablation) | ResNet-18 · PACS (DG, avg) | accuracy: 80.1 -> 82.8 | Improved |
| also apply MixStyle after res4 (all four blocks) (ablation) | ResNet-18 · PACS (DG, avg) | accuracy: 82.8 -> 75.6 | Hurt |
| apply MixStyle after res1-2 instead of res1-4 (ablation) | ResNet-50 · Market1501->Duke re-ID | mAP: 10.2 -> 23.8 | Improved |
| mix styles instead of replacing them (ablation) | MixStyle · PACS (DG, avg) | accuracy: 82.1 -> 82.8 | Improved |
| random shuffle at each layer instead of fixed shuffle (ablation) | MixStyle · PACS (DG, avg) | accuracy: 82.4 -> 82.8 | 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.
- FixMatch — FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence (NeurIPS 2020) · better on 2 of 2 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
- Domain Generalization with MixStyle (ICLR 2021 · similarity 0.96)
- Style Neophile: Constantly Seeking Novel Styles for Domain Generalization (CVPR 2022 · similarity 0.87)
- Mixup-Induced Domain Extrapolation for Domain Generalization (AAAI 2024 · similarity 0.85)
- Selective Mixup Helps with Distribution Shifts, But Not (Only) because of Mixup (ICML 2024 · similarity 0.84)
- On Unsupervised Domain Adaptation: Pseudo Label Guided Mixup for Adversarial Prompt Tuning (AAAI 2024 · similarity 0.84)
- Style Adaptation and Uncertainty Estimation for Multi-Source Blended-Target Domain Adaptation (NeurIPS 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.
