Domain Adaptation 벤치마크
Domain Adaptation on MNIST-to-MNIST-M
Accuracy
- 2015-02-10 — MMD [tzeng2015ddc]; [long2015learning]: Accuracy 76.9
- 2015-05-28 — DANN [ganin2016domain]: Accuracy 77.4
- 2016-08-22 — DSN (DANN): Accuracy 83.2
- 2018-03-27 — DeepJDOT: Accuracy 92.4
- 2021-03-24 — DRANet: Accuracy 98.7
| Rank | Model | Accuracy | Paper | Code | Year |
|---|---|---|---|---|---|
| 1 | DRANet | 98.7 | DRANet: Disentangling Representation and Adaptation Networks for Unsupervised Cross-Domain Adaptation | Seung-Hun-Lee/DRANet | 2021 |
| 2 | DeepJDOT | 92.4 | DeepJDOT: Deep Joint Distribution Optimal Transport for Unsupervised Domain Adaptation | bbdamodaran/deepJDOT · asahi417/DeepDomainAdaptation · CtrlZ1/Domain-Adaptation-Algorithms · +1 | 2018 |
| 3 | DSN (DANN) | 83.2 | Domain Separation Networks | tensorflow/models · tensorflow/models · fungtion/DSN · +3 | 2016 |
| 4 | DANN [ganin2016domain] | 77.4 | Domain-Adversarial Training of Neural Networks | PaddlePaddle/PaddleSpeech · thuml/Transfer-Learning-Library · facebookresearch/DomainBed · +34 | 2015 |
| 5 | MMD [tzeng2015ddc]; [long2015learning] | 76.9 | Learning Transferable Features with Deep Adaptation Networks | thuml/Transfer-Learning-Library · kevinmusgrave/pytorch-adapt · CtrlZ1/Domain-Adaptation-Algorithms · +2 | 2015 |