TensorFlow Version Used: 2.0
This is an implementation of the paper Domain-Adversarial Training of Neural Networks(https://arxiv.org/abs/1505.07818). I have tested my implementation on 3 datasets ( MNIST, SVHN, SynNumbers). Paper focuses upon Unsupervised Domain Adaptation via Backpropagation.
| MNIST | SVHN | SynNumbers | |
|---|---|---|---|
| Source Training Accuracy | 99.9% | 99.7% | 99.8% |
| Source Test Accuracy | 99.9% | 91.5% | 96.4% |
| Target Accuracy(MNIST) | - | 71.3% | 89.2% |
| Target Accuracy(SVHN) | 20.2% | - | 84.8% |
| Target Accuracy(SynNumbers) | 24.5% | 91.1% | - |
I have used this implementation my research work for CT lung Images for Interstitial lung diseases (ILD).
This Repo will be updated soon with full description.