Name
Abstract | ||
|---|---|---|
Unsupervised domain adaptation techniques have been successful for a wide range of problems where supervised labels are limited. The task is to classify an unlabeled `targetu0027 dataset by leveraging a labeled `sourceu0027 dataset that comes from a slightly similar distribution. We propose metric-based adversarial discriminative domain adaptation (M-ADDA) which performs two main steps. First, it uses a metric learning approach to train the source model on the source dataset by optimizing the triplet loss function. This results in clusters where embeddings of the same label are close to each other and those with different labels are far from one another. Next, it uses the adversarial approach (as that used in ADDA cite{2017arXiv170205464T}) to make the extracted features from the source and target datasets indistinguishable. Simultaneously, we optimize a novel loss function that encourages the target datasetu0027s embeddings to form clusters. While ADDA and M-ADDA use similar architectures, we show that M-ADDA performs significantly better on the digits adaptation datasets of MNIST and USPS. This suggests that using metric-learning for domain adaptation can lead to large improvements in classification accuracy for the domain adaptation task. The code is available at url{this https URL}. |
Year | DOI | Venue |
|---|---|---|
2018 | 10.1007/978-3-030-30671-7_2 | arXiv: Learning |
Field | DocType | Volume |
MNIST database,Domain adaptation,Source model,Artificial intelligence,Discriminative model,Mathematics,Machine learning | Journal | abs/1807.02552 |
Citations | PageRank | References |
3 | 0.38 | 21 |
Authors | ||
2 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Issam H. Laradji | 1 | 79 | 9.40 |
| Reza Babanezhad | 2 | 17 | 1.36 |