Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2022
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2205.15523 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866929229284769792 |
|---|---|
| author | Hou, Jinyong Deng, Jeremiah D. Cranefield, Stephen Din, Xuejie |
| author_facet | Hou, Jinyong Deng, Jeremiah D. Cranefield, Stephen Din, Xuejie |
| contents | To successfully apply trained neural network models to new domains, powerful transfer learning solutions are essential. We propose to introduce a novel cross-domain latent modulation mechanism to a variational autoencoder framework so as to achieve effective transfer learning. Our key idea is to procure deep representations from one data domain and use it to influence the reparameterization of the latent variable of another domain. Specifically, deep representations of the source and target domains are first extracted by a unified inference model and aligned by employing gradient reversal. The learned deep representations are then cross-modulated to the latent encoding of the alternative domain, where consistency constraints are also applied. In the empirical validation that includes a number of transfer learning benchmark tasks for unsupervised domain adaptation and image-to-image translation, our model demonstrates competitive performance, which is also supported by evidence obtained from visualization. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2205_15523 |
| institution | arXiv |
| publishDate | 2022 |
| record_format | arxiv |
| spellingShingle | Variational Transfer Learning using Cross-Domain Latent Modulation Hou, Jinyong Deng, Jeremiah D. Cranefield, Stephen Din, Xuejie Machine Learning Artificial Intelligence Computer Vision and Pattern Recognition To successfully apply trained neural network models to new domains, powerful transfer learning solutions are essential. We propose to introduce a novel cross-domain latent modulation mechanism to a variational autoencoder framework so as to achieve effective transfer learning. Our key idea is to procure deep representations from one data domain and use it to influence the reparameterization of the latent variable of another domain. Specifically, deep representations of the source and target domains are first extracted by a unified inference model and aligned by employing gradient reversal. The learned deep representations are then cross-modulated to the latent encoding of the alternative domain, where consistency constraints are also applied. In the empirical validation that includes a number of transfer learning benchmark tasks for unsupervised domain adaptation and image-to-image translation, our model demonstrates competitive performance, which is also supported by evidence obtained from visualization. |
| title | Variational Transfer Learning using Cross-Domain Latent Modulation |
| topic | Machine Learning Artificial Intelligence Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2205.15523 |