Aux-NAS: Exploiting Auxiliary Labels with Negligibly Extra Inference Cost
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913345257340928 |
|---|---|
| author | Gao, Yuan Zhang, Weizhong Luo, Wenhan Ma, Lin Yu, Jin-Gang Xia, Gui-Song Ma, Jiayi |
| author_facet | Gao, Yuan Zhang, Weizhong Luo, Wenhan Ma, Lin Yu, Jin-Gang Xia, Gui-Song Ma, Jiayi |
| contents | We aim at exploiting additional auxiliary labels from an independent (auxiliary) task to boost the primary task performance which we focus on, while preserving a single task inference cost of the primary task. While most existing auxiliary learning methods are optimization-based relying on loss weights/gradients manipulation, our method is architecture-based with a flexible asymmetric structure for the primary and auxiliary tasks, which produces different networks for training and inference. Specifically, starting from two single task networks/branches (each representing a task), we propose a novel method with evolving networks where only primary-to-auxiliary links exist as the cross-task connections after convergence. These connections can be removed during the primary task inference, resulting in a single-task inference cost. We achieve this by formulating a Neural Architecture Search (NAS) problem, where we initialize bi-directional connections in the search space and guide the NAS optimization converging to an architecture with only the single-side primary-to-auxiliary connections. Moreover, our method can be incorporated with optimization-based auxiliary learning approaches. Extensive experiments with six tasks on NYU v2, CityScapes, and Taskonomy datasets using VGG, ResNet, and ViT backbones validate the promising performance. The codes are available at https://github.com/ethanygao/Aux-NAS. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2405_05695 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Aux-NAS: Exploiting Auxiliary Labels with Negligibly Extra Inference Cost Gao, Yuan Zhang, Weizhong Luo, Wenhan Ma, Lin Yu, Jin-Gang Xia, Gui-Song Ma, Jiayi Machine Learning Artificial Intelligence Computer Vision and Pattern Recognition We aim at exploiting additional auxiliary labels from an independent (auxiliary) task to boost the primary task performance which we focus on, while preserving a single task inference cost of the primary task. While most existing auxiliary learning methods are optimization-based relying on loss weights/gradients manipulation, our method is architecture-based with a flexible asymmetric structure for the primary and auxiliary tasks, which produces different networks for training and inference. Specifically, starting from two single task networks/branches (each representing a task), we propose a novel method with evolving networks where only primary-to-auxiliary links exist as the cross-task connections after convergence. These connections can be removed during the primary task inference, resulting in a single-task inference cost. We achieve this by formulating a Neural Architecture Search (NAS) problem, where we initialize bi-directional connections in the search space and guide the NAS optimization converging to an architecture with only the single-side primary-to-auxiliary connections. Moreover, our method can be incorporated with optimization-based auxiliary learning approaches. Extensive experiments with six tasks on NYU v2, CityScapes, and Taskonomy datasets using VGG, ResNet, and ViT backbones validate the promising performance. The codes are available at https://github.com/ethanygao/Aux-NAS. |
| title | Aux-NAS: Exploiting Auxiliary Labels with Negligibly Extra Inference Cost |
| topic | Machine Learning Artificial Intelligence Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2405.05695 |