Deep Nets with Subsampling Layers Unwittingly Discard Useful Activations at Test-Time
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_ | 1866929523750076416 |
|---|---|
| author | Yang, Chiao-An Liu, Ziwei Yeh, Raymond A. |
| author_facet | Yang, Chiao-An Liu, Ziwei Yeh, Raymond A. |
| contents | Subsampling layers play a crucial role in deep nets by discarding a portion of an activation map to reduce its spatial dimensions. This encourages the deep net to learn higher-level representations. Contrary to this motivation, we hypothesize that the discarded activations are useful and can be incorporated on the fly to improve models' prediction. To validate our hypothesis, we propose a search and aggregate method to find useful activation maps to be used at test time. We applied our approach to the task of image classification and semantic segmentation. Extensive experiments over nine different architectures on multiple datasets show that our method consistently improves model test-time performance, complementing existing test-time augmentation techniques. Our code is available at https://github.com/ca-joe-yang/discard-in-subsampling. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_01083 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Deep Nets with Subsampling Layers Unwittingly Discard Useful Activations at Test-Time Yang, Chiao-An Liu, Ziwei Yeh, Raymond A. Computer Vision and Pattern Recognition Machine Learning Subsampling layers play a crucial role in deep nets by discarding a portion of an activation map to reduce its spatial dimensions. This encourages the deep net to learn higher-level representations. Contrary to this motivation, we hypothesize that the discarded activations are useful and can be incorporated on the fly to improve models' prediction. To validate our hypothesis, we propose a search and aggregate method to find useful activation maps to be used at test time. We applied our approach to the task of image classification and semantic segmentation. Extensive experiments over nine different architectures on multiple datasets show that our method consistently improves model test-time performance, complementing existing test-time augmentation techniques. Our code is available at https://github.com/ca-joe-yang/discard-in-subsampling. |
| title | Deep Nets with Subsampling Layers Unwittingly Discard Useful Activations at Test-Time |
| topic | Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2410.01083 |