Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2402.12624 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909112100454400 |
|---|---|
| author | Menezes, Angelo G. Peterlevitz, Augusto J. Chinelatto, Mateus A. de Carvalho, André C. P. L. F. |
| author_facet | Menezes, Angelo G. Peterlevitz, Augusto J. Chinelatto, Mateus A. de Carvalho, André C. P. L. F. |
| contents | Continual Object Detection is essential for enabling intelligent agents to interact proactively with humans in real-world settings. While parameter-isolation strategies have been extensively explored in the context of continual learning for classification, they have yet to be fully harnessed for incremental object detection scenarios. Drawing inspiration from prior research that focused on mining individual neuron responses and integrating insights from recent developments in neural pruning, we proposed efficient ways to identify which layers are the most important for a network to maintain the performance of a detector across sequential updates. The presented findings highlight the substantial advantages of layer-level parameter isolation in facilitating incremental learning within object detection models, offering promising avenues for future research and application in real-world scenarios. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_12624 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Efficient Parameter Mining and Freezing for Continual Object Detection Menezes, Angelo G. Peterlevitz, Augusto J. Chinelatto, Mateus A. de Carvalho, André C. P. L. F. Computer Vision and Pattern Recognition Artificial Intelligence Continual Object Detection is essential for enabling intelligent agents to interact proactively with humans in real-world settings. While parameter-isolation strategies have been extensively explored in the context of continual learning for classification, they have yet to be fully harnessed for incremental object detection scenarios. Drawing inspiration from prior research that focused on mining individual neuron responses and integrating insights from recent developments in neural pruning, we proposed efficient ways to identify which layers are the most important for a network to maintain the performance of a detector across sequential updates. The presented findings highlight the substantial advantages of layer-level parameter isolation in facilitating incremental learning within object detection models, offering promising avenues for future research and application in real-world scenarios. |
| title | Efficient Parameter Mining and Freezing for Continual Object Detection |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2402.12624 |