Salvato in:
| Autori principali: | , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2023
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2308.06948 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911887616114688 |
|---|---|
| author | Liang, Yu Zhang, Yufeng Zhang, Shiliang Wang, Yaowei Xiao, Sheng Xiao, Rong Wang, Xiaoyu |
| author_facet | Liang, Yu Zhang, Yufeng Zhang, Shiliang Wang, Yaowei Xiao, Sheng Xiao, Rong Wang, Xiaoyu |
| contents | Backward-compatible training circumvents the need for expensive updates to the old gallery database when deploying an advanced new model in the retrieval system. Previous methods achieved backward compatibility by aligning prototypes of the new model with the old one, yet they often overlooked the distribution of old features, limiting their effectiveness when the low quality of the old model results in a weakly feature discriminability. Instance-based methods like L2 regression take into account the distribution of old features but impose strong constraints on the performance of the new model itself. In this paper, we propose MixBCT, a simple yet highly effective backward-compatible training method that serves as a unified framework for old models of varying qualities. We construct a single loss function applied to mixed old and new features to facilitate backward-compatible training, which adaptively adjusts the constraint domain for new features based on the distribution of old features. We conducted extensive experiments on the large-scale face recognition datasets MS1Mv3 and IJB-C to verify the effectiveness of our method. The experimental results clearly demonstrate its superiority over previous methods. Code is available at https://github.com/yuleung/MixBCT . |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2308_06948 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | MixBCT: Towards Self-Adapting Backward-Compatible Training Liang, Yu Zhang, Yufeng Zhang, Shiliang Wang, Yaowei Xiao, Sheng Xiao, Rong Wang, Xiaoyu Computer Vision and Pattern Recognition Backward-compatible training circumvents the need for expensive updates to the old gallery database when deploying an advanced new model in the retrieval system. Previous methods achieved backward compatibility by aligning prototypes of the new model with the old one, yet they often overlooked the distribution of old features, limiting their effectiveness when the low quality of the old model results in a weakly feature discriminability. Instance-based methods like L2 regression take into account the distribution of old features but impose strong constraints on the performance of the new model itself. In this paper, we propose MixBCT, a simple yet highly effective backward-compatible training method that serves as a unified framework for old models of varying qualities. We construct a single loss function applied to mixed old and new features to facilitate backward-compatible training, which adaptively adjusts the constraint domain for new features based on the distribution of old features. We conducted extensive experiments on the large-scale face recognition datasets MS1Mv3 and IJB-C to verify the effectiveness of our method. The experimental results clearly demonstrate its superiority over previous methods. Code is available at https://github.com/yuleung/MixBCT . |
| title | MixBCT: Towards Self-Adapting Backward-Compatible Training |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2308.06948 |