Task-Adaptive Saliency Guidance for Exemplar-free Class Incremental Learning
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2022
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866913285252579328 |
|---|---|
| author | Liu, Xialei Zhai, Jiang-Tian Bagdanov, Andrew D. Li, Ke Cheng, Ming-Ming |
| author_facet | Liu, Xialei Zhai, Jiang-Tian Bagdanov, Andrew D. Li, Ke Cheng, Ming-Ming |
| contents | Exemplar-free Class Incremental Learning (EFCIL) aims to sequentially learn tasks with access only to data from the current one. EFCIL is of interest because it mitigates concerns about privacy and long-term storage of data, while at the same time alleviating the problem of catastrophic forgetting in incremental learning. In this work, we introduce task-adaptive saliency for EFCIL and propose a new framework, which we call Task-Adaptive Saliency Supervision (TASS), for mitigating the negative effects of saliency drift between different tasks. We first apply boundary-guided saliency to maintain task adaptivity and \textit{plasticity} on model attention. Besides, we introduce task-agnostic low-level signals as auxiliary supervision to increase the \textit{stability} of model attention. Finally, we introduce a module for injecting and recovering saliency noise to increase the robustness of saliency preservation. Our experiments demonstrate that our method can better preserve saliency maps across tasks and achieve state-of-the-art results on the CIFAR-100, Tiny-ImageNet, and ImageNet-Subset EFCIL benchmarks. Code is available at \url{https://github.com/scok30/tass}. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2212_08251 |
| institution | arXiv |
| publishDate | 2022 |
| record_format | arxiv |
| spellingShingle | Task-Adaptive Saliency Guidance for Exemplar-free Class Incremental Learning Liu, Xialei Zhai, Jiang-Tian Bagdanov, Andrew D. Li, Ke Cheng, Ming-Ming Computer Vision and Pattern Recognition Exemplar-free Class Incremental Learning (EFCIL) aims to sequentially learn tasks with access only to data from the current one. EFCIL is of interest because it mitigates concerns about privacy and long-term storage of data, while at the same time alleviating the problem of catastrophic forgetting in incremental learning. In this work, we introduce task-adaptive saliency for EFCIL and propose a new framework, which we call Task-Adaptive Saliency Supervision (TASS), for mitigating the negative effects of saliency drift between different tasks. We first apply boundary-guided saliency to maintain task adaptivity and \textit{plasticity} on model attention. Besides, we introduce task-agnostic low-level signals as auxiliary supervision to increase the \textit{stability} of model attention. Finally, we introduce a module for injecting and recovering saliency noise to increase the robustness of saliency preservation. Our experiments demonstrate that our method can better preserve saliency maps across tasks and achieve state-of-the-art results on the CIFAR-100, Tiny-ImageNet, and ImageNet-Subset EFCIL benchmarks. Code is available at \url{https://github.com/scok30/tass}. |
| title | Task-Adaptive Saliency Guidance for Exemplar-free Class Incremental Learning |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2212.08251 |