TS-ACL: Closed-Form Solution for Time Series-oriented Continual Learning
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866917986532589568 |
|---|---|
| author | Li, Jiaxu Fan, Kejia Lai, Songning Lv, Linpu Xu, Jinfeng Tang, Jianheng Liu, Anfeng Song, Houbing Herbert Yue, Yutao Liu, Yunhuai Zhuang, Huiping |
| author_facet | Li, Jiaxu Fan, Kejia Lai, Songning Lv, Linpu Xu, Jinfeng Tang, Jianheng Liu, Anfeng Song, Houbing Herbert Yue, Yutao Liu, Yunhuai Zhuang, Huiping |
| contents | Time series classification underpins critical applications such as healthcare diagnostics and gesture-driven interactive systems in multimedia scenarios. However, time series class-incremental learning (TSCIL) faces two major challenges: catastrophic forgetting and intra-class variations. Catastrophic forgetting occurs because gradient-based parameter update strategies inevitably erase past knowledge. And unlike images, time series data exhibits subject-specific patterns, also known as intra-class variations, which refer to differences in patterns observed within the same class. While exemplar-based methods fail to cover diverse variation with limited samples, existing exemplar-free methods lack explicit mechanisms to handle intra-class variations. To address these two challenges, we propose TS-ACL, which leverages a gradient-free closed-form solution to avoid the catastrophic forgetting problem inherent in gradient-based optimization methods while simultaneously learning global distributions to resolve intra-class variations. Additionally, it provides privacy protection and efficiency. Extensive experiments on five benchmark datasets covering various sensor modalities and tasks demonstrate that TS-ACL achieves performance close to joint training on four datasets, outperforming existing methods and establishing a new state-of-the-art (SOTA) for TSCIL. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_15954 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | TS-ACL: Closed-Form Solution for Time Series-oriented Continual Learning Li, Jiaxu Fan, Kejia Lai, Songning Lv, Linpu Xu, Jinfeng Tang, Jianheng Liu, Anfeng Song, Houbing Herbert Yue, Yutao Liu, Yunhuai Zhuang, Huiping Machine Learning Artificial Intelligence I.2.6 Time series classification underpins critical applications such as healthcare diagnostics and gesture-driven interactive systems in multimedia scenarios. However, time series class-incremental learning (TSCIL) faces two major challenges: catastrophic forgetting and intra-class variations. Catastrophic forgetting occurs because gradient-based parameter update strategies inevitably erase past knowledge. And unlike images, time series data exhibits subject-specific patterns, also known as intra-class variations, which refer to differences in patterns observed within the same class. While exemplar-based methods fail to cover diverse variation with limited samples, existing exemplar-free methods lack explicit mechanisms to handle intra-class variations. To address these two challenges, we propose TS-ACL, which leverages a gradient-free closed-form solution to avoid the catastrophic forgetting problem inherent in gradient-based optimization methods while simultaneously learning global distributions to resolve intra-class variations. Additionally, it provides privacy protection and efficiency. Extensive experiments on five benchmark datasets covering various sensor modalities and tasks demonstrate that TS-ACL achieves performance close to joint training on four datasets, outperforming existing methods and establishing a new state-of-the-art (SOTA) for TSCIL. |
| title | TS-ACL: Closed-Form Solution for Time Series-oriented Continual Learning |
| topic | Machine Learning Artificial Intelligence I.2.6 |
| url | https://arxiv.org/abs/2410.15954 |