Disentangled Training with Adversarial Examples For Robust Small-footprint Keyword Spotting
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866916368811556864 |
|---|---|
| author | Wang, Zhenyu Wan, Li Zhang, Biqiao Huang, Yiteng Li, Shang-Wen Sun, Ming Lei, Xin Yang, Zhaojun |
| author_facet | Wang, Zhenyu Wan, Li Zhang, Biqiao Huang, Yiteng Li, Shang-Wen Sun, Ming Lei, Xin Yang, Zhaojun |
| contents | A keyword spotting (KWS) engine that is continuously running on device is exposed to various speech signals that are usually unseen before. It is a challenging problem to build a small-footprint and high-performing KWS model with robustness under different acoustic environments. In this paper, we explore how to effectively apply adversarial examples to improve KWS robustness. We propose datasource-aware disentangled learning with adversarial examples to reduce the mismatch between the original and adversarial data as well as the mismatch across original training datasources. The KWS model architecture is based on depth-wise separable convolution and a simple attention module. Experimental results demonstrate that the proposed learning strategy improves false reject rate by $40.31%$ at $1%$ false accept rate on the internal dataset, compared to the strongest baseline without using adversarial examples. Our best-performing system achieves $98.06%$ accuracy on the Google Speech Commands V1 dataset. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2408_13355 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Disentangled Training with Adversarial Examples For Robust Small-footprint Keyword Spotting Wang, Zhenyu Wan, Li Zhang, Biqiao Huang, Yiteng Li, Shang-Wen Sun, Ming Lei, Xin Yang, Zhaojun Sound Artificial Intelligence Audio and Speech Processing A keyword spotting (KWS) engine that is continuously running on device is exposed to various speech signals that are usually unseen before. It is a challenging problem to build a small-footprint and high-performing KWS model with robustness under different acoustic environments. In this paper, we explore how to effectively apply adversarial examples to improve KWS robustness. We propose datasource-aware disentangled learning with adversarial examples to reduce the mismatch between the original and adversarial data as well as the mismatch across original training datasources. The KWS model architecture is based on depth-wise separable convolution and a simple attention module. Experimental results demonstrate that the proposed learning strategy improves false reject rate by $40.31%$ at $1%$ false accept rate on the internal dataset, compared to the strongest baseline without using adversarial examples. Our best-performing system achieves $98.06%$ accuracy on the Google Speech Commands V1 dataset. |
| title | Disentangled Training with Adversarial Examples For Robust Small-footprint Keyword Spotting |
| topic | Sound Artificial Intelligence Audio and Speech Processing |
| url | https://arxiv.org/abs/2408.13355 |