Frequency-mix Knowledge Distillation for Fake Speech Detection
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866917693956816896 |
|---|---|
| author | Fan, Cunhang Dong, Shunbo Xue, Jun Chen, Yujie Yi, Jiangyan Lv, Zhao |
| author_facet | Fan, Cunhang Dong, Shunbo Xue, Jun Chen, Yujie Yi, Jiangyan Lv, Zhao |
| contents | In the telephony scenarios, the fake speech detection (FSD) task to combat speech spoofing attacks is challenging. Data augmentation (DA) methods are considered effective means to address the FSD task in telephony scenarios, typically divided into time domain and frequency domain stages. While each has its advantages, both can result in information loss. To tackle this issue, we propose a novel DA method, Frequency-mix (Freqmix), and introduce the Freqmix knowledge distillation (FKD) to enhance model information extraction and generalization abilities. Specifically, we use Freqmix-enhanced data as input for the teacher model, while the student model's input undergoes time-domain DA method. We use a multi-level feature distillation approach to restore information and improve the model's generalization capabilities. Our approach achieves state-of-the-art results on ASVspoof 2021 LA dataset, showing a 31\% improvement over baseline and performs competitively on ASVspoof 2021 DF dataset. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_09664 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Frequency-mix Knowledge Distillation for Fake Speech Detection Fan, Cunhang Dong, Shunbo Xue, Jun Chen, Yujie Yi, Jiangyan Lv, Zhao Sound Audio and Speech Processing In the telephony scenarios, the fake speech detection (FSD) task to combat speech spoofing attacks is challenging. Data augmentation (DA) methods are considered effective means to address the FSD task in telephony scenarios, typically divided into time domain and frequency domain stages. While each has its advantages, both can result in information loss. To tackle this issue, we propose a novel DA method, Frequency-mix (Freqmix), and introduce the Freqmix knowledge distillation (FKD) to enhance model information extraction and generalization abilities. Specifically, we use Freqmix-enhanced data as input for the teacher model, while the student model's input undergoes time-domain DA method. We use a multi-level feature distillation approach to restore information and improve the model's generalization capabilities. Our approach achieves state-of-the-art results on ASVspoof 2021 LA dataset, showing a 31\% improvement over baseline and performs competitively on ASVspoof 2021 DF dataset. |
| title | Frequency-mix Knowledge Distillation for Fake Speech Detection |
| topic | Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2406.09664 |