Attentive Merging of Hidden Embeddings from Pre-trained Speech Model for Anti-spoofing Detection
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914834936758272 |
|---|---|
| author | Pan, Zihan Liu, Tianchi Sailor, Hardik B. Wang, Qiongqiong |
| author_facet | Pan, Zihan Liu, Tianchi Sailor, Hardik B. Wang, Qiongqiong |
| contents | Self-supervised learning (SSL) speech representation models, trained on large speech corpora, have demonstrated effectiveness in extracting hierarchical speech embeddings through multiple transformer layers. However, the behavior of these embeddings in specific tasks remains uncertain. This paper investigates the multi-layer behavior of the WavLM model in anti-spoofing and proposes an attentive merging method to leverage the hierarchical hidden embeddings. Results demonstrate the feasibility of fine-tuning WavLM to achieve the best equal error rate (EER) of 0.65%, 3.50%, and 3.19% on the ASVspoof 2019LA, 2021LA, and 2021DF evaluation sets, respectively. Notably, We find that the early hidden transformer layers of the WavLM large model contribute significantly to anti-spoofing task, enabling computational efficiency by utilizing a partial pre-trained model. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_10283 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Attentive Merging of Hidden Embeddings from Pre-trained Speech Model for Anti-spoofing Detection Pan, Zihan Liu, Tianchi Sailor, Hardik B. Wang, Qiongqiong Computation and Language Sound Audio and Speech Processing Self-supervised learning (SSL) speech representation models, trained on large speech corpora, have demonstrated effectiveness in extracting hierarchical speech embeddings through multiple transformer layers. However, the behavior of these embeddings in specific tasks remains uncertain. This paper investigates the multi-layer behavior of the WavLM model in anti-spoofing and proposes an attentive merging method to leverage the hierarchical hidden embeddings. Results demonstrate the feasibility of fine-tuning WavLM to achieve the best equal error rate (EER) of 0.65%, 3.50%, and 3.19% on the ASVspoof 2019LA, 2021LA, and 2021DF evaluation sets, respectively. Notably, We find that the early hidden transformer layers of the WavLM large model contribute significantly to anti-spoofing task, enabling computational efficiency by utilizing a partial pre-trained model. |
| title | Attentive Merging of Hidden Embeddings from Pre-trained Speech Model for Anti-spoofing Detection |
| topic | Computation and Language Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2406.10283 |