LlamaPartialSpoof: An LLM-Driven Fake Speech Dataset Simulating Disinformation Generation
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_ | 1866916551885586432 |
|---|---|
| author | Luong, Hieu-Thi Li, Haoyang Zhang, Lin Lee, Kong Aik Chng, Eng Siong |
| author_facet | Luong, Hieu-Thi Li, Haoyang Zhang, Lin Lee, Kong Aik Chng, Eng Siong |
| contents | Previous fake speech datasets were constructed from a defender's perspective to develop countermeasure (CM) systems without considering diverse motivations of attackers. To better align with real-life scenarios, we created LlamaPartialSpoof, a 130-hour dataset that contains both fully and partially fake speech, using a large language model (LLM) and voice cloning technologies to evaluate the robustness of CMs. By examining valuable information for both attackers and defenders, we identify several key vulnerabilities in current CM systems, which can be exploited to enhance attack success rates, including biases toward certain text-to-speech models or concatenation methods. Our experimental results indicate that the current fake speech detection system struggle to generalize to unseen scenarios, achieving a best performance of 24.49% equal error rate. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_14743 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | LlamaPartialSpoof: An LLM-Driven Fake Speech Dataset Simulating Disinformation Generation Luong, Hieu-Thi Li, Haoyang Zhang, Lin Lee, Kong Aik Chng, Eng Siong Audio and Speech Processing Sound Previous fake speech datasets were constructed from a defender's perspective to develop countermeasure (CM) systems without considering diverse motivations of attackers. To better align with real-life scenarios, we created LlamaPartialSpoof, a 130-hour dataset that contains both fully and partially fake speech, using a large language model (LLM) and voice cloning technologies to evaluate the robustness of CMs. By examining valuable information for both attackers and defenders, we identify several key vulnerabilities in current CM systems, which can be exploited to enhance attack success rates, including biases toward certain text-to-speech models or concatenation methods. Our experimental results indicate that the current fake speech detection system struggle to generalize to unseen scenarios, achieving a best performance of 24.49% equal error rate. |
| title | LlamaPartialSpoof: An LLM-Driven Fake Speech Dataset Simulating Disinformation Generation |
| topic | Audio and Speech Processing Sound |
| url | https://arxiv.org/abs/2409.14743 |