Membership Inference Attack Against Music Diffusion Models via Generative Manifold Perturbation
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866908804845666304 |
|---|---|
| author | Liu, Yuxuan Zhang, Peihong Sang, Rui Li, Zhixin Tan, Yizhou Cai, Yiqiang Li, Shengchen |
| author_facet | Liu, Yuxuan Zhang, Peihong Sang, Rui Li, Zhixin Tan, Yizhou Cai, Yiqiang Li, Shengchen |
| contents | Membership inference attacks (MIAs) test whether a specific audio clip was used to train a model, making them a key tool for auditing generative music models for copyright compliance. However, loss-based signals (e.g., reconstruction error) are weakly aligned with human perception in practice, yielding poor separability at the low false-positive rates (FPRs) required for forensics. We propose the Latent Stability Adversarial Probe (LSA-Probe), a white-box method that measures a geometric property of the reverse diffusion: the minimal time-normalized perturbation budget needed to cross a fixed perceptual degradation threshold at an intermediate diffusion state. We show that training members, residing in more stable regions, exhibit a significantly higher degradation cost. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_01645 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Membership Inference Attack Against Music Diffusion Models via Generative Manifold Perturbation Liu, Yuxuan Zhang, Peihong Sang, Rui Li, Zhixin Tan, Yizhou Cai, Yiqiang Li, Shengchen Sound Membership inference attacks (MIAs) test whether a specific audio clip was used to train a model, making them a key tool for auditing generative music models for copyright compliance. However, loss-based signals (e.g., reconstruction error) are weakly aligned with human perception in practice, yielding poor separability at the low false-positive rates (FPRs) required for forensics. We propose the Latent Stability Adversarial Probe (LSA-Probe), a white-box method that measures a geometric property of the reverse diffusion: the minimal time-normalized perturbation budget needed to cross a fixed perceptual degradation threshold at an intermediate diffusion state. We show that training members, residing in more stable regions, exhibit a significantly higher degradation cost. |
| title | Membership Inference Attack Against Music Diffusion Models via Generative Manifold Perturbation |
| topic | Sound |
| url | https://arxiv.org/abs/2602.01645 |