Escaping the Linearity Trap: Manifold Detours for Black-Box Adversarial Attacks on Singing Audio Deepfake Detection

arXiv 2605.30366•05de5459f2c1cd7317c30374f24f621dde106f64cf4d9916dc453e1fd79c465c
FinTech-AI-securityLLM-agentsadversarial-MLadversarial-patchesaudio-watermark-removalblack-box-attacksdeepfake-detectiondiffusion-attacksdynamic-separatorsprompt-cachingprompt-injectionregulated-SOCscene-robustnesssecurity-architecturesoftware-reverse-engineering

Paper metadata

arXiv ID
2605.30366
Version
Not specified by this published record
Category
Computer Science — Cryptography and Security (cs.CR)

The PDF link points to arxiv.org. Baitaphish does not expose a private stored PDF.

Evidence and limitations

Source ID
arxiv_cs_cr
Record identifier
05de5459f2c1cd7317c30374f24f621dde106f64cf4d9916dc453e1fd79c465c
Enrichment time
2026-06-01T07:23:34Z
AI-assisted enrichment
Yes

This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.

Escaping the Linearity Trap: Manifold Detours for Black-Box Adversarial Attacks on Singing Audio Deepfake Detection · Baitaphish