Back to AI News
arxiv

AUTEUR: Acoustic Plan Learning for Red-Teaming Text-to-Audio Models

AUTEUR introduces acoustic plan learning, a red-teaming framework that represents test prompts as structured acoustic plans covering foreground sounds, context, and vocal attributes. It trains a plan generator with supervised fine-tuning and reinforcement learning using waveform feedback while keeping the target text-to-audio model fixed. Experiments across four target systems report attack success rates of 79.4% on TangoFlux and 95.6% on Kling, with generated prompts transferring across open-source and closed-source models.

AI Brief

It presents a concrete red-teaming method with reported attack success rates and transfer findings relevant to text-to-audio model safety.

arXiv:2611.09351v1 Announce Type: new Abstract: Automated red-teaming of text-to-audio models requires test prompts that pass textual screening and expose harmful acoustic content. The central challenge is to connect the sounds produced by the target model to both reusable test generation and prompt revision. Our key idea is to represent test descriptions as structured acoustic plans that organize foreground sounds, context, and vocal attributes. To this end, we propose acoustic plan learning, a framework that learns and refines these plans using waveform feedback. We train the plan generator through supervised fine-tuning followed by reinforcement learning with audio feedback while keeping the target audio model fixed. At inference time, the same trained model generates initial plans and uses feedback to revise their context and vocal attributes while preserving the foreground. Experiments across four target systems show attack success rates of 79.4% on TangoFlux and 95.6% on Kling. The resulting test prompts also transfer across open-source and closed-source models, indicating shared acoustic vulnerabilities.
Be the first to rate this article.

Comments & Questions (0)

Captcha
Please be respectful — let's keep the conversation friendly.

No comments yet

Be the first to comment!