Exploiting Neural Audio Codec Latents for Adversarial Audio Attacks
利用神经音频编解码器潜变量进行对抗性音频攻击
机构 * Dept. of Computer Science and Engineering, University of North Texas(北德克萨斯大学计算机科学与工程系)
专题命中 扩散模型 :diffusion(abstract)
AI总结 提出在神经音频编解码器连续潜空间中生成对抗性扰动的方法,实现单次前向传播、低延迟(<7ms)的高成功率(99%)目标攻击。
Comments Accepted to Interspeech 2026. 5 Pages with references containing 2 figures and 4 tables. Code is available at https://github.com/vcbsl/DAC-GAN or https://github.com/Bharath-K3/DAC-GAN