BriefGPT.xyz
Sep, 2024
重新评估对抗性语音中的噪声增强方法
Reassessing Noise Augmentation Methods in the Context of Adversarial Speech
HTML
PDF
Karla Pizzi, Matías P. Pizarro B, Asja Fischer
TL;DR
本研究探讨了噪声增强训练是否能够同时提高自动语音识别（ASR）系统的对抗鲁棒性。通过对四种最先进的ASR架构在不同增强条件下的对抗鲁棒性进行比较分析，结果表明，噪声增强不仅改善了模型在噪声语音上的表现，还提升了其抵御对抗攻击的能力。
Abstract
In this study, we investigate if noise-augmented training can concurrently improve
Adversarial Robustness
in
Automatic Speech Recognition
(ASR) systems. We conduct a
→