Universal Adversarial Examples In Speech Command Classification

Abstract

Adversarial examples are inputs intentionally perturbed with the aim of forcing a machine learning model to produce a wrong prediction, while the changes are not easily detectable by a human. Although this topic has been intensively studied in the image domain, classification tasks in the audio domain have received less attention. In this paper we address the existence of universal perturbations for speech command classification. We provide evidence that universal attacks can be generated for speech command classification tasks, which are able to generalize across different models to a significant extent. Additionally, a novel analytical framework is proposed for the evaluation of universal perturbations under different levels of universality, demonstrating that the feasibility of generating effective perturbations decreases as the universality level increases. Finally, we propose a more detailed and rigorous framework to measure the amount of distortion introduced by the perturbations

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keyvadillo2019universal

Related papers

• Universal Adversarial Perturbations For Speech Recognition Systems (2019)14.11
• Exploring Targeted Universal Adversarial Perturbations To End-to-end ASR Models (2021)8.35
• Universal Adversarial Perturbations Generative Network For Speaker Recognition (2020)12.33
• Inaudible Adversarial Perturbations For Targeted Attack In Speaker Recognition (2020)12.33
• Characterizing The Adversarial Vulnerability Of Speech Self-supervised Learning (2021)4.52
• Training Augmentation With Adversarial Examples For Robust Speech Recognition (2018)11.39
• Targeted Adversarial Examples For Black Box Audio Systems (2018)15.75
• Adversarial Machine Learning And Speech Emotion Recognition: Utilizing Generative Adversarial Networks For Robustness (2018)0.00