Explore the vulnerabilities in Automatic Speech Recognition (ASR) and Speaker Identification Systems in this 15-minute IEEE conference presentation. Gain insights into various attack methods, evaluation metrics, and mediums used to exploit these systems. Delve into the concept of knowledge transferability and examine real-world experiments that demonstrate the potential risks. Learn why a Systematization of Knowledge (SoK) approach is crucial for understanding and addressing these security challenges in voice-based technologies.
SoK - The Faults in Our ASRs - An Overview of Attacks Against Automatic Speech Recognition and Speaker Identification Systems
Overview
Syllabus
Introduction
Attack
Why SoK
Overview
Attacks
Metrics
Mediums
Knowledge
Target Transferability
Experiments
Taught by
IEEE Symposium on Security and Privacy
Tags
Related Courses
-
Hear "No Evil", See "Kenansville" - Efficient and Transferable Black-Box Attacks on Automatic Speech Recognition Systems
-
TrojanModel - A Practical Trojan Attack against Automatic Speech Recognition Systems
-
SoK - Taxonomy of Attacks on Open-Source Software Supply Chains
-
Audio Adversarial Examples - Targeted Attacks on Speech-to-Text
-
Targeted Adversarial Examples for Black Box Audio Systems
-
ImU - Physical Impersonating Attack for Face Recognition System with Natural Style Changes
