Adversarial Command Detection Using Parallel Speech Recognition Systems

  • Peng Cheng
  • , M. S. Arun Sankar
  • , Ibrahim Ethem Bagci
  • , Utz Roedig

Research output: Chapter in Book/Report/Conference proceedingsConference proceedingpeer-review

Abstract

Personal Voice Assistants (PVAs) such as Apple’s Siri, Amazon’s Alexa and Google Home are now commonplace. PVAs are susceptible to adversarial commands; an attacker is able to modify an audio signal such that humans do not notice this modification but the Speech Recognition (SR) will recognise a command of the attacker’s choice. In this paper we describe a defence method against such adversarial commands. By using a second SR in parallel to the main SR of the PVA it is possible to detect adversarial commands. It is difficult for an attacker to craft an adversarial command that is able to force two different SR into recognising the adversarial command while ensuring inaudibility. We demonstrate the feasibility of this defence mechanism for practical setups. For instance, our evaluation shows that such system can be tuned to detect 50% of adversarial commands while not impacting on normal PVA use.

Original languageEnglish
Title of host publicationComputer Security. ESORICS 2021 International Workshops - CyberICPS, SECPRE, ADIoT, SPOSE, CPS4CIP, and CDT and SECOMANE, 2021
EditorsSokratis Katsikas, Costas Lambrinoudakis, Nora Cuppens, John Mylopoulos, Christos Kalloniatis, Weizhi Meng, Steven Furnell, Frank Pallas, Jörg Pohle, M. Angela Sasse, Habtamu Abie, Silvio Ranise, Luca Verderame, Enrico Cambiaso, Jorge Maestre Vidal, Marco Antonio Sotelo Monge
PublisherSpringer Science and Business Media Deutschland GmbH
Pages238-255
Number of pages18
ISBN (Print)9783030954833
DOIs
Publication statusPublished - 2022
Event7th Workshop on the Security of Industrial Control Systems and of Cyber-Physical Systems, CyberICPS 2021, 5th International Workshop on Security and Privacy Requirements Engineering, SECPRE 2021, 4th International Workshop on Attacks and Defenses for Internet-of-Things, ADIoT 2021, 3rd Workshop on Security, Privacy, Organizations, and Systems Engineering, SPOSE 2021, 2nd Cyber-Physical Security for Critical Infrastructures Protection, CPS4CIP 2021 and 1st International Workshop on Cyber Defence Technologies and Secure Communications at the Network Edge, CDT and SECOMANE 2021 held in conjunction with 26th European Symposium on Research in Computer Security, ESORICS 2021 - Virtual, Online
Duration: 4 Oct 20218 Oct 2021

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13106 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference7th Workshop on the Security of Industrial Control Systems and of Cyber-Physical Systems, CyberICPS 2021, 5th International Workshop on Security and Privacy Requirements Engineering, SECPRE 2021, 4th International Workshop on Attacks and Defenses for Internet-of-Things, ADIoT 2021, 3rd Workshop on Security, Privacy, Organizations, and Systems Engineering, SPOSE 2021, 2nd Cyber-Physical Security for Critical Infrastructures Protection, CPS4CIP 2021 and 1st International Workshop on Cyber Defence Technologies and Secure Communications at the Network Edge, CDT and SECOMANE 2021 held in conjunction with 26th European Symposium on Research in Computer Security, ESORICS 2021
CityVirtual, Online
Period4/10/218/10/21

Fingerprint

Dive into the research topics of 'Adversarial Command Detection Using Parallel Speech Recognition Systems'. Together they form a unique fingerprint.

Cite this