March 29, 2023, 1:10 a.m. | Haodong Zhao, Wei Du, Junjie Guo, Gongshen Liu

cs.CR updates on arXiv.org arxiv.org

Speaker verification has been widely used in many authentication scenarios.
However, training models for speaker verification requires large amounts of
data and computing power, so users often use untrustworthy third-party data or
deploy third-party models directly, which may create security risks. In this
paper, we propose a backdoor attack for the above scenario. Specifically, for
the Siamese network in the speaker verification system, we try to implant a
universal identity in the model that can simulate any enrolled speaker and …

attack authentication backdoor computing data identity large may network party power risks scenario security security risks system third third-party training verification

SOC 2 Manager, Audit and Certification

@ Deloitte | US and CA Multiple Locations

Director, Threat and Attack Research

@ Singtel | Macquarie Park, Australia

Manager Information Security

@ Diebold Nixdorf | Remote, United States

Senior Analyst, IT Information Security

@ IHG | GA, United States

Eurizon Capital SGR - Compliance Senior Specialist

@ Intesa Sanpaolo | Milano, IT

Tier 1 Fusion Security Analyst

@ Nielsen | Bengaluru, India