Securing Voice Authentication Applications Against Targeted Data Poisoning

TL;DR

This paper presents a robust voice authentication framework resilient to targeted data poisoning attacks, maintaining accuracy even with minimal poisoned data, by using realistic datasets and attack scenarios.

Contribution

It introduces an enhanced voice authentication method that effectively counters targeted data poisoning attacks with limited poisoned data.

Findings

Robust authentication with only 5% poisoned data

Effective in realistic attack scenarios

Maintains high accuracy under attack

Abstract

Deep neural network-based voice authentication systems are promising biometric verification techniques that uniquely identify biological characteristics to verify a user. However, they are particularly susceptible to targeted data poisoning attacks, where attackers replace legitimate users' utterances with their own. We propose an enhanced framework using realworld datasets considering realistic attack scenarios. The results show that the proposed approach is robust, providing accurate authentications even when only a small fraction (5% of the dataset) is poisoned.