Adjust-free adversarial example generation in speech recognition using evolutionary multi-objective optimization under black-box condition

TL;DR

This paper introduces a black-box adversarial attack method for speech recognition that uses evolutionary multi-objective optimization to generate robust, adjust-free adversarial examples unaffected by timing lag.

Contribution

It presents a novel EMO-based approach for creating robust adversarial speech examples that do not require timing adjustments, improving attack robustness under black-box conditions.

Findings

Successfully generated adjust-free adversarial examples.

Examples are robust against timing lag.

Method outperforms previous approaches in robustness.

Abstract

This paper proposes a black-box adversarial attack method to automatic speech recognition systems. Some studies have attempted to attack neural networks for speech recognition; however, these methods did not consider the robustness of generated adversarial examples against timing lag with a target speech. The proposed method in this paper adopts Evolutionary Multi-objective Optimization (EMO)that allows it generating robust adversarial examples under black-box scenario. Experimental results showed that the proposed method successfully generated adjust-free adversarial examples, which are sufficiently robust against timing lag so that an attacker does not need to take the timing of playing it against the target speech.