Adversarial Music: Real World Audio Adversary Against Wake-word Detection System

TL;DR

This paper demonstrates a real-world audio adversarial attack using inconspicuous background music to deactivate commercial voice assistants like Alexa by significantly reducing wake-word detection accuracy.

Contribution

It introduces the first real-world adversarial attack on a commercial-grade wake-word detection system using an emulated model and over-the-air testing with music as an adversarial signal.

Findings

Effective reduction of wake-word detection F1 score from over 93% to 11%

Successful over-the-air attack against Alexa, reducing F1 score from 92.5% to 11%

Non-adversarial music does not disable Alexa as effectively

Abstract

Voice Assistants (VAs) such as Amazon Alexa or Google Assistant rely on wake-word detection to respond to people's commands, which could potentially be vulnerable to audio adversarial examples. In this work, we target our attack on the wake-word detection system, jamming the model with some inconspicuous background music to deactivate the VAs while our audio adversary is present. We implemented an emulated wake-word detection system of Amazon Alexa based on recent publications. We validated our models against the real Alexa in terms of wake-word detection accuracy. Then we computed our audio adversaries with consideration of expectation over transform and we implemented our audio adversary with a differentiable synthesizer. Next, we verified our audio adversaries digitally on hundreds of samples of utterances collected from the real world. Our experiments show that we can effectively reduce the recognition F1 score of our emulated model from 93.4% to 11.0%. Finally, we tested our audio adversary over the air, and verified it works effectively against Alexa, reducing its F1 score from 92.5% to 11.0%.; We also verified that non-adversarial music does not disable Alexa as effectively as our music at the same sound level. To the best of our knowledge, this is the first real-world adversarial attack against a commercial-grade VA wake-word detection system. Our code and demo videos can be accessed at \url{https://www.junchengbillyli.com/AdversarialMusic}