Columbia University computer scientists developed an algorithm that can block smart devices from spying on users by generating extremely quiet sounds.
Columbia's Carl Vondrick said the algorithm can block a rogue microphone from listening in 80% of the time by camouflaging a person's voice.
Columbia's Mia Chiquier explained that the program applies predictive attacks, signals that can disrupt any word automatic speech recognition models are taught to transcribe.
The system performs this task in near-real time by predicting an attack on the future of the signal, or word, based on two seconds of input speech.
The attack's volume closely resembles normal background noise, so people can converse naturally.
From Columbia Engineering News
View Full Article
Abstracts Copyright © 2022 SmithBucklin, Washington, DC, USA
No entries found