Privacy-Preserving Speech Representation Learning using Vector Quantization
This addresses privacy concerns for users of virtual assistants like Siri and Alexa, but it is incremental as it builds on existing vector quantization techniques.
The paper tackled the problem of speech signals containing sensitive speaker identity information by proposing a vector quantization method to produce anonymous representations, achieving a trade-off between speech recognition performance and privacy with configurable dictionary sizes.
With the popularity of virtual assistants (e.g., Siri, Alexa), the use of speech recognition is now becoming more and more widespread.However, speech signals contain a lot of sensitive information, such as the speaker's identity, which raises privacy concerns.The presented experiments show that the representations extracted by the deep layers of speech recognition networks contain speaker information.This paper aims to produce an anonymous representation while preserving speech recognition performance.To this end, we propose to use vector quantization to constrain the representation space and induce the network to suppress the speaker identity.The choice of the quantization dictionary size allows to configure the trade-off between utility (speech recognition) and privacy (speaker identity concealment).