Biologically Inspired Artificial Intelligence

Screen Shot 2018-02-14 at 15.42.59.png

Dr. David M. Horowitz, CTO and my co-founder of My Voice AI (www.myvoice.ai), has led a distinguished career focussed on Biologically Inspired Artificial Intelligence. Biologically Inspired Artificial Intelligence is an AI whose algorithms are rooted in human physiological models of Sensory Perception and Cognition. For example, when tackling the problem of Voice Biometrics, Dr. Horowitz applies models that capture the form and function of human voice physiology. The principle is that nothing performs better at a sensory act like voice identification than the human speech communication chain where the speech signal is produced by the mouth and vocal tract and received by the auditory system or ear. Therefore, if one wants to build a system to identify a person by his or her voice quality, the system should be based on models of human voice physiology that is responsible for creating unique voice quality. Such models are called source-filter models of speech communication where the source is in the larynx, namely the glottis - two fleshy folds at the opening to the lungs that vibrates during air-flow from the lungs and gives speech its pitch and tone and this source is filtered or shaped by the vocal tract including the lips. It is the vocal tract that produces the individual sounds of speech or so-called phonemes and it is the voicing source or glottis that produces the voice quality that enables us to identify who the speaker is.

davidHpic.jpg

Dr. Horowitz crafts My Voice AI voice biometrics system called ACV (Acoustic Correlates of Voice Quality) with the source filter model of speech communication but also employs machine learning algorithms as well. Indeed, Biologically Inspired Artificial Intelligence does not reject machine learning but rather compliments it and ACV works with Biologically Inspired AI by complimenting machine learning algorithms. We first showed in late 2016 that adding ACV to a GMM increased accuracy of text independent voice identification by nearly 10%.

My Voice AI has taken this technology to incorporate it in an SDK and platform for rapid commercial distribution and integration. In doing so, My Voice AI is positioned to be the world leader in phrase (text) independent voice biometrics.

officepicarticle.jpg
officepicamsterdam.jpg
Nikola Andelic