Prof Narayanan's Research Teaches AI To Find Anyone's Emotional State By Voice

Prof Narayanan’s Research Teaches AI To Find Anyone’s Emotional State By Voice

Human voice is a rich medium with which we not only communicate our wants and needs and intent but it’s also a very special way of expressing our emotions and identity in emerging times where AI is increasingly becoming part of our lives, people want to use that to interact with other digital systems not only for transactional purposes, such as asking about weather, but more so as a social companion for elderly Human speech is rather special as it carries information about our intent, emotions, identity and several other data, like our health state A true AI system, therefore, should not only recognise the intonational properties of human speech, the words we speak and the way we interact but also take into account the nonverbal cues It’s almost like a painting, where we mix different basic colours to create a landscape with different possibilities.

Shrikanth Narayanan is a Professor at the University of Southern California and an interdisciplinary engineer-scientist with a focus on human-centered signal processing and machine intelligence as well as informatics with speech and spoken language processing at its core.

A prolific award-winning researcher, educator and inventor, with hundreds of publications to his credit, his work translates to using speech and audio to identify mental health and wellness issues, analyzing the health and stress level of workers and developing AI tools for understanding how stories are told in film and TV from a social lens. Flickr So sit back, plug in your headphones, tune in to your favourite lo-fi channel and read on this interesting conversation that we had with Professor Narayanan.

What do we understand about voice as a medium of communication in the age of AI?

Human voice is a rich medium with which we not only communicate our wants and needs and intent but it’s also a very special way of expressing our emotions and identity. And spoken language in humans is particularly remarkable. It allows us to easily communicate all our thoughts, ideas and desires through voice; and in emerging times where AI is increasingly becoming part of our lives, people want to use that to interact with other digital systems not only for transactional purposes, such as asking about weather, but more so as a social companion for elderly and learning systems for children. Representative image While today’s AI like Echo, Siri and Google Assistant does an outstanding job of word recognition and analysis, its dependence on speech alone is an inherent limitation–it, kind of, seems mechanical. When can we have an AI system that is truly capable of sensing and reacting to a user’s emotions?

Human speech is rather special as it carries information about our intent, emotions, identity and several […]

Continue reading the rest at www.indiatimes.com