Mining voice for health indicators
Dale Joachim
(SondeHealth)
Abstract:
The cost of healthcare reported around 20% of GDP in the US incentives institutions in the health care eco system to look for new ways to decrease cost through more affordable health monitoring, early detection and intervention methods. Speech research has for many decades shown evidence of health indicators in human voice. Recent advancements in AI and speech processing coupled with large scale speech and health data collections have accelerated the mining of health indicators in speech, while the ubiquitous use of smartphones has provided the necessary platform for large-scale deployment of speech health applications. This talk will provide health and economic motivations for mining speech for health indicators, then will focus on methods for measuring speech (speech features), methods for mining speech, recent advancements and forecast.
Biographical Sketch:
D. Joachim is the VP of Speech and Data Science at SondeHealth, a startup company leveraging vocal biomarkers for screening various health conditions including mental health. Prior to SondeHealth, Joachim held a number of research positions, as well as faculty, visiting faculty and adjunct faculty at Tulane University, MIT and Northeastern University. Joachim holds a PhD in Electrical Engineering.
Date and time:
Saturday, October 29, 2022
12:45 PM—2:30 PM
Zoom Meeting Link:
https://umassboston.zoom.us/j/92555836040