Title: “Explainable Artificial Intelligence (XAI): Applications in Voice, Speech Science and Education”
Speaker: Mark Berardi, PhD, Department of Communication Sciences and Disorders, University of Iowa
Artificial intelligence is rapidly transforming speech-language pathology research and practice, yet the "black box" nature of many AI systems limits their practical use and educational value. This presentation demonstrates how explainable artificial intelligence (XAI) techniques enhance transparency and real-world application across voice and speech research and classroom experiences. Through voice fatigue studies, I show how interpretable machine learning reveals distinct response patterns among speakers, enabling personalized risk assessment for teachers and other occupational voice users. In psychiatric applications, XAI analysis demonstrates that speech science expertise significantly improves AI model performance, with clinical voice measures enhancing generic acoustic features. Additionally, an AI literacy project shows how students can develop critical skills for evaluating AI technologies through hands-on learning activities. This work establishes XAI as both a research tool and educational approach, ensuring AI advances remain understandable and applicable while supporting personalized medicine in communication disorders.
The seminars take place every Friday from noon to 1 p.m. on Zoom.
Zoom link for online seminars (used throughout the semester): https://uiowa.zoom.us/j/97359327852?pwd=0HDya18oxt4Yg8FiBr4WAn0F5jnzMw.1
Meeting ID: 973 5932 7852 Passcode: 219629