Dr. Ashley R Keaton

Education

• University of Nevada, Reno – B.A., Spanish Language and Society

About

I am interested in what features of synthetic speech are most memorable or distinctive to a listener. My Master's Thesis investigated how phonetic information from synthesized speech is encoded into memory, and how this process differs from the encoding of natural speech.

My doctoral research investigated how listeners evaluate computer voices when they use sociolinguistic variation that is typically associated with human voices. I use the sociolinguistic (ING) variant (e.g., talking vs talkin') to test how listeners perceive synthetic voices that use nonstandard variants. In particular, I am interested in how listener's perceptions change if they believe the computer voice is a human speaker. 

I also collaborate on projects that use automatic speech recognition in applied contexts. In collaboration with the University of Iceland, I investigate "linguistic biomarkers" that indicate early signs of cognitive decline that may develop into dementia or Alzheimer's disease. At the University of Maryland, Baltimore County, I collaborate on a project that leverages sociolinguistic variation to create sophisticated machine learning models to identify voice deepfakes, voice clones, and spoofed audio. 

More broadly, I have an avid interest in cognitive science and love thinking about how our linguistic research provides insights into human cognition. 

Research Focus

Human-Computer Interaction; Speech Perception; Phonetics; Cognitive Science