- Research Program Mentor
MS Master of Science candidate
Natural Language Processing, Computer Vision, Deep Learning
Explorations of gendered voice and its ethics
As voice assistants (VAs) permeate throughout our society, more and more people are bringing Alexa or Siri or their favorite VAs home. Currently, there is research being done on affect prediction: eg if someone responds to the VA angrily, should the VA assume it was wrong? A potential project in this area could look at whether a person's place on the gender spectrum affects the way they interact with voice assistants. One of the biggest issues with technology is that it is being trained to support a single type of person- unfortunately, different groups of people respond quite differently to indicate the same thing. If the student wanted to look more closely at gendered expressions of emotion, they could look at audio representations and how emotion can be isolated from them. A student could further explore whether it is ethical for technology companies to vary interactions based on predicted gender of user. There are many ethical questions that lie in this space.