Assistant Professor of Linguistics Betsy Sneller was a faculty sponsor of a Computer Science and Engineering Capstone Experience this spring. In partnership with Russ Werner, Sneller supervised a group of undergraduate students in Computer Science in the development of an on-premises automated speech recognition (ASR) system.
The team, named MSU Linguistics, included members Jacob Caurdy, Yichen Ding, Maria Irimie, Kyle Reinhart, and Eden Seo. Team members developed an open-source all-in-one speech-to-text software package that creates time-aligned transcripts from audio files of speech. This system will be beneficial to the research done by the MI Diaries project, which collects audio diaries from Michiganders, but will also be beneficial to any researchers working with long-form audio data that needs to be transcribed.
The team’s software is easily re-trainable, meaning that any researcher will be able to fine-tune the ASR to better fit their own language data. Best of all, since the software runs locally and will be available open-source, researchers will be able to save costs associated with using existing ASR services like Google.
CSE 498 capstone teams present their projects on Design Day in a variety of ways. Teams create and set up an exhibit where they demonstrate their projects and answer questions from Design Day attendees, which include corporate clients, families, friends, faculty, and students, along with hundreds of local middle and high school students. In addition, as part of the award competition, teams present their projects to a panel of judges.
Team MSU Linguistics came away with the Exposition Award for the best overall Design Day performance, for their impressive in-person exhibit of their project. Their exhibit included a hands-on portion, for visitors to test out how their ASR software works in real time. One visitor tried to stymie the software by reciting the Jabberwocky, and it still transcribed the speech pretty closely!
Congratulations to Team MSU Linguistics for their win!
