Summary
In collaboration with the UK Nunn Center for Oral History and Dr. Doug Boyd, we’ve developed a system known as SpeakEZ. There is an archive with a backlog of thousands of hours of interviews that need to have time-stamped, speaker-labeled, accurate transcriptions. This process used to be done manually, but it would take hours. SpeakEZ is a one-stop shop for generating transcriptions and speaker labeling, outputting one merged time-stamped speaker-labeled transcript.
In addition, the system can extract themes from the oral history interviews, identify places/names mentioned during the interview, and flag risqué topics that might need to be censored before an interview is released publicly.
Access
Email ai@uky.edu for more information.
Ownership
Currently, SpeakEZ is in its prototype stage. SpeakEZ is a tenant of CAT-Talk.
Resources
Vaiden Logan was funded 50% on this project.