Live Voice Assistant Inquiries

This is another dataset in Defined.ai's collection of “live” (unscripted or unsimulated) speech data. The dataset features English (US) and Portuguese (Portugal) speakers talking to their voice assistants. The recordings are made via far field devices. 

The inquiries are initiated by the users themselves and are not pre-scripted or simulated, and the data is cleansed of any personal identifiable information. Every inquiry has a precise, and valuable NLP annotations. Each inquiry has tags for its intent and entity. The data includes diverse accents and background noises.