A music expert visual assistant

Modern music providers offer multiple services such as artist recommendations, browsing, playlist generation etc. Users interact with these services using graphical interfaces on computers and mobile devices. There are applications that need voice-only interactions such as smart watches or when the user’s hands and eyes are occupied. Virtual assistants, such as Siri by Apple and Alexa by Amazon, are commercial examples of voice driven assistants. The proposed research investigates the development of a virtual assistant that is designed for meaningful music service interactions. In addition to basic music playback commands the assistant will engage in music-specific dialogs supported by a knowledge base of music facts, automatic music content analysis, and effective speech recognition using domain-specific constraints. This work is an example of integrative artificial intelligence and a human-in-the loop system. Our goal is to achieve complex and meaningful interactions within the constrained but semantically rich domain of music listening.

Faculty Supervisor:

George Tzanetakis

Student:

Partner:

Ergonomyx

Discipline:

Computer science

Sector:

Manufacturing

University:

University of Victoria

Program: