CASE STUDIES

Industrial and Sociaetal applications of Machine Learning for Music by the UnSupervised Group

  • Music Biometrics in Education to accelerate human learning

    This case study by UnSupervised proposes realtime biometric data, Machine Learning and Interactive music practice using wearable sensors as a new route for musical education. Let us say a music student meets their teacher once per week. How do they practice at home on their own? This tool can become a smart tool to monitor musical exercises set by the teacher, adding ultimate rigour to pattern matching and modes of musical expression.

    Techncial Paper

  • Sonifying Machine Learning Algorithms for Economic Decisions

    This case study proposes SonOpt (open source) as a data sonification application for monitoring the progress of Sonifying Bi-objective Population-Based Optimization Algorithms during search, to facilitate algorithm understanding. Economic decisions can now be better understood thanks to sonification processes of the ML algorithm (not only visualisation)

    Technical Paper

  • A.I. Beatboxer by Keep•It•Human - A gamified tool for Diversity and Inclusion

    We presented our A.I. Beatboxer to ca. 22.000 visitors during the Manchester Science Festival 2022. Visitors could have a conversation with an intelligent mind embodied as a beatboxer.
    A key conclusion was that virtual embodiment of different beatboxing personalities provides new routes for experiencing Diversity and Inclusion in a safe, positive, and nurturing manner.
    Users (mostly very young people) select their hilarious beatboxing avatar and start making funny noises or talk to the microphone. The AI converts any voice input into professional beatboxing music in real time and replies back with a similar pattern. Then the big laughter! Parents automatically pickup their phones to start filming!

    Videogames for Good