Posted May 23, 2019
MIKA: Music Knowledge Intelligence Agent
Team Members and Titles
Summary

Machine learning systems coupling symbolic- and signal-based inputs have been almost exclusively used for speech recognition MIKA extends this work to more complex signals, testing on a historical corpus of jazz.

Key Findings

Developed a software architecture where multiple machine learning components work together to algorithmically compose new chord progressions based on an analysis of historical models.

Project Start Date
Project End Date

Publications

David Dahlbom and Jonas Braasch (2019): "Multiple f0 pitch estimation for musical applications using dynamic Bayesian networks and learned priors." The Journal of the Acoustical Society of America 145, 1814.

Jeremy Stewart, Matthew Goodheart, Curtis Bahn, Mary Simoni, Jonas Braasch (2019). "Music Intelligence and Knowledge Agent (MIKA)." Proceedings of the International Computer Music Conference.

Nathan Keil, David Dahlbom, Jeremy Stewart, Matthew Goodheart, Curtis Bahn, Mary Simoni, Michael Perrone, Jonas Braasch (2019). "Polyphonic pitch perception in rooms using deep learning networks with data rendered in auditory virtual environments." Proceedings of the Acoustical Society of America.