Georgia Tech Music Intelligence Group (GTMIG)

We are interested in creating systems that can understand musical sound. Our guiding principles are that:

1) Machine understanding of music will serve as a foundation for applications that help us to enjoy music. For example:

• helping us find new music
• creating automatic accompanists
• developing more intuitive compositional tools

Technology will play a role in making more people musicians as well as allowing musicians to more fully express themselves.

2) Understanding human cognition of musical sound will help us to develop these technologies and provides an excellent window into human cognition more generally.

• By understanding our complex multi-layered response to music we will be able to start to see the connection between abstract sound patterns and the profoundly moving experiences that music can engender.
• Because music involves so many aspects of cognition, from complex auditory pattern recognition, to attention, memory and emotion, its understanding provides a window into the mind.

3) Modeling music will help us to understand human creativity and lead to new creative partnerships between musicians as well as between musicians and technology. In our group, we are particularly focused on realtime situations such as musical improvisation.

The tools we use to address these problems come from many disciplines. We often start at the signal level and attempt to extract information that can then be used to drive generative or decision models. For these tasks we use techniques from digital signal processing and machine learning. Many of our working principles and musical models come from the insights of music theory or from our own intuitive understanding as practicing musicians. In many cases, musical insights suggest computational approaches. Our cognitive studies draw upon behavioral and imaging methods drawn from psychology as well as theoretical principles from artificial intelligence.

Whenever possible we seek to showcase our work in performances and in functional applications. This allows us to evaluate research that is often subjective and cannot easily be compared with and established ground-truth, such as when creating an improvising system or an application for music recommendation.

Below we describe some of our current and proposed research. A full list of publications and accompanying presentations is below.