AI

At MIT, on a Bell Labs fellowship with Prof Marvin Minsky, Levitt pioneered the field of algorithmic music generation for traditional classical and jazz styles. His AI description languages ranged across rhythmic concepts like syncopation, swing, and waltz, melody and counterpoint terms, and the harmonic cadences and progressions common to so many styles.

 

Levitt and Stephan Schwanauer co-edited the MIT Press book Machine Models of Music, a collection of AI papers about software that shows how we perceive and create music, including Levitt’s thesis A Representation of Musical Dialects.