My research interests centre on the modeling of symbolic processes in musical improvisation: on how the choices a soloist makes in real time can be described, formalised, and perhaps anticipated.
Most of my work sits where computational musicology meets human–machine co-improvisation. I read improvised lines through statistical and information-theoretic analysis, looking for the regularities that sit beneath an apparently free performance: which gestures recur, how much a phrase tells us about the one that follows, and where a player departs from their own habits. The same descriptions that explain what a soloist did can also tell a machine how to take part in the conversation rather than merely accompany it.
Underneath all of this lies a more practical concern, the construction of aligned symbolic corpora. Serious claims about improvisation depend on transcriptions whose notes, harmony, and form are aligned closely enough to be compared across takes, across tunes, and across players, so a good share of my time goes into building and curating that material.