Making the connection
As a concert-level pianist, an expert in judo, a motorcycle racer, a certified emergency medical technician, a mathematician, and a professor of computer science, Graeme Bailey is clearly a man of many applications. What is perhaps more remarkable, however, is the unifying enthusiasm he brings to problems that, on the surface, seem to have little to do with each other.
For instance, as he pursued certification as a paramedic, Bailey learned of uncertainties associated with caring for medical patients who have lung injuries. It would be of great clinical importance, he realized, if physicians had a good computer model of how human lungs function. The combination of Bailey's medical training and his formal education in the mathematical field of topology—in essence, the study of geometric figures under distortion—uniquely positions him to tackling this problem.
Alas, it is a non-trivial computational challenge to model an organ with some 300 million individual air sacs—alveoli—each exquisitely delicate. “It's very hard to do direct imaging to see what's going on,” he says. While photomicrographs from living animal studies can give some information, “the trouble is you're looking at a little patch, just on the surface, and you don't know if you're looking at the same alveoli, so getting factual information about what's happening there is difficult.”
In response, Bailey has based his model in mathematical formalisms that are clearly too simple to represent actual lung function, but begin to cause the same effects. From there he can compare the model's output with empirical data and refine it by adding other, real-life properties, such as a third dimension, limited mobility of the alveoli, and mutual pressure among the microsacs. “It's a standard reverse reasoning model,” he says. “where you say, ‘I've created something that behaves the same way as a duck, and it quacks, but unfortunately that doesn't guarantee that it's a duck.'”
Without this guarantee, work must continue., and so far Bailey has helped to confirm a so-called “recuitment model” of lung function, where the alveoli do not all inflate at the same time, but evoke more capacity as needed. All this would seem remote from Bailey's other major interest—to devise an artificial accompanist for vocalists and other instrumentalists—but he sees a common theme in the economy of information. Both in modelling the lungs and modelling how accompanists exchange cues with their living partners, the problem is in “trying to understand the information. … In both cases we begin to form a hierarchy of information, and then decide what pieces are needed” to accomplish the task. In this work he is collaborating with Carol Krumhansl, a violinist and Cornell Professor of Psychology, and with composer and PhD student Spencer Topel.
Bailey's articulateness and enthusiasm have translated into success in the classroom. His multiple teaching awards, including the Kenneth A. Goldman '71 Excellence in Teaching Award, and the Kendall S. Carpenter Memorial Advising Award, are clear evidence of his effectiveness in challenging his students. They, in turn, provide Bailey with ongoing opportunities to bend and stretch what is clearly a versatile talent.