State estimation for neurocritical care patients

Stuart Russell

UC Berkeley, Computer Science Divsion


Abstract

[joint work with Geoff Manley, Department of Neurological Surgery, UCSF, and Norm Aleks]

I will describe a new project, joint between UCB and UCSF, to investigate the application of probabilistic inference to interpret sensor data from patients in the neurotrauma intensive care unit. The project will severely test, and probably break, all known methods for probabilistic modelling, inference, and learning, and has led to a much broader research agenda in multi-scale, multi-purpose biosystems modelling. Needless to say, numerous PhD thesis topics are likely to emerge.