Applying Generative Models to find Critical Boundaries for Alzheimer?s Patients Lead Investigator: Daniel de Marchi Institution : University of North Carolina - Chapel Hill E-Mail : demarch@live.unc.edu Proposal ID : 1254 Proposal Description: We will apply a generative model to forecast the cognitive decline of Alzheimer's patients. The model's initial conditions will be perturbed so the transition points, where the model's forecast is most unstable, can be identified. This will allow prioritization of the most critical patients, as well as accurate projections of cognitive decline.