RESEARCH STAY AT UC3M: DEPARTMENT OF MECHANICAL ENGINEERING
Project: Statistical signal processing as a discipline has grown to a point where it can take on addressing very challenging problems.
With this work, we aim at exploring the population Monte Carlo (PMC) sampling methodology in solving problems that so far have been beyond the reach of existing methods. More specifically, our interest is in inference related to complex nonlinear models with large numbers of unknowns and where the underlying probability distributions may be non-Gaussian and multimodal.
The research will involve application of PMC sampling on graphical models for improved nonparametric belief propagation and for dealing with highly complex models, embedding PMC sampling within PF for improved PF performance, addressing global optimization with PMC sampling, conducting model selection with PMC sampling, studying theoretical issues including convergence of estimates obtained by PMC sampling, and application of PMC sampling to a case study of in vitro evolution of cancer stem cells.
Stay Period: SEP 11 - FEB 12