1400-1500: Sparse graphical models in genomics: an application to censored qPCR data
Speaker: Veronica Vinciotti, Reader in Statistics, College of Engineering, Brunel University London, UK.
Abstract
Regularized inference of networks using graphical modelling approaches has seen many applications in biology, most notably in the recovery of regulatory networks from high-dimensional gene expression data. Various extensions to the standard graphical lasso approach have been proposed, such as dynamic and hierarchical graphical models. In this talk, I will focus on a latest extension to censored graphical models in order to deal with censored data such as qPCR data. We propose a computationally efficient EM-like algorithm for the estimation of the conditional independence graph and thus the recovery of the underlying regulatory network.
1500-1600: Probabilistic modelling with the experts
Speaker: Sami Kaski, Professor, Department of Computer Science, Aalto University, Finland.
Abstract
I will discuss multiple-data-source prediction problems typical of omics-based precision medicine. What is less typical is that some of the data sources are expert users, whose time is costly, changing the problem to active learning or experimental design for prediction. We have addressed this setup as a probabilistic modelling problem, where different types of sources need different modelling assumptions. I will demonstrate that promising results can be achieved in treatment effectiveness prediction tasks in restricted settings, even by explaining human variation with noise models. Richer behaviour requires richer models that draw from cognitive science.