Random Objects and Metric Statistics for Connectomics

A workshop in IMSI's Connectomics program on the statistical analysis of non-Euclidean data, where random objects are random variables and data samples taking values in general metric spaces, with emphasis on applications to brain connectomics. It covers theory, methodology and applications for samples of networks, distributions, compositional data, covariance matrices and surfaces, tree data, data on manifolds, and high-dimensional and functional data viewed as elements of a metric space. Topics include means, barycenters and variance, conditional Fréchet means and Fréchet regression, and extensions towards metric principal component analysis, uncertainty quantification, deep learning, causal inference and longitudinal and time-series methods. The programme includes lightning talks and a poster session for early-career researchers. All registrants, online and in person, receive Zoom links and can attend online. In-person seats are limited and in-person registrants must wait for an invitation from IMSI before travelling; invited in-person participants pay a non-refundable fee of $25 (students) or $50 (non-students).