The third meeting of CRN Seminar will take place on Thursday, December 19, 2024, at 14:15–17:45 in room A.4.1 in building C-19. Agenda includes three lectures:

14:15–15:15
Integration meeting

15:15–15:55
Prof. Jarosław Harężlak (Indiana University)
Statistical regularization methods with applications in brain imaging

Abstract:

Regularization methods play a crucial role in the analysis of brain imaging data, where the number of observations is frequently much smaller than the number of covariates. This is even more crucial in multi-modal imaging, where combining data from different sources (e.g., sMRI, fMRI, and dMRI) can enhance insights into brain structure and function. In these settings, regularization helps to address the challenges posed by high-dimensional, noisy data by imposing constraints that promote stability and interpretability in model estimation. Off-the-shelf techniques such as Lasso, ridge regression, and elastic net are commonly employed to control overfitting and improve prediction accuracy, while spatial and informed regularization methods can leverage the inherent structure of imaging data, allowing for the integration of multiple imaging modalities. We describe the common regularization methods, and their extensions developed by us, including PEER (partially empirical eigenvectors for regression) and SpINNEr (Sparsity Inducing Nuclear-Norm Estimator).

16:05–16:45
Prof. Adam Nowak (Polish Academy of Sciences)
On the heat equation on a sphere and on other manifolds (in Polish)

16:55–17:35
Prof. Bartosz Trojan (Wrocław University of Science and Technology)
Compactifications of affine buildings (in Polish)