The main goal is to develop new fundamental approaches to global optimization and learning systems with focus on artificial neural networks, modern methods of multiscale modeling and neuro-inspired computations based on systems of adaptive nonlinear differential equations. The project consists of several coherent stages with a focus on the research and development of collaborative neurodynamic approaches to optimization problems in supervised/unsupervised learning with applications to robotics, robust control, feature extraction tasks, as well as surrogate models, inverse problems, deformable shape modeling and optimization.