The talk will focus on efficient data-driven modeling associated with the inverse problem and feature extraction. We show how the theories of 1-manifolds and F-transforms contribute to these delineated areas.