Applied Mathematical Seminar28: Plate bending with isometry constraint: FEM vs machine learning

Seminar| Institute of Mathematical Sciences

Abstract:  We study a minimization problem that arises from a plate bending model with isometry constraints. Both finite element method and machine learning oriented method are exploited to investigate the convergence of the minimizer, and to what degree the isometry constraints are met by the proposed methods. We found that FEM may have certain a-priori theoretical convergence, while machine learning method seems much more efficient in most cases under study.