MS-22 Matrices, Tensors and Manifold Optimization (Tensors and Manifolds)

Organizers: Daniel Kressner (Ecole Polytechnique Fédérale de Lausanne, Switzerland) and Bart Vandereycken (University of Geneva, Switzerland)

Optimization on manifolds and tensor decompositions are popular research topics in numerical linear algebra. The purpose of this mini-symposium is to bring together recent advances in both fields that highlight geometrically inspired algorithms and analyses for rank-structured matrix problems. Example topics include global recovery guarantees for matrix completion and phase retrieval, optimization on bounded rank matrices, continuous low-rank decompositions, manifolds for shape analysis, and computational complexity of manifold optimization.

Speakers

Nicolas Boumal, Semidefinite Programs with a Dash of Smoothness: Why and When the Low-Rank Approach Works

Sutanoy Dasgupta, A Geometric Framework For Density Modeling

Alex Gorodetsky, Low-rank functional decompositions with applications to stochastic optimal control

Wen Huang, Intrinsic Representation of Tangent Vectors and Vector transport on Matrix Manifolds

Ivan Oseledets, Optimization over low-rank manifold: new results and applications

Ju Sun, When are nonconvex optimization problems not scary?

Ke Wei, Guarantees of Riemannian Optimization for Low Rank Matrix Reconstruction

Ke Ye, Tensor network ranks