Geometry of Nonnegative Rank

ended

This proposal is centered on nonnegative rank - a notion that modifies the definition of matrix rank. This project is basic research aimed to better understand the geometry of matrices and tensors of nonnegative rank at most r, how this knowledge can be employed to improve exact and numerical nonnegative matrix algorithms, obtain an algorithm for nested polytopes in dimension three, and address questions about mixture models in statistics. Computing and understanding nonnegative rank is truly interdisciplinary. Completing this project requires techniques from polyhedral geometry, real algebraic geometry, optimization, statistics, symbolic algebra, and numerics.