Semidefinite embedding

Semidefinite embedding is a concept from mathematical optimization and, more specifically, from the field of semidefinite programming. It is used in various applications, including optimization, control theory, and machine learning. At a high level, a semidefinite embedding refers to a representation of certain types of problems or structures in a higher-dimensional space using semidefinite matrices. A semidefinite matrix is a symmetric matrix that has non-negative eigenvalues, which means it defines a convex cone.