Ayoub Belhadji

DPPKQ is a Python library for reproducing the figures of the article Kernel quadrature with DPPs . The article compares kernel quadrature on nodes that follows a certain projection DPP to other distributions such as ridge leverage score distribution introduced in the article On the Equivalence between Kernel Quadrature Rules, uniform grid, sequential Bayesian quadrature and many more...

FrameBuilder is a Python library for random finite frames with prescribed lengths and spectrum. The algorithm is based on a parametrization of the set of finite frames with prescribed lengths and spectrum (the restricted Gelfand-Tsetlin polytope) given in the article Constructing finite frames of a given spectrum and set of lengths. The algorithm outputs finite frames without further structure as it is the case with the classic algorithm based on the application of Given's rotations proposed in the article Generalized Finite Algorithms for Constructing Hermitian Matrices with Prescribed Diagonal and Spectrum. This randomized algorithm was used in the article A determinantal point process for column subset selection to generate toy datasets.

A Python library for low rank approximations using column subset selection algorithms. It compares several algorithms: projection DPPs, volume sampling, double phase, determinisitic selection based on leverage scores and pivoted QR decomposition.