Github Browse our lab repository here




NuMax Learning near-isometric embeddings using semi-definite programming

A linear dimensionality reduction technique that is near-isometric, i.e., preserves pairwise distances on a dataset. The codebase has an efficient implementation that uses a greedy procedure to obtain a superset of the active constraints. It is also possible to incorporate classification promoting objectives.

(code)   (paper)

CS-MUVI Video compressive sensing using motion-flow models
(code)   (paper)

SpaRCS Compressive recovery of low rank + sparse matrices

A greedy algorithm for compressive recovery of a matrix modeled as a sum of low-rank and sparse matrices. This approach uses a variant of CoSamp and Admira to achieve this.

(code)   (paper)

CS-LDS Linear dynamical system parameters from compressive measurements
(code)   (paper)