Adam Coates et al. NIPS 2010: k-means for scalable learning of filter banks. Very efficient at tiling the space of possible inputs, and works very well with a sparse-coding-like activation function in SVM's! MATLAB code is available on Adam's website.
Test_l1_constraint_02 Basis Pursuit with Douglas Rachford By Gabriel Peyre Included in Toolbox Sparse Optmization Code covered by the BSD License View in: File Exchange Example Code Test for DR algorithm for L1 minimization (BP). We do here a compressed sensing resolution (random matrix