Locality Regularized Reconstruction: Structured Sparsity and Delaunay Triangulations

Linear representation learning is widely studied due to its conceptual simplicity and empirical utility in tasks such as compression, classification, and feature extraction. Given a set of points and a vector , the goal is to find coefficients so that , subject to some desired structure on . In this work we seek that forms a local reconstruction of by solving a regularized least squares regression problem. We obtain local solutions through a locality function that promotes the use of columns of that are close to when used as a regularization term. We prove that, for all levels of regularization and under a mild condition that the columns of have a unique Delaunay triangulation, the optimal coefficients' number of non-zero entries is upper bounded by , thereby providing local sparse solutions when . Under the same condition we also show that for any contained in the convex hull of there exists a regime of regularization parameter such that the optimal coefficients are supported on the vertices of the Delaunay simplex containing . This provides an interpretation of the sparsity as having structure obtained implicitly from the Delaunay triangulation of . We demonstrate that our locality regularized problem can be solved in comparable time to other methods that identify the containing Delaunay simplex.
View on arXiv