Abstract

Reverse Engineering Sparse Network Reconstruction via Compressed Sensing: A Systems Identification Approach

Andrecut, M. and Madni, A.M.

Abstract

Compressed sensing enables the recovery of a sparse signal from a small number of projections onto random vectors. Here, we propose a new extension of compressed sensing to the sparse network reconstruction problem. This is a reverse engineering problem that requires the identification of connections between the elements of a sparse network, given a set of random input-output signals, processed by the network. This problem is of interest in many fields such as genetic regulatory networks, neural networks, and social networks, where reverse engineering plays an important role. An application to the sparse reconstruction of gene networks is also presented.

From:

Andrecut, M. and Madni, A.M., Conference on System Engineering Research (CSER), 2008.