Abstract:Compressive sensing (CS) has been studied and applied in structural health monitoring for data acquisition and reconstruction, wireless data transmission, structural modal identification, and spare damage identification. The key issue in CS is finding the optimal solution for sparse optimization. In the past several years, many algorithms have been proposed in the field of applied mathematics. In this paper, we propose a machine-learning-based approach to solve the CS data-reconstruction problem. By treating a computation process as a data flow, the solving process of CS-based data reconstruction is formalized into a standard supervised-learning task. The prior knowledge, i.e., the basis matrix and the CS-sampled signals, are used as the input and the target of the network; the basis coefficient matrix is embedded as the parameters of a certain layer; and the objective function of conventional compressive sensing is set as the loss function of the network. Regularized by l1-norm, these basis coefficients are optimized to reduce the error between the original CS-sampled signals and the masked reconstructed signals with a common optimization algorithm. In addition, the proposed network is able to handle complex bases, such as a Fourier basis. Benefiting from the nature of a multi-neuron layer, multiple signal channels can be reconstructed simultaneously. Meanwhile, the disassembled use of a large-scale basis makes the method memory-efficient. A numerical example of multiple sinusoidal waves and an example of field-test wireless data from a suspension bridge are carried out to illustrate the data-reconstruction ability of the proposed approach. The results show that high reconstruction accuracy can be obtained by the machine learning-based approach. In addition, the parameters of the network have clear meanings; the inference of the mapping between input and output is fully transparent, making the CS data reconstruction neural network interpretable.

Compressive Sensing Reconstruction of Feed-Forward Connectivity in Pulse-Coupled Nonlinear Networks.

EiCSNet: Efficient Iterative Neural Network for Compressed Sensing Reconstruction

Reconstruction of sparse recurrent connectivity and inputs from the nonlinear dynamics of neuronal networks

THE ROLE OF SPARSITY IN INVERSE PROBLEMS FOR NETWORKS WITH NONLINEAR DYNAMICS

Compressive Sensing Inference of Neuronal Network Connectivity in Balanced Neuronal Dynamics

Network Dynamics for Optimal Compressive-Sensing Input-Signal Recovery.

Compressed Sensing for Network Reconstruction

Efficient Image Processing Via Compressive Sensing of Integrate-and-fire Neuronal Network Dynamics

Exact reconstruction of gene regulatory networks using compressive sensing

Cascaded Reconstruction Network for Compressive image sensing

Electrocardiogram Reconstruction Based on Compressed Sensing

Network reconstruction based on compressive sensing

Reconstructing Heterogeneous Networks via Compressive Sensing and Clustering

Detecting Synaptic Connections in Neural Systems Using Compressive Sensing

Causal and Structural Connectivity of Pulse-Coupled Nonlinear Networks

Generalized Sparse Recovery Model and Its Neural Dynamical Optimization Method for Compressed Sensing

Reconstruction of Complex Network Based on the Noise Via QR Decomposition and Compressed Sensing.

Global Reconstruction of Complex Network Topology Via Structured Compressive Sensing

2 Machine-learning approach for CS data reconstruction

Sparse Recovery Beyond Compressed Sensing: Separable Nonlinear Inverse Problems