Optimized Reconstruction Parameters for Noise Modeling in Multi-Task Bayesian Compressed Sensing for Sparse 2D Spectroscopy

Keywords

accommodates added aliased aliasing appears application approximate assume assuming basis best blue brain choice close complex compressed concentrates conditioned conditions considered considering corresponds coupled coupling dashed decoupling defined detect determined diagonal dimension discrepancy distribution division draw dropouts encoding equally evaluated evaluating evaluation even expense experiment exploration expression extend feature flexible focusing formed free frequency full fully function gamma glutamate hand health ideal ideally imperfect implemented improved include included incorporating indicate institute insufficient interfering introduced inverse lead leads likelihood limiting magnitude matched maximization mayer medical metabolite metabolites minimize model modeling models molecules naturally noise noising optimized overlap patterns peaks perfectly performance placed posterior postulate precision previous previously prior priori priors promising proposed random readily realizations reconstructed reconstruction reconstructions recovery represent respectively sampled sampling scanning sciences sensing seven simulated solutions solved sparse spectra spectral spectroscopy spectrum step successful successfully suggested summed supported taken task technology trans trials uncertain uniform uniformly upon variance zero