Yoshinori Nakanishi-Ohno

Sparse approximation based on a random overcomplete basis

We discuss a strategy of sparse approximation that is based on the use of an overcomplete basis, and evaluate its performance when a random matrix is used as this basis. A small combination of basis vectors is chosen from a given overcomplete basis, according to a given compression rate, such that they compactly represent the target data with as small a distortion as possible. As a selection method, we study the

Distribution estimation of hyperparameters in Markov random field models

\ell_0

Bayesian inference of metal oxide ultrathin film structure based on crystal truncation rod measurements

- and

Theory of Distribution Estimation of Hyperparameters in Markov Random Field Models

\ell_1

Statistical mechanical analysis of sparse linear regression as a variable selection problem

-based methods, which employ the exhaustive search and

Compressed Sensing in Scanning Tunneling Microscopy/Spectroscopy for Observation of Quasi-Particle Interference

\ell_1

Three levels of data-driven science

-norm regularization techniques, respectively. The performance is assessed in terms of the trade-off relation between the representation distortion and the compression rate. First, we evaluate the performance analytically in the case that the methods are carried out ideally, using methods of statistical mechanics. Our result clarifies the fact that the

Spatial coherence of the insulating phase in quasi-two-dimensional LaNiO3 films

\ell_0

ES-DoS: Exhaustive search and density-of-states estimation as a general framework for sparse variable selection

-based method greatly outperforms the

Exact Calculation of Typical Hyper-Parameter Posterior Distribution of Gaussian Markov Random Field model.

\ell_1