In the scope of "AI for Science", solving inverse problems is a longstanding challenge in materials and drug discovery, where goal to determine hidden structures given set desirable properties. Deep generative models are recently proposed solve problems, but these currently struggling expensive forward operators, precisely localizing exact solutions fully exploring parameter spaces without miss...

We present a thermal velocity sampling method for calculating Doppler-broadened atomic spectra, which more efficiently reaches convergence than regular weighted sampling. The uses equal-population of the 1D distribution, “inverse transform” cumulative distribution function, and is broadly applicable to normal distributions. also discuss efficiency by eliminating classes, do not significantly co...

In this paper, we have established a new framework of truncated inverse sampling for estimating mean values of non-negative random variables such as binomial, Poisson, hypergeometrical, and bounded variables. We have derived explicit formulas and computational methods for designing sampling schemes to ensure prescribed levels of precision and confidence for point estimators. Moreover, we have d...

Bandpass sampling algorithm is effectively adopted to obtain the digital signal with significantly reduced sampling rate for a single radio frequency(RF) signal. In order to apply the concept to multiple RF signals, we propose bandpass sampling algorithms with the normal and the inverse placements since we are interested in uniform order of the spectrum in digital domain after bandpass sampling...

We discuss the effects of fast nuclear excitations on heavy-ion fusion reactions at energies near and below the Coulomb barrier. Using the fusion of two 40 Ca nuclei as an example and the inversion method, we show that the mass renormalization induced by fast nuclear excitations leads to a large surface diffuseness in the effective potential for heavy-ion fusion reactions.

In order to recover the movements of usually hidden articulators such as tongue or velum, we have developed a data-based speech inversion method. HMMs are trained, in a multistream framework, from two synchronous streams: articulatory movements measured by EMA, and MFCC + energy from the speech signal. A speech recognition procedure based on the acoustic part of the HMMs delivers the chain of p...