Bayesian Deconvolution Research Articles

Transient temperature measurement of silicon carbide Schottky barrier diode based on thermal reflection.

In this article, we present a transient temperature detection device for silicon carbide (SiC) Schottky barrier diodes (SBDs) based on thermal reflection theory. We constructed a thermal reflection temperature measurement device based on a 530-nm green laser. This device is more suitable for transient temperature measurement of SiC SBDs than previous thermal reflection equipment. The accuracy of temperature measurement by our device was confirmed by comparison with the results of infrared thermal imaging. The high temporal resolution characteristics of the thermal reflection technology allowed the detection of millisecond-level transient temperature changes in SiC SBDs. In addition, we investigated the complementarity of transient temperature change curves during heating and cooling processes, as well as the reasons for the differences between these curves. Finally, we used the structural function method combined with the Bayesian deconvolution algorithm to obtain the thermal resistance along the heat flow path of the device and validated the results using an established thermal resistance testing method.

Single-cell Bayesian deconvolution

Individual cells exhibit substantial heterogeneity in protein abundance and activity, which is frequently reflected in broad distributions of fluorescently labeled reporters. Since all cellular components are intrinsically fluorescent to some extent, the observed distributions contain background noise that masks the natural heterogeneity of cellular populations. This limits our ability to characterize cell-fate decision processes that are key for development, immune response, tissue homeostasis, and many other biological functions. It is therefore important to separate the contributions from signal and noise in single-cell measurements. Addressing this issue rigorously requires deconvolving the noise distribution from the signal, but approaches in that direction are still limited. Here, we present a non-parametric Bayesian formalism that performs such a deconvolution efficiently on multidimensional measurements, providing unbiased estimates of the resulting confidence intervals. We use this approach to study the expression of the mesodermal transcription factor Brachyury in mouse embryonic stem cells undergoing differentiation.

BEDwARS: a robust Bayesian approach to bulk gene expression deconvolution with noisy reference signatures

Differential gene expression in bulk transcriptomics data can reflect change of transcript abundance within a cell type and/or change in the proportions of cell types. Expression deconvolution methods can help differentiate these scenarios. BEDwARS is a Bayesian deconvolution method designed to address differences between reference signatures of cell types and corresponding true signatures underlying bulk transcriptomic profiles. BEDwARS is more robust to noisy reference signatures and outperforms leading in-class methods for estimating cell type proportions and signatures. Application of BEDwARS to dihydropyridine dehydrogenase deficiency identified the possible involvement of ciliopathy and impaired translational control in the etiology of the disorder.

Gaussian process deconvolution

Let us consider the deconvolution problem, i.e. to recover a latent source x ( ⋅ ) from the observations y = [ y 1 , … , y N ] of a convolution process y = x ⋆ h + η , where η is an additive noise, the observations in y might have missing parts with respect to y , and the filter h could be unknown. We propose a novel strategy to address this task when x is a continuous-time signal: we adopt a Gaussian process prior on the source x , which allows for closed-form Bayesian non-parametric deconvolution. We first analyse the direct model to establish the conditions under which the model is well-defined. Then, we turn to the inverse problem, where we study (i) some necessary conditions under which Bayesian deconvolution is feasible and (ii) to which extent the filter h can be learned from data or approximated for the blind deconvolution case. The proposed approach, termed Gaussian process deconvolution, is compared to other deconvolution methods conceptually, via illustrative examples, and using real-world datasets.

SiC MOSFET Trap Characterization Based on the Self-built Test Platform

The interface state of SiC/SiO2 is one of the key factors limiting the reliability and performance of SiC MOSFETs. In this paper, we built a dedicated detrap measurement platform to improve the accuracy of trap measurement. Based on the excellent voltage switching time of this platform, the entire voltage switching process can be controlled within 1 µs, and the time resolution is also improved. The original millisecond-level acquisition accuracy is improved to a microsecond level, and the identification range of traps is expanded. The traps were extracted by using the transient current method based on the Bayesian deconvolution algorithm. With this approach, we investigated the trapping mechanism of SiC MOSFETs and mainly characterized the trap locations, energy levels, and trapping time constants. The findings revealed the existence of three different types of traps/defects, Dp1, Dp2, and Dp3, with activation energies of 0.28 eV, 0.035 eV, and 0.084 eV, respectively. The non-destructive characterization of SiC MOSFET defects can be realized through the test platform and Bayesian deconvolution algorithm in this paper, which brings great convenience for trap characterization.

Trap Characterization of Trench-Gate SiC MOSFETs Based on Transient Drain Current

The SiC/SiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> interface state is one of the main factors that limit the performance and reliability of the SiC metal-oxide-semiconductor field-effect transistor (MOSFET). In this paper, we use a Bayesian deconvolution algorithm to optimize trap feature extraction based on the transient current method and improve the trap extraction accuracy. Using this method, we study the trap capture mechanism in SiC MOSFETs and mainly characterize the trap position, the trap energy level, and the capture time constant. The results obtained show that there are three different types of traps and defects, two of which are SiC interface traps at the gate-source and gate-drain interfaces, with activation energies of 0.089 eV and 0.035 eV, respectively; the third trap type is an oxide trap, and its time constant does not vary with temperature. The characterization results are verified via deep level transient spectroscopy, and the results show reasonable agreement with those obtained by the method proposed in this paper. This method can be combined with electrical stress testing in long-term reliability research to realize nondestructive characterization of the defects of SiC MOSFETs.

Model training periods impact estimation of COVID-19 incidence from wastewater viral loads

Wastewater-based epidemiology (WBE) has been deployed broadly as an early warning tool for emerging COVID-19 outbreaks. WBE can inform targeted interventions and identify communities with high transmission, enabling quick and effective responses. As the wastewater (WW) becomes an increasingly important indicator for COVID-19 transmission, more robust methods and metrics are needed to guide public health decision-making. This research aimed to develop and implement a mathematical framework to infer incident cases of COVID-19 from SARS-CoV-2 levels measured in WW. We propose a classification scheme to assess the adequacy of model training periods based on clinical testing rates and assess the sensitivity of model predictions to training periods. A testing period is classified as adequate when the rate of change in testing is greater than the rate of change in cases. We present a Bayesian deconvolution and linear regression model to estimate COVID-19 cases from WW data. The effective reproductive number is estimated from reconstructed cases using WW. The proposed modeling framework was applied to three Northern California communities served by distinct WW treatment plants. The results showed that training periods with adequate testing are essential to provide accurate projections of COVID-19 incidence.

Model Design in Bayesian Spectral Deconvolution

Model Design in Bayesian Spectral Deconvolution

Thermal Modelling an LED Bulb using IR Thermography and Bayesian Deconvolution

This study focuses on LED bulbs where LEDs are typically mounted on a metal core printed circuit board. A sample of LED bulbs having base E27 and a power range between 3.5-11.5 W have been measured. A thermal model will be found using thermal transient testing, infrared thermal measurements, and the Bayesian deconvolution. LED surface temperature will be measured without contact using a low-resolution IR imager. A test facility has been built which automatically switches power on and measures and records samples during the measurement cycle which typically lasts around one hour. This time duration was long enough for reaching the thermal steady-state condition. The procedure uses a measured thermal transient, first calculates the thermal impedance Zth, then the time constant spectrum Rζ(z) using the Bayesian deconvolution. An equivalent thermal network model will be found in Foster-presentation. This can be transformed to Cauer-presentation. The model is a driving point model for thermal impedance and includes thermal resistances and thermal capacitances. FEM simulation gives more detailed information about heat transfer paths and internal temperatures of an LED bulb.

Transformed Gaussian Random Fields for Unsupervised Image Deconvolution

This paper deals with the problem of Bayesian deconvolution. Starting from the classical Gaussian Markov Random Fields (GMRF) prior, we present a broader model referred as <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">transformed</i> GMRF (TGRF) in which the latent field results from a nonlinear transformation of a GMRF. We propose a Bayesian inference method to estimate TGRF from an observed image with known parameters, and introduce methods inspired from expectation-maximization in order to jointly deconvolve and estimate the statistical model's parameter for both GMRF and TGRF. Numerical results allow to determine the best inference method among several possibilities on fully synthetic data, on a <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">phantom</i> image and on real fluorescence microscopy images.

Deconvolving Native and Intact Protein Mass Spectra with UniDec.

Intact protein, top-down, and native mass spectrometry (MS) generally requires the deconvolution of electrospray ionization (ESI) mass spectra to assign the mass of components from their charge state distribution. For small, well-resolved proteins, the charge can usually be assigned based on the isotope distribution. However, it can be challenging to determine charge states with larger proteins that lack isotopic resolution, in complex mass spectra with overlapping charge states, and in native spectra that show adduction. To overcome these challenges, UniDec uses Bayesian deconvolution to assign charge states and to create a zero-charge mass distribution. UniDec is fast, user-friendly, and includes a range of advanced tools to assist in intact protein, top-down, and native MS data analysis. This chapter provides a step-by-step protocol and an in-depth explanation of the UniDec algorithm, and highlights the parameters that affect the deconvolution. It also covers advanced data analysis tools, such as macromolecular mass defect analysis and tools for assigning potential PTMs and bound ligands. Overall, this chapter provides users with a deeper understanding of UniDec, which will enhance the quality of deconvolutions and allow for more intricate MS experiments.

Optimization-Based Network Identification for Thermal Transient Measurements

Network identification by deconvolution is a proven method for determining the thermal structure function of a given device. The method allows to derive the thermal capacitances as well as the resistances of a one-dimensional thermal path from the thermal step response of the device. However, the results of this method are significantly affected by noise in the measured data, which is unavoidable to a certain extent. In this paper, a post-processing procedure for network identification from thermal transient measurements is presented. This so-called optimization-based network identification provides a much more accurate and robust result compared to approaches using Fourier or Bayesian deconvolution in combination with Foster-to-Cauer transformation. The thermal structure function obtained from network identification by deconvolution is improved by repeatedly solving the inverse problem in a multi-dimensional optimization process. The result is a non-diverging thermal structure function, which agrees well with the measured thermal impedance. In addition, the associated time constant spectrum can be calculated very accurately. This work shows the potential of inverse optimization approaches for network identification.

Improving Constraints on Planetary Interiors With PPs Receiver Functions.

Seismological constraints obtained from receiver function (RF) analysis provide important information about the crust and mantle structure. Here, we explore the utility of the free‐surface multiple of the P‐wave (PP) and the corresponding conversions in RF analysis. Using earthquake records, we demonstrate the efficacy of PPs‐RFs before illustrating how they become especially useful when limited data is available in typical planetary missions. Using a transdimensional hierarchical Bayesian deconvolution approach, we compute robust P‐to‐S (Ps)‐ and PPs‐RFs with InSight recordings of five marsquakes. Our Ps‐RF results verify the direct Ps converted phases reported by previous RF analyses with increased coherence and reveal other phases including the primary multiple reverberating within the uppermost layer of the Martian crust. Unlike the Ps‐RFs, our PPs‐RFs lack an arrival at 7.2 s lag time. Whereas Ps‐RFs on Mars could be equally well fit by a two‐ or three‐layer crust, synthetic modeling shows that the disappearance of the 7.2 s phase requires a three‐layer crust, and is highly sensitive to velocity and thickness of intra‐crustal layers. We show that a three‐layer crust is also preferred by S‐to‐P (Sp)‐RFs. While the deepest interface of the three‐layer crust represents the crust‐mantle interface beneath the InSight landing site, the other two interfaces at shallower depths could represent a sharp transition between either fractured and unfractured materials or thick basaltic flows and pre‐existing crustal materials. PPs‐RFs can provide complementary constraints and maximize the extraction of information about crustal structure in data‐constrained circumstances such as planetary missions.

Quantitative Performance Comparison of Thermal Structure Function Computations

The determination of thermal structure functions from transient thermal measurements using network identification by deconvolution is a delicate process as it is sensitive to noise in the measured data. Great care must be taken not only during the measurement process but also to ensure a stable implementation of the algorithm. In this paper, a method is presented that quantifies the absolute accuracy of network identification on the basis of different test structures. For this purpose, three measures of accuracy are defined. By these metrics, several variants of network identification are optimized and compared against each other. Performance in the presence of noise is analyzed by adding Gaussian noise to the input data. In the cases tested, the use of a Bayesian deconvolution provided the best results.

Measurements at laser materials processing machines: spectrum deconvolution including uncertainties and model selection

Abstract. Laser materials processing of workpieces using ultra-short pulsed lasers can lead to unwanted X-ray emission. Their dose rate and spectral distribution have been precisely determined. The measurements were carried out using a thermoluminescence detector (TLD)-based spectrometer in which 30 TLD planes are arranged one behind the other, the first 10 layers made of polymethyl methacrylate, while the remaining 20 layers are interspaced by absorbers with, from the front to the back, increasing atomic charge and thickness. The penetration depth of the radiation into the spectrometer depends on its energy, so that the energy-resolved spectrum of the radiation can be calculated from the TLD dose values by means of mathematical methods (Bayesian deconvolution). The evaluation process also takes into account both the uncertainties of all input quantities and the possibility of adopting different models for the spectrum form. This allowed the resulting spectra to be associated with their realistic uncertainty. The measurements are traceable to the Système international d'unités (SI), i.e. the International System of Units. The results not only provide manufacturers and users of ultra-short pulsed lasers with important information on the design of the machines with regard to radiation protection, but were also included in the recently concluded legislative procedure in the field of radiation protection in Germany.

FEDReD

Context.WhileGaiaenables us to probe the extended local neighbourhood in great detail, the thin disc structure at larger distances remains sparsely explored.Aims.We aim here to build a non-parametric 3D model of the thin disc structures handling both the extinction and the stellar density simultaneously.Methods.We developed a Bayesian deconvolution method in two dimensions: extinction and distance. It uses a reference catalogue whose completeness information defines the selection function. It is designed so that any complementary information from other catalogues can be added. It has also been designed to be robust to outliers, which are frequent in crowded fields, and differential extinction. The prior information is designed to be minimal: only a reference H-R diagram. We derived for this an empirical H-R diagram of the thin disc usingGaiaDR2 data, but synthetic isochrone-based H-R diagrams can also be used.Results.We validated the method on simulations and real fields using 2MASS and UKIDSS data complemented byGaiaDR2 photometry and parallaxes. We detail the results of two test fields: a 2MASS field centred around the NGC 4815 open cluster, which shows an over-density of both extinction and stellar density at the cluster distance, and a UKIDSS field atl = 10° where we recover the position of the Galactic bar.

FEDReD

Aims. We aim to map the 3D distribution of the interstellar extinction of the Milky Way disc up to distances larger than those probed with the Gaia parallax alone. Methods. We applied the FEDReD (Field Extinction-Distance Relation Deconvolver) algorithm to the 2MASS near-infrared photometry together with the Gaia DR2 astrometry and photometry. This algorithm uses a Bayesian deconvolution approach, based on an empirical HR-diagram representative of the local thin disc, in order to map the extinction as a function of distance of various fields of view. Results. We analysed more than 5.6 million stars to obtain an extinction map of the entire Galactic disc within |b| &lt; 0.24°. This map provides information up to 5 kpc in the direction of the Galactic centre and more than 7 kpc in the direction of the anticentre. This map reveals the complete shape of structures that are known locally, such as the Vela complex and the split of the local arm. Furthermore, our extinction map shows many large “clean bubbles”, especially the one in the Sagittarius-Carina complex, and four others, which define a structure that we nickname the butterfly.

Bayesian Deconvolution and Quantification of Metabolites from J-Resolved NMR Spectroscopy

Two-dimensional (2D) nuclear magnetic resonance (nmr) methods have become increasingly popular in metabolomics, since they have considerable potential to accurately identify and quantify metabolites within complex biological samples. 2D 1H J-resolved (jres) nmr spectroscopy is a widely used method that expands overlapping resonances into a second dimension. However, existing analytical processing methods do not fully exploit the information in the jres spectrum and, more importantly, do not provide measures of uncertainty associated with the estimates of quantities of interest, such as metabolite concentration. Combining the data-generating mechanisms and the extensive prior knowledge available in online databases, we develop a Bayesian method to analyse 2D jres data, which allows for automatic deconvolution, identification and quantification of metabolites. The model extends and improves previous work on one-dimensional nmr spectral data. Our approach is based on a combination of B-spline tight wavelet frames and theoretical templates, and thus enables the automatic incorporation of expert knowledge within the inferential framework. Posterior inference is performed through specially devised Markov chain Monte Carlo methods. We demonstrate the performance of our approach via analyses of datasets from serum and urine, showing the advantages of our proposed approach in terms of identification and quantification of metabolites.

Bayesian angular super‐resolution for sea‐surface target in forward‐looking scanning radar

As the traditional imaging dead zone, how to achieve high angular resolution in forward-looking region is always a big challenge in radar imaging field. In recent years, many approaches are proposed to solve this problem. However, most of the proposed forward-looking imaging approaches fail to work in the complex sea-surface situation. In this study, a Bayesian deconvolution method aimed at the sea-surface target is presented, which relies on the convolution model between the target distribution and the antenna pattern of scanning radar. In view of the situation of sea surface, Rayleigh distribution is adopted to describe the statistical property of sea-surface clutter, and lognormal distribution is employed to represent the prior distribution of sea-surface target. Then, the optimisation theory is utilised to obtain the iterative estimated solution. The simulation results are given to verify the superior performance of the presented method.

Randomized Phase III Oncology Trials: A Survey and Empirical Bayes Inference

Mandatory registration of eligible clinical trials at clinicaltrials.gov provides an unprecedented opportunity to survey existing trials that is immune to potential publication bias. Phase III oncology clinical trials represent the gold-standard evaluation on new therapeutic interventions to fight one of the most deadly diseases. Yet, the collective performance of these trials is not well understood. Through clinicaltrials.gov, we identified 130 eligible randomized phase III oncology trials in 2008–2012 and extracted results from 122 using clinicaltrials.gov and other sources. We estimated the distribution of the effect size of randomized phase III oncology trials through a new Bayesian deconvolution method, which allows the calculation of several performance measures. We found that about 22% of the interventions for cancer treatment that reach phase III have null or negative efficacy, and another 30% with strong positive efficacy. For the rest interventions, the majority have modest efficacy, which leads to insufficient power. The false positive rate is low at 0.13%, whereas as high as 33.5% of the trials could be false negatives. The distribution of the effect sizes also provide a tangible prior for Bayesian inference of future trials.