BackgroundMR fingerprinting (MRF) is a novel method for quantitative assessment of in vivo MR relaxometry that has shown high precision and accuracy. However, the method requires data acquisition using customized, complex acquisition strategies and dedicated post processing methods thereby limiting its widespread application.ObjectiveTo develop a deep learning (DL) network for synthesizing MRF signals from conventional magnitude-only MR imaging data and to compare the results to the actual MRF signal acquired.MethodsA U-Net DL network was developed to synthesize MRF signals from magnitude-only 3D T1-weighted brain MRI data acquired from 37 volunteers aged between 21 and 62 years of age. Network performance was evaluated by comparison of the relaxometry data (T1, T2) generated from dictionary matching of the deep learning synthesized and actual MRF data from 47 segmented anatomic regions. Clustered bootstrapping involving 10,000 bootstraps followed by calculation of the concordance correlation coefficient were performed for both T1 and T2 MRF data pairs. 95% confidence limits and the mean difference between true and DL relaxometry values were also calculated.ResultsThe concordance correlation coefficient (and 95% confidence limits) for T1 and T2 MRF data pairs over the 47 anatomic segments were 0.8793 (0.8136–0.9383) and 0.9078 (0.8981–0.9145) respectively. The mean difference (and 95% confidence limits) were 48.23 (23.0–77.3) s and 2.02 (−1.4 to 4.8) s.ConclusionIt is possible to synthesize MRF signals from MRI data using a DL network, thereby creating the potential for performing quantitative relaxometry assessment without the need for a dedicated MRF pulse sequence.

We present Knowledge Engine for Genomics (KnowEnG), a free-to-use computational system for analysis of genomics data sets, designed to accelerate biomedical discovery. It includes tools for popular bioinformatics tasks such as gene prioritization, sample clustering, gene set analysis, and expression signature analysis. The system specializes in "knowledge-guided" data mining and machine learning algorithms, in which user-provided data are analyzed in light of prior information about genes, aggregated from numerous knowledge bases and encoded in a massive "Knowledge Network." KnowEnG adheres to "FAIR" principles (findable, accessible, interoperable, and reuseable): its tools are easily portable to diverse computing environments, run on the cloud for scalable and cost-effective execution, and are interoperable with other computing platforms. The analysis tools are made available through multiple access modes, including a web portal with specialized visualization modules. We demonstrate the KnowEnG system's potential value in democratization of advanced tools for the modern genomics era through several case studies that use its tools to recreate and expand upon the published analysis of cancer data sets.

For many years, the computer industry has relied on steady progress in the exponential rate of scaling MOSFETs in integrated circuits. The usual expectation, based on Moore's law, is that the number of transistors able to be packed on a chip doubles roughly every 18 months. Sustaining this pace requires aggressive research into the numerous bottlenecks that threaten to slow it down. Much research has gone into the photolithography needed to produce such dense circuits, device structures that would allow smaller channel lengths, and a plethora of other materials and device advances that help sustain the present rate of scaling. In the past decade, however, another issue has emerged that threatens to impose an absolute limit on how many transistors can be packed onto a die. This is the issue of heat dissipation.

In this paper, we develop a convolutional neural network model to predict the mechanical properties of a two-dimensional checkerboard composite quantitatively. The checkerboard composite possesses two phases: one phase is soft and ductile while the other is stiff and brittle. The ground-truth data used in the training process are obtained from finite element analyses under the assumption of plane stress. Monte Carlo simulations and central limit theorem are used to find the size of the dataset needed. Once the training process is completed, the developed model is validated using data unseen during training. The developed neural network model captures the stiffness, strength, and toughness of checkerboard composites with high accuracy. Also, we integrate the developed model with a genetic algorithm (GA) optimizer to identify the optimal microstructural designs. The genetic algorithm optimizer adopted here has several operators: selection, crossover, mutation, and elitism. The optimizer converges to configurations with highly enhanced properties. For the case of the modulus and starting from randomly-initialized generation, the GA optimizer converges to the global maximum which involves no soft elements. Also, the GA optimizers, when used to maximize strength and toughness, tend towards having soft elements in the region next to the crack tip.

In this paper, we describe the development of moving mesh adaptation framework and its application to charge transport simulation of semiconductor devices, with emphasis on its relevance to power semiconductor devices. Mesh adaptivity in the context of semiconductor device simulation is an important problem and can help deal with the convergence and numerical stability issues, as well as automate the meshing process. We demonstrate the efficacy of our proposed meshing scheme through the simulation of a GaN-based power diode, as well as a Si diode with a non-rectangular doping profile, by externally coupling our framework to Sentaurus Device TCAD. We perform error analysis and compare our results with simulations based on high-resolution uniform structured meshes as well as manually refined axis-aligned meshes. In addition to the benefits in terms of accuracy, automation, and generality, our method can be regarded as a stepping stone toward computationally scalable and adaptive semiconductor device simulations.

There have been many efforts to identify the relation between the geometric structure of carbon-nanomaterials and the intensity profile of Raman spectra. As a result, researchers have reported intensity profiles changing with the geometric properties of carbon-nanomaterials, i.e. the length of carbon nanotubes and the thickness of graphene. Based on these measured data, we constructed an autonomous framework that can deduce the geometric property from a Raman spectra pattern using a machine learning algorithm. In this work, we focus on the Raman peak shift recognition using principal component analysis (PCA) to identify the number of graphene layers and this framework can accelerate processes in both measurement and geometric property analysis.

New designs for vertical 2D-materials-based TFETs are proposed in this paper adopting asymmetric layer numbers for the top and bottom layer with undoped source/drain using Black Phosphorus as an example. The results show that abrupt turn-on and I on /I off > 10 5 can be sustained when the channel length is down to sub-5 nm. The results are benchmarked against other TFETs based on promising 2D materials homo-/hetero-structures, meanwhile, the limitations, as well as guidelines, are presented.

P2PLoc envisions wearable Internet of Things devices that compute the relative positions of each user, resulting in a topology or configuration of mobile users that can be tracked in real time for group-motion applications.

Background Clustering is one of the most common techniques in data analysis and seeks to group together data points that are similar in some measure. Although there are many computer programs available for performing clustering, a single web resource that provides several state-of-the-art clustering methods, interactive visualizations and evaluation of clustering results is lacking. Methods ClusterEnG (acronym for Clustering Engine for Genomics) provides a web interface for clustering data and interactive visualizations including 3D views, data selection and zoom features. Eighteen clustering validation measures are also presented to aid the user in selecting a suitable algorithm for their dataset. ClusterEnG also aims at educating the user about the similarities and differences between various clustering algorithms and provides tutorials that demonstrate potential pitfalls of each algorithm. Conclusions The web resource will be particularly useful to scientists who are not conversant with computing but want to understand the structure of their data in an intuitive manner. The validation measures facilitate the process of choosing a suitable clustering algorithm among the available options. ClusterEnG is part of a bigger project called KnowEnG (Knowledge Engine for Genomics) and is available at http://education.knoweng.org/clustereng.

Among efforts made to improve thermoelectric efficiency, the use of structurally modified graphene nanomaterials as thermoelectric matter are one of the promising strategies owing to their fascinating physical and electrical properties, and these materials are anticipated to be less thermally conductive than regular graphene structures, as a result of an additional phonon scattering introduced at the modified surfaces. In this study, we explore the thermal conductivity behaviors of strain-induced rippled graphene sheets by varying the ripple amplitude, periodicity, and dimensions of the structure. We introduce a technique which enables creation of a graphene sheet with evenly distributed ripples in molecular dynamics simulation, and the Green-Kubo linear response theory is used to calculate the thermal conductivity of the structures of interest. The results reveal the reduction of thermal conductivity with the greater degree of strain, the smaller system dimension, and the shorter ripple wavelength, which, in turn, could lead to the thermoelectric efficiency enhancement. This work has significance in that it presents the capability of generating repeated and controllable patterns in molecular dynamics, and so, it enables the atomic-level transport study in the regularly patterned two-dimensional surface or in any structures with a specified degree of strain.

Summary Clustering is one of the most common techniques used in data analysis to discover hidden structures by grouping together data points that are similar in some measure into clusters. Although there are many programs available for performing clustering, a single web resource that provides both state-of-the-art clustering methods and interactive visualizations is lacking. ClusterEnG (acronym for Clustering Engine for Genomics) provides an interface for clustering big data and interactive visualizations including 3D views, cluster selection and zoom features. ClusterEnG also aims at educating the user about the similarities and differences between various clustering algorithms and provides clustering tutorials that demonstrate potential pitfalls of each algorithm. The web resource will be particularly useful to scientists who are not conversant with computing but want to understand the structure of their data in an intuitive manner. Availability ClusterEnG is part of a bigger project called KnowEnG (Knowledge Engine for Genomics) and is available at http://education.knoweng.org/clustereng . Contact songi@illinois.edu

A new vertical tunnel FET design based on black phosphorus is presented in this paper adopting asymmetric layer numbers for top and bottom layer with undoped drain. The results show that the SS and I-on/I-off can be maintained below 10 mV/dec and beyond 10(5), respectively, when channel length is down to 3 nm.

This paper explores self-heating effects on junctionless gate-all-around nanowire MOSFET using self-consistently coupled 3D full band electro-thermal transport. The self-consistent algorithm begins by supplying the heat generation data from the 3D electron Monte Carlo with 2D quantum correction to the phonon Monte Carlo. Subsequently, the phonon Monte Carlo transports the phonons introduced from the electron simulation and considers their scattering through the anharmonic three-phonon processes. The anharmonic three-phonon decay and the use of full dispersion facilitate a detailed description of heat transfer and the determination of the temperature map. We compare the performance of gate-all-around junctionless against the conventional inversion mode gate-all-around MOSFET. Our results indicate that junctionless MOSFET has less self-heating effects than the conventional inversion mode device, particularly at the limits of high currents.

In this work, we calculate the thermal conductivity of layered bismuth telluride (Bi2Te3) thin films by solving the Boltzmann transport equation in the relaxation-time approximation using full phonon dispersion and compare our results with recently published experimental data and molecular dynamics simulation. The group velocity of each phonon mode is readily extracted from the full phonon dispersion obtained from first-principle density-functional theory calculation and is used along with the phonon frequency to compute the various scattering terms. Our model incorporates the typical interactions impeding thermal transport (e.g., umklapp, isotope, and boundary scatterings) and introduces a new interaction capturing the reduction of phonon transmission through van der Waals interfaces of adjacent Bi2Te3 quintuple layers forming the virtual superlattice thin film. We find that this novel approach extends the empirical Klemens-Callaway relaxation model in such anisotropic materials and recovers the experimental anisotropy while using a minimal set of parameters.

Thermoelectric efficiency has been limited to a non-practical range because of the electronic material properties that are not easily controlled. Therefore, limiting thermal transport has been the best approach to enhance efficiency, and efforts to reduce the thermal conductivity even further have been extensively reported. In this work, we investigate thermoelectric transport coefficients of uniformly straight and sinusoidally undulated nanowires and compare the resulting thermoelectric figure-of-merit. The improved efficiency should suggest another strategy to effectively control thermal transport in semiconductor nanostructures.

Improvement of thermoelectric efficiency has been very challenging in the solid-state industry due to the interplay among transport coefficients which measure the efficiency. In this work, we modulate the geometry of nanowires to interrupt thermal transport with causing only a minimal impact on electronic transport properties, thereby maximizing the thermoelectric power generation. As it is essential to scrutinize comprehensively both electronic and thermal transport behaviors for nano-scale thermoelectric devices, we investigate the Seebeck coefficient, the electrical conductance, and the thermal conductivity of sinusoidally corrugated silicon nanowires and eventually look into an enhancement of the thermoelectric figure-of-merit [Formula: see text] from the modulated nanowires over typical straight nanowires. A loss in the electronic transport coefficient is calculated with the recursive Green function along with the Landauer formalism, and the thermal transport is simulated with the molecular dynamics. In contrast to a small influence on the thermopower and the electrical conductance of the geometry-modulated nanowires, a large reduction of the thermal conductivity yields an enhancement of the efficiency by 10% to 35% from the typical nanowires. We find that this approach can be easily extended to various structures and materials as we consider the geometrical modulation as a sole source of perturbation to the system.

A 3-D full-band particle Monte Carlo (MC) simulator, with full electron and phonon dispersion and a 2-D quantum correction is self-consistently coupled to a phonon MC simulator. The coupling entails feeding the phonon data obtained from the 3-D electrical MC to the phonon MC. The phonon MC reciprocates by providing the resulting spatial temperature map, which is used in the electron MC, with tempe...