Kernel Integration Research Articles

Bounding Seed Loss from Isolated Habitat Patches

Dispersal of propagules (seeds, spores) from a geographically isolated habitat into an uninhabitable matrix can play a decisive role in driving population dynamics. ODE and integrodifference models of these dynamics commonly feature a “dispersal success” parameter representing the average proportion of dispersing propagules that remain in viable habitat. While dispersal success can be estimated by empirical measurements or by integration of dispersal kernels, one may lack resources for fieldwork or details on dispersal kernels for numerical computation. Here we derive simple upper bounds on the proportion of propagule loss—the complement of dispersal success—that require only habitat area, habitat perimeter, and the mean dispersal distance of a propagule. Using vector calculus in a probabilistic framework, we rigorously prove bounds for the cases of both symmetric and asymmetric dispersal. We compare the bounds to simulations of integral models for the population of Asclepias syriaca (common milkweed) at McKnight Prairie—a 14 hectare reserve surrounded by agricultural fields in Goodhue County, Minnesota—and identify conditions under which the bounds closely estimate propagule loss.

Strategies for designing machine learning models in renewable energy with insufficient data

In the area of renewable energy including wind, solar, and electricity, Machine Learning has established a far-reaching conceptual and algorithmic design framework. However, the design of Machine Learning architectures typically relies on substantial amounts of data. The data acquisition process often involves complex tasks of gathering, cleaning, and labeling data. This challenge is exacerbated in the energy sector due to concerns related to data security and the uniqueness of development initiatives, leading to a potential scarcity of available data. Quite commonly, researchers encounter a vast array of innovative projects that have not been studied in the past and therefore only limited data become available. Therefore, there is an urgent and highly motivated need to develop machine learning strategies that can construct reliable models in the presence of limited data. In this study, three machine learning models tailored for small-scale data scenarios are developed and analyzed. The first model, a fuzzy cluster ensemble approach, focuses on utilizing small training samples treated as the core of clustering. By applying clustering algorithms, one derives membership functions reflecting the structure of data that facilitate accurate predictions. The second model, a kernel integration framework, combines kernel functions with limited training data to calculate activation levels, enabling accurate prediction outputs. For both models, the final predictions are generated by applying weighted aggregations, where the weights are derived from the respective membership grades and activation levels of the data. Additionally, a neural network contrastive learning model is introduced, which enhances the training process by employing pairwise comparisons among data points. This method effectively increases the amount of useable training data, improving the model's ability to make accurate predictions. Together, these approaches provide a robust solution for making reliable predictions even in data-scarce scenarios within renewable energy research. Moreover, the predictive performance of these models is quantified using the Root Mean Square Error criterion. This study also assesses the performance of the developed models by forming a granular characterization of results delivered by the designed models. A comprehensive suite of experimental studies is included, accompanied by an exhaustive comparative analysis. Compared to conventional machine learning models, especially neural networks, the designed models offer reliable results even with limited training data, showing an improvement of around 20 %, thus laying the groundwork for digitization processes in the realm of sustainable energy.

NPTS-PK: A new point kernel code for fast calculation of 3D gamma radiation field

The point kernel integration method is commonly utilized for the analytical calculation of gamma radiation fields in the field of radiation protection and shielding design. This study introduces NPTS-PK, a program developed for rapid calculation of 3D gamma radiation fields based on the NPTS program and the point kernel integration method. Based on the geometric construction and input method of the NPTS program, NPTS-PK supports point kernel calculations for radiation sources and shielding structures of various complex shapes and materials. By calculating material attenuation coefficients using continuous energy cross-section parameters, combined with a more precise buildup factor calculation method, the accuracy of calculation results has been enhanced. Improvements in ray tracing processes and the implementation of the probability neighbor list method have accelerated geometric processing. To address the challenge of excessive computation time associated with large-scale grid counting in point kernel programs, NPTS-PK integrates several efficient acceleration techniques, elevating the speed of radiation field calculation by an order of magnitude. Tests on typical model and engineering scenario demonstrate that the deviation between NPTS-PK results and the reference values is within a few tens of percent, and the computational efficiency and accuracy are improved compared to the standard point kernel programs.

Enhanced Intrusion Detection in IoT with a Novel PRBF Kernel and Cloud Integration

Enhanced Intrusion Detection in IoT with a Novel PRBF Kernel and Cloud Integration

Continuous mapping of forest canopy height using ICESat-2 data and a weighted kernel integration of multi-temporal multi-source remote sensing data aided by Google Earth Engine.

Forest Canopy Height (FCH) is a crucial parameter that offers valuable insights into forest structure. Spaceborne LiDAR missions provide accurate FCH measurements, but a significant challenge is their point-based measurements lacking spatial continuity. This study integrated ICESat-2's ATL08-derived FCH values with multi-temporal and multi-source remote sensing (RS) datasets to generate continuous FCH maps for northern forests in Iran. Sentinel-1/2, ALOS-2 PALSAR-2, and FABDEM datasets were prepared in Google Earth Engine (GEE) for FCH mapping, each possessing unique spatial and geometrical characteristics that differ from those of the ATL08 product. Given the importance of accurately representing the geometrical characteristics of the ATL08 segments in modeling FCH, a novel Weighted Kernel (WK) approach was proposed in this paper. The WK approach could better represent the RS datasets within the ATL08 ground segments compared to other commonly used resampling approaches. The correlation between all RS data features improved by approximately 6% compared to previously employed approaches, indicating that the RS data features derived after convolving the WK approach are more predictive of FCH values. Furthermore, the WK approach demonstrated superior performance among machine learning models, with random forests outperforming other models, achieving a coefficient of determination (R2) of 0.71, root mean square error (RMSE) of 4.92m, and mean absolute percentage error (MAPE) of 29.95%. Furthermore, in contrast to previous studies using only summer datasets, this study included spring and autumn data from Sentinel-1/2, resulting in a 6% increase in R2 and a 0.5-m decrease in RMSE. The proposed methodology filled the research gaps and improved the accuracy of FCH estimations.

Establishment and Application of a Radiation Dose Rate Model for Nuclear Medicine Examinees.

Traditional methods for determining radiation dose in nuclear medicine include the Monte Carlo method, the discrete ordinate method, and the point kernel integration method. This study presents a new mathematical model for predicting the radiation dose rate in the vicinity of nuclear medicine patients. A new algorithm was created by combining the physical model of "cylinder superposition" of the human body with integral analysis to assess the radiation dose rate in the vicinity of nuclear medicine patients. The model accurately predicted radiation dose rates within distances of 0.1-3.0 m, with a deviation of less than 11% compared to observed rates. The model demonstrated greater accuracy at shorter distances from the radiation source, with a deviation of only 1.55% from observed values at 0.1 m. The model proposed in this study effectively represents the spatial and temporal distribution of the radiation field around nuclear medicine patients and demonstrates good agreement with actual measurements. This model has the potential to serve as a radiation dose rate alert system in hospital environments.

An ARF gene mutation creates flint kernel architecture in dent maize

Dent and flint kernel architectures are important characteristics that affect the physical properties of maize kernels and their grain end uses. The genes controlling these traits are unknown, so it is difficult to combine the advantageous kernel traits of both. We found mutation of ARFTF17 in a dent genetic background reduces IAA content in the seed pericarp, creating a flint-like kernel phenotype. ARFTF17 is highly expressed in the pericarp and encodes a protein that interacts with and inhibits MYB40, a transcription factor with the dual functions of repressing PIN1 expression and transactivating genes for flavonoid biosynthesis. Enhanced flavonoid biosynthesis could reduce the metabolic flux responsible for auxin biosynthesis. The decreased IAA content of the dent pericarp appears to reduce cell division and expansion, creating a shorter, denser kernel. Introgression of the ARFTF17 mutation into dent inbreds and hybrids improved their kernel texture, integrity, and desiccation, without affecting yield.

PDATC-NCPMKL: Predicting drug's Anatomical Therapeutic Chemical (ATC) codes based on network consistency projection and multiple kernel learning

The development and discovery of new drugs is time-consuming and needs lots of human and material resources. Therefore, discovery of novel effects of existing drugs is an important alternative way, which can accelerate the process of designing “new” drugs. The anatomical Therapeutic Chemical (ATC) classification system recommended by World Health Organization (WHO) is a basic research area in this regard. A novel ATC code of an existing drug suggests its novel effects. Some computational models have been proposed, which can predict the drug-ATC code associations. However, their performance is not very high. There still exist spaces for improvement. In this study, a new recommendation system (named PDATC-NCPMKL), which incorporated network consistency projection and multi-kernel learning, was designed to identify drug-ATC code associations. For drugs or ATC codes, several kernels were constructed, which were fused by a multiple kernel learning method and an additional kernel integration scheme. To enhance the performance, the drug-ATC code association adjacency matrix was reformulated by a variant of weighted K nearest known neighbors (WKNKN). The reformulated adjacency matrix, drug and ATC code kernels were fed into network consistency projection to generate the association score matrix. The proposed recommendation system was tested on the ATC codes at the second, third and fourth levels in drug ATC classification system using ten-fold cross-validation. The results indicated that all AUROC and AUPR values were close to or exceeded 0.96. Such performance was higher than some existing computational models. Some additional tests were conducted to prove the utility of adjacency matrix reformulation and to analyze the importance of drug and ATC code kernels.

A single-phase GPU-accelerated surface tension model using SPH

This paper presents an accelerated single-phase surface tension smoothed particle hydrodynamics (SPH) solver developed to run entirely on a graphics processing unit (GPU) capable of simulating millions of particles in three dimensions on a single GPU. The single-phase surface tension model is augmented with a contact line force to improve the prediction of the physics at the liquid-solid contact point. The surface tension model uses the modified dynamic boundary condition (mDBC) to impose no-slip conditions at the wall boundary. To enable simulations with millions of particles, the single-phase surface tension model has been implemented in the open-source SPH code DualSPHysics to exploit the GPU acceleration, paying special attention to the size of the kernel support and integration of the neighbour lists. The new scheme is validated using 2-D and 3-D test cases including drop deformation, drop oscillation, Rayleigh-Plateau instability and surface contact angles. The results show a good agreement with the analytical solutions with a standard spatial convergence behaviour. Profiling the new surface tension solver shows an additional computational complexity. The performance analysis shows that the new code has a speed up of up two orders of magnitude (x70-80) compared to the CPU-only code. Profiling the new CUDA kernels shows they have the near identical performance metrics with the main CUDA kernels in the original DualSPHysics solver.

Peanut kernel integrity detection based on deep learning convolution neural network

The edge features of peanut grain image are not considered in peanut grain integrity detection, and there are many noises, resulting in low accuracy of peanut grain integrity detection and poor effect of image edge noise removal. Therefore, this paper designs a peanut seed integrity detection method based on deep learning convolution neural network. The edge contour of peanut grain image is closed, the edge contour curve characteristics of peanut grain image are extracted, the noise area of peanut grain edge image is determined according to the filter window of Gaussian kernel function, and the wavelet descriptor in contour description operator is used to reduce the edge noise of peanut grain image. Input the image into the convolution neural network, update the weight by gradient descent method, construct the peanut seed integrity detection model, output the peanut seed integrity detection results, and realize the peanut seed integrity detection. The experimental results show that the proposed detection method can improve the accuracy of peanut grain detection and has certain feasibility.

Study on the Technology of Grapefruit Capsules Removement by Enzymatic

<p>Taking grapefruit as the experimental raw material, the fruit kernel integrity and extraction rate were comprehensively evaluated, and the grapefruit enzymatic de-coating process was explored to provide a basis for the industrial production of grapefruit fruit. The enzymatic de-coating process parameters were optimized by orthogonal test. The results showed that the water bath temperature was 40℃, the total amount of pectinase and cellulase was 1.25 mL, Study on the Technology of Grapefruit Capsules Removement by Enzymatic the ratio was 1:3, and the ratio of material to liquid (West) When the ratio of pomelo and mixed enzyme solution was 80g/100mL, the effect of decapsulation was the best. The grapefruit obtained in this case was relatively complete, and the extraction rate reached 93%.</p>

FRACTIONAL CALCULUS OPERATORS–BLOCH–TORREY PARTIAL DIFFERENTIAL EQUATION–ARTIFICIAL NEURAL NETWORKS–COMPUTATIONAL COMPLEXITY MODELING OF THE MICRO–MACROSTRUCTURAL BRAIN TISSUES WITH DIFFUSION MRI SIGNAL PROCESSING AND NEURONAL MULTI-COMPONENTS

Fractional calculus and fractional-order calculus are arranged in lineage as regards the mathematical models with complexity-theoretical tenets capable of capturing the subtle molecular dynamics by the integration of power-law convolution kernels into time- and space-related derivatives emerging in equations concerning the Magnetic Resonance Imaging (MRI) phenomena to which the fractional models of diffusion and relaxation are applied. Endowed with an intricate level of complexity and a unique physical and structural scaffolding at molecular and cellular levels with numerous synapses forming elaborate neural networks which entail in-depth probing and computing of patterns and signatures in individual cells and neurons, human brain as a heterogeneous medium is constituted of tissues with cells of different sizes and shapes, distributed across an extra-cellular space. Characterization of the unique brain cells is sought after to unravel the connections between different cells and tissues for accurate, reliable, robust and optimal models and computing. Accordingly, Diffusion Magnetic Resonance Imaging (DMRI), as a noninvasive and experimental imaging technique with clinical and research applications, provides a measure related to the diffusion characteristics of water in biological tissues, particularly in the brain tissues. Compatible with these aspects and beyond the diffusion coefficients’ measurement, DMRI technique aims to exceed the spatial resolution of the MRI images and draw inferences from the microstructural properties of the related medium. Thus, novel tools become essential for the description of the biological (organelles, membranes, macromolecules and so on) and neurological (axons, dendrites, neurons and so forth) tissues’ complexity. Mathematical model-based computational analyses with multifaceted methods to extract information from the DMRI with SpinDoctor into neuronal dynamics can provide quantitative parametric instruments in order to reflect the tissue properties focusing on the precise link between the tissue microstructure and signals acquired by employing advanced medical imaging technologies. Coalesced with accurate neuron geometry models as well as numerical DMRI simulations, a novel extended and multifaceted predictive mathematical model based on SpinDoctor and Bloch–Torrey partial differential equation (BTPDE) with the Caputo fractional-order derivative (FOD) with three-parameter [Formula: see text] Mittag-Leffler function (MLF) has been proposed and developed in our study by extending for the application on Brain Neuron Spin Unit dataset with the relevant multi-stage application-related steps. The feedforward neural networks (FFNNs) with BFGS Quasi-Newton equation, as one of the artificial neural network (ANN) algorithms, are applied on BTPDE with Caputo fractional-order derivative for the neurons and their algorithmic complexity is computed by building a BTPDE with Caputo FOD Neuron model based on different fractional orders. The fractional-order degree of the proposed and developed model is applied in relation to their corresponding complexity degrees. Consequently, experimentation and observations from the simulation-driven FFNN (with BFGS Quasi-Newton equation) learning scheme applied to the Bloch–Torrey PDE–Caputo FOD with MLF Neuron model (named as FFNN–BTPDE–CFODMLF Neuron model) proposed in this study, are made. Thus, by investigating whether the mathematical models based on the accurate neuron geometry models obtained can be optimized by comparing the errors in order to define the order parameter and identify to what optimal extent the errors are in relation to the prediction results with a particular focus on the neuron model, we have been able to estimate and predict brain microstructure through DMRI, accentuating mathematical and medical contributions based on the exploitation and corroboration of powerful modeling as well as computational capabilities.

Application of HAM for Nonlinear Integro-Differential Equations of Order Two

In this work, we consider the second order nonlinear integro-differential Equation (IDEs) of the Volterra-Fredholm type. One of the popular methods for solving Volterra or Fredholm type IDEs is the method of quadrature while the problem of consideration is a linear problem. If IDEs are nonlinear or integral kernel is complicated, then quadrature rule is not most suitable; therefore, other types of methods are needed to develop. One of the suitable and effective method is homotopy analysis method (HAM) developed by Liao in 1992. To apply HAM, we firstly reduced the IDEs into nonlinear integral Equation (IEs) of Volterra-Fredholm type; then the standard HAM was applied. Gauss-Legendre quadrature formula was used for kernel integrations. Obtained system of algebraic equations was solved numerically. Moreover, numerical examples demonstrate the high accuracy of the proposed method. Comparisons with other methods are also provided. The results show that the proposed method is simple, effective and dominated other methods.

3rdParTEE: Securing Third-Party IoT Services Using the Trusted Execution Environment

Advancements in the Internet of Things (IoT) have resulted in the connection and deployment of numerous smart and embedded devices. Although such devices enable various services such as smart grids, they attract more attackers to the IoT world. A trusted execution environment (TEE), which can be created by using TrustZone technology, is a promising security artifact for protecting critical operations and sensitive data of IoT devices. Unfortunately, although TrustZone is available in most ARM architecture-based devices ranging from microcontrollers to high-end smart devices, it has not been widely adopted by third-party IoT service providers because of its limited access. That is, because the TEE is maintained by the TEE platform vendors to preserve its security. Therefore, third parties must adhere to strict policies and procedures to ensure the deployment of trusted services in the TEE. This aspect hinders the fast development and deployment of IoT services. In this work, we propose 3rdParTEE to address this problem by enabling third-party IoT service providers to readily run their trusted services, thereby minimizing their dependency on the TEE maintainers. Specifically, 3rdParTEE facilitates the secure running of the third-party’s native kernel driver in the TEE without hampering the security of the existing TEE components. To demonstrate the effectiveness of our approach, we ran three kernel drivers for maintaining the IoT services platform (e.g., kernel integrity check) in the TEE. Additionally, during the performance evaluation, we observed a reasonable performance overhead of up to 7% when running the kernel drivers in such a secure manner.

A Hardware Platform for Ensuring OS Kernel Integrity on RISC-V

The OS kernel is typically preassumed as a trusted computing base in most computing systems. However, it also implies that once an attacker takes control of the OS kernel, the attacker can seize the entire system. Because of such security importance of the OS kernel, many works have proposed security solutions for the OS kernel using an external hardware module located outside the processor. By doing this, these works can realize the physical isolation of security solutions from the OS kernel running in the processor, but they cannot access the inner state of the processor, which attackers can manipulate. Thus, they elaborated several methods to overcome such limited capability of external hardware. However, those methods usually come with several side effects, such as high-performance overhead, kernel code modifications, and/or excessively complicated hardware designs. In this paper, we introduce RiskiM, a new hardware-based monitoring platform to ensure kernel integrity from outside the host system. To deliver the inner state of the host to RiskiM, we have devised a hardware interface architecture, called PEMI. Through PEMI, RiskiM is supplied with all internal states of the host system essential for fulfilling its monitoring task to protect the kernel. To empirically validate our monitoring platform’s security strength and performance, we have fully implemented PEMI and RiskiM on a RISC-V based processor and FPGA, respectively. Our experiments show that RiskiM succeeds in the host kernel protection by detecting even the advanced attacks which could circumvent previous solutions, yet suffering from virtually no aforementioned side effects.

Dual-mode Integrity Measurement System Based on Virtualized TEE

The Trusted Execution Environment (TEE) is mainly used to protect sensitive information and ensure information security. TEE has been used in the system structure of various mobile devices. With the development of TEE, attack methods against TEE gradually appear. The virtualized TEE architecture proposed by existing research can solve some security problems caused by TEE architecture, but there are still some deficiencies. To enhance the TEE architecture, a virtualized TEE-based kernel integrity measurement scheme is proposed, and identity-based and neural network-based integrity measurements are used in different scenarios. The experimental results show that the designed dual-mode integrity measurement system can effectively detect attacks against different kernel objects and has little impact on system performance, thus achieving security enhancements for virtualized TEE.

Open-Source Multi-GPU-Accelerated QM/MM Simulations with AMBER and QUICK.

The quantum mechanics/molecular mechanics (QM/MM) approach is an essential and well-established tool in computational chemistry that has been widely applied in a myriad of biomolecular problems in the literature. In this publication, we report the integration of the QUantum Interaction Computational Kernel (QUICK) program as an engine to perform electronic structure calculations in QM/MM simulations with AMBER. This integration is available through either a file-based interface (FBI) or an application programming interface (API). Since QUICK is an open-source GPU-accelerated code with multi-GPU parallelization, users can take advantage of "free of charge" GPU-acceleration in their QM/MM simulations. In this work, we discuss implementation details and give usage examples. We also investigate energy conservation in typical QM/MM simulations performed at the microcanonical ensemble. Finally, benchmark results for two representative systems in bulk water, the N-methylacetamide (NMA) molecule and the photoactive yellow protein (PYP), show the performance of QM/MM simulations with QUICK and AMBER using a varying number of CPU cores and GPUs. Our results highlight the acceleration obtained from a single or multiple GPUs; we observed speedups of up to 53× between a single GPU vs a single CPU core and of up to 2.6× when comparing four GPUs to a single GPU. Results also reveal speedups of up to 3.5× when the API is used instead of FBI.

Effects of Field Processing of Sorghum Grain on Popping Traits

Interest in the use of popped sorghum in food products has resulted in a niche market for sorghum hybrids with high popping quality but little work has been done to assess the relative effects of field processing methods of grain on popping quality. This study evaluated the relative effects of harvest moisture and threshing methods on the popping quality of sorghum grain. A grain sorghum hybrid with good popping quality was produced during two different years in Texas wherein it was harvested at two moisture levels (low and high) and grain was removed from panicles using five different threshing methods (hand, rubber belt, metal brushes and two metal concave bar systems). Years, harvest moisture content and threshing method influenced all three popping quality measurements (popping efficacy, expansion ratio and flake size), but threshing method had an order of magnitude larger effect than either moisture level or year. While many of the interactions were significant, they did not influence the general trends observed. As such, the threshing methods with less direct impact force on the grain (hand and rubber belt) had higher popping quality than those samples threshed with greater impact force on the grain (metal-based systems). The popping quality differences between threshing system are likely due to a reduction in kernel integrity caused by the impacts to the kernel that occurred while threshing the grain. The results herein indicate that field processing of the grain, notably threshing method has significant impacts on the popping quality and should be taken into consideration when grain sorghum is harvested for popping purposes.

MSU-Net: Multi-Scale U-Net for 2D Medical Image Segmentation.

Aiming at the limitation of the convolution kernel with a fixed receptive field and unknown prior to optimal network width in U-Net, multi-scale U-Net (MSU-Net) is proposed by us for medical image segmentation. First, multiple convolution sequence is used to extract more semantic features from the images. Second, the convolution kernel with different receptive fields is used to make features more diverse. The problem of unknown network width is alleviated by efficient integration of convolution kernel with different receptive fields. In addition, the multi-scale block is extended to other variants of the original U-Net to verify its universality. Five different medical image segmentation datasets are used to evaluate MSU-Net. A variety of imaging modalities are included in these datasets, such as electron microscopy, dermoscope, ultrasound, etc. Intersection over Union (IoU) of MSU-Net on each dataset are 0.771, 0.867, 0.708, 0.900, and 0.702, respectively. Experimental results show that MSU-Net achieves the best performance on different datasets. Our implementation is available at https://github.com/CN-zdy/MSU_Net.

Comparison of variance-reduction techniques for gamma dose rate determination

Three-dimensional computer simulation and virtual reality technology enable the visualization of dose encountered by workers during dismantling operations by using simplified real-time dose computation tools. Such tools generally use a macroscopic approach for gamma dose rate calculation, namely the point kernel integration method with build-up factors. This simplified physical model enhances calculation performance but presents also some restrictions. In contrast, stochastic Monte Carlo methods enable a precise estimation of gamma dose rate, but computing time is prohibitive for real-time dose applications. To speed up the simulation, Monte Carlo codes can be used in combination with variance-reduction techniques, which have to be used very cautiously to stay within their limits of validity. This paper presents a comparison between two variance-reduction techniques implemented in the Monte Carlo code TRIPOLI-4 $$\circledR $$ , the exponential transform and the adaptive multilevel splitting, testing their efficiency in dismantling-like configurations.Both methods behave better in deep penetration problems but require a good amount of user experience in the creation of the importance map. This study shows the need to develop a new type of algorithm capable to tackle configurations where the lack of collisions can limit the efficiency of the current VRT.