Low Computational Cost Research Articles

Multiscale enhanced non‐uniform transformation field analysis

AbstractEnhanced transformation field analysis (E‐TFA), recently proposed for reduced‐order modeling, is here formulated for and applied to multiscale analysis. The approach is able to reproduce a highly complex nonlinear macroscale behavior, resulting from a microstructure with cohesive interfaces embedded in an elasto‐plastic bulk. E‐TFA features a consistent tangent matrix in its solution procedure, which enables a straightforward definition of the upscaled tangent stiffness tensor. Numerical tests show that, compared to FE, the proposed approach yields accurate solutions at a lower computational cost.

A mathematical model to design pillar-shaped bioinspired mechanical sensors

Soft robots and sensor/actuator systems are often based on bioinspired designs to leverage nature patterns. Specifically, pillar-shaped sensors are useful for human activity monitoring, locomotion of soft robots or treatment of cardiovascular diseases. If electric or magnetic particles are added in the manufacturing process, these structures can be tuned through remote fields to attain a specific mechanical behaviour. This promising technique has direct applications in high-impact fields such as bioengineering, soft robots or sensor designing. Filament-shaped smart sensors can send electrical signals when subjected to an external mechanical stimulus or provide a mechanical response when a remote and controllable field is applied broadening their possibilities of action. As the efficient design of these structures is highly challenging, developing a technical tool with a low computational cost to help throughout layout processes (i.e. inverse engineering) is pivotal. Theoretical modelling of the kinematics and dynamics of a wire-shaped structure under an external action is the first step to provide a methodology to help designing mechanical sensors in an efficient, understandable and low time-consuming way. The event of mechanical deformation after receiving the external stimulus and before sending the corresponding output signal is key in the conceptualisation process of smart sensors. This work intends to give insight into the dynamics of a deformable pillar-shaped sensor component under an external action without addressing or coupling its causes and, hence, provide the general mechanical framework to serve as basis for multiphysics formulations for pillar-shaped sensor design.

Adaptive mesh refinement towards optimized mesh generation for large eddy simulation of turbulent combustion in a typical micro gas turbine combustor

Over the last years, several advanced micro Gas Turbine (mGT) cycle developments have been proposed, aiming at making the mGT more fuel and operational flexible. However, accurate data on real industrial combustors, assessing the performances and emissions of the combustion under unconventional diluted conditions or fuels involved in these novel cycles, are still missing. In this framework, Large Eddy Simulations, who allow to accurately assess the unsteady effects coupled to turbulent-chemistry interaction of reacting flows, offer an opportunity to better assess the combustion behaviour under these specific conditions. However, the computational cost remains much higher compared to RANS simulations. The mesh generation process might be complex, especially when the region of interest is not intuitively known. Therefore, Adaptive Mesh Refinement (AMR) methods allow the mesh to be refined only in the region where finer cells are required to capture important phenomena, i.e. in the flame front. By dynamically refining the mesh all along the simulation, the mesh is optimized in terms of cell quantity and distribution for more accurate results at potentially lower computational costs. In this work, an adaptive mesh refinement method based on a predefined criterion is successfully applied to the LES of a typical industrial mGT combustor, the Turbec T100. The results show that the adaptation strategy allows to automatically generate a dynamic mesh that is able to capture correctly the flame over time for an increased computational cost of 15%. Therefore, the human meshing effort is reduced while the automatic meshing leads to an optimized mesh.

Method for scalable and performant GPU-accelerated simulation of multiphase compressible flow

Multiphase compressible flows are often characterized by a broad range of space and time scales, entailing large grids and small time steps. Simulations of these flows on CPU-based clusters can thus take several wall-clock days. Offloading the compute kernels to GPUs appears attractive but is memory-bound for many finite-volume and -difference methods, damping speedups. Even when realized, GPU-based kernels lead to more intrusive communication and I/O times owing to lower computation costs. We present a strategy for GPU acceleration of multiphase compressible flow solvers that addresses these challenges and obtains large speedups at scale. We use OpenACC for directive-based offloading of all compute kernels while maintaining low-level control when needed. An established Fortran preprocessor and metaprogramming tool, Fypp, enables otherwise hidden compile-time optimizations. This strategy exposes compile-time optimizations and high memory reuse while retaining readable, maintainable, and compact code. Remote direct memory access realized via CUDA-aware MPI and GPUDirect reduces halo-exchange communication time. We implement this approach in the open-source solver MFC [1]. Metaprogramming results in an 8-times speedup of the most expensive kernels compared to a statically compiled program, reaching 46% of peak FLOPs on modern NVIDIA GPUs and high arithmetic intensity (about 10 FLOPs/byte). In representative simulations, a single NVIDIA A100 GPU is 7-times faster compared to an Intel Xeon Cascade Lake (6248) CPU die, or about 300-times faster compared to a single such CPU core. At the same time, near-ideal (97%) weak scaling is observed for at least 13824 GPUs on OLCF Summit. A strong scaling efficiency of 84% is retained for an 8-times increase in GPU count. Collective I/O, implemented via MPI3, helps ensure the negligible contribution of data transfers (<1% of the wall time for a typical, large simulation). Large many-GPU simulations of compressible (solid-)liquid-gas flows demonstrate the practical utility of this strategy.

Metaheuristic and Heuristic Algorithms-Based Identification Parameters of a Direct Current Motor

Direct current motors are widely used in industry applications, and it has become necessary to carry out studies and experiments for their optimization. In this manuscript, a comparison between heuristic and metaheuristic algorithms is presented, specifically, the Steiglitz–McBride, Jaya, Genetic Algorithm (GA), and Grey Wolf Optimizer (GWO) algorithms. They were used to estimate the parameters of a dynamic model that approximates the actual responses of current and angular velocity of a DC motor. The inverse of the Euclidean distance between the current and velocity errors was defined as the fitness function for the metaheuristic algorithms. For a more comprehensive comparison between algorithms, other indicators such as mean squared error (MSE), standard deviation, computation time, and key points of the current and velocity responses were used. Simulations were performed with MATLAB/Simulink 2010 using the estimated parameters and compared to the experiments. The results showed that Steiglitz–McBride and GWO are better parametric estimators, performing better than Jaya and GA in real signals and nominal parameters. Indicators say that GWO is more accurate for parametric estimation, with an average MSE of 0.43%, but it requires a high computational cost. On the contrary, Steiglitz–McBride performed with an average MSE of 3.32% but required a much lower computational cost. The GWO presented an error of 1% in the dynamic response using the corresponding indicators. If a more accurate parametric estimation is required, it is recommended to use GWO; however, the heuristic algorithm performed better overall. The performance of the algorithms presented in this paper may change if different error functions are used.

TL‐GNN: Android Malware Detection Using Transfer Learning

ABSTRACTMalware growth has accelerated due to the widespread use of Android applications. Android smartphone attacks have increased due to the widespread use of these devices. While deep learning models offer high efficiency and accuracy, training them on large and complex datasets is computationally expensive. Hence, a method that effectively detects new malware variants at a low computational cost is required. A transfer learning method to detect Android malware is proposed in this research. Because of transferring known features from a source model that has been trained to a target model, the transfer learning approach reduces the need for new training data and minimizes the need for huge amounts of computational power. We performed many experiments on 1.2 million Android application samples for performance evaluation. In addition, we evaluated how well our framework performed in comparison with traditional deep learning and standard machine learning models. In comparison with state‐of‐the‐art Android malware detection methods, the proposed framework offers improved classification accuracy of 98.87%, a precision of 99.55%, recall of 97.30%, F1‐measure of 99.42%, and a quicker detection rate of 5.14 ms using the transfer learning strategy.

Efficient pulse sequence design framework for high-dimensional MR fingerprinting scans using systematic error index.

For effective optimization of MR fingerprinting (MRF) pulse sequences, estimating and minimizing errors from actual scan conditions are crucial. Although virtual-scan simulations offer an approximation to these errors, their computational demands become expensive for high-dimensional MRF frameworks, where interactions between more than two tissue properties are considered. This complexity makes sequence optimization impractical. We introduce a new mathematical model, the systematic error index (SEI), to address the scalability challenges for high-dimensional MRF sequence design. By eliminating the need to perform dictionary matching, the SEI model approximates quantification errors with low computational costs. The SEI model was validated in comparison with virtual-scan simulations. The SEI model was further applied to optimize three high-dimensional MRF sequences that quantify two to four tissue properties. The optimized scans were examined in simulations and healthy subjects. The proposed SEI model closely approximated the virtual-scan simulation outcomes while achieving hundred- to thousand-times acceleration in the computational speed. In both simulation and in vivo experiments, the optimized MRF sequences yield higher measurement accuracy with fewer undersampling artifacts at shorter scan times than the heuristically designed sequences. We developed an efficient method for estimating real-world errors in MRF scans with high computational efficiency. Our results illustrate that the SEI model could approximate errors both qualitatively and quantitatively. We also proved the practicality of the SEI model of optimizing sequences for high-dimensional MRF frameworks with manageable computational power. The optimized high-dimensional MRF scans exhibited enhanced robustness against undersampling and system imperfections with faster scan times.

Computing Diversity Paradigm for the Utilization of Unused Telephony and Marine Infrastructure

Computing systems play a significant role in data storage and processing. They enable the execution of data processing in data intensive domains such as astronomy. Therefore, having access to computing platforms is important for capital constrained astronomy organizations (CCAOs). The CCAOs need computing facilities to process the data acquired from an increasing number of telescopes. It is important to ensure that CCAOs execute the required computing at low cost. The research proposes the astronomy computing diversity paradigm to address this challenge for the CCAO. The proposed astronomy computing diversity paradigm uses disaggregated servers to design Class A and Class B data centres. The low–cost computing is realized by reducing computing platform operational costs. In the proposed solution, the Class A data centre uses unutilized telephony infrastructure intended for telephony. The Class B data centre is hosted in aquaria arrays. The Class B data centre reduces the cooling costs by using water in big aquaria. The performance evaluation is done focusing on the power usage effectiveness (PUE) and the angular resolution. Results show that using the aquaria-based data centre enhances the PUE and angular resolution by 22.3% and 40.4% on average, respectively. It reduces the operational costs by 79% on average.

Efficient global sensitivity analysis method for dynamic models in high dimensions

AbstractDynamic models generating time‐dependent model predictions are typically associated with high‐dimensional input spaces and high‐dimensional output spaces, in particular if time is discretized. It is computationally prohibitive to apply traditional global sensitivity analysis (SA) separately on each time output, as is common in the literature on multivariate SA. As an alternative, we propose a novel method for efficient global SA of dynamic models with high‐dimensional inputs by combining a new polynomial chaos expansion (PCE)‐driven partial least squares (PLS) algorithm with the analysis of variance. PLS is used to simultaneously reduce the dimensionality of the input and output variables spaces, by identifying the input and output latent variables that account for most of their joint variability. PCE is incorporated into the PLS algorithm to capture the non‐linear behavior of the physical system. We derive the sensitivity indices associated with each output latent variable, based on which we propose generalized sensitivity indices that synthesize the influence of each input on the variance of entire output time series. All sensitivities can be computed analytically by post‐processing the coefficients of the PLS‐PCE representation. Hence, the computational cost of global SA for dynamic models essentially reduces to the cost for estimating these coefficients. We numerically compare the proposed method with existing methods by several dynamic models with high‐dimensional inputs. The results show that the PLS‐PCE method can obtain accurate sensitivity indices at low computational cost, even for models with strong interaction among the inputs.

Benchmark Investigation of SCC-DFTB against Standard and Hybrid DFT to Model Electronic Properties in Two-Dimensional MOFs for Thermoelectric Applications.

Recent studies have shown that metal-organic frameworks (MOFs) have potential as thermoelectric materials, and the topic has received increasing attention. The main motivation for this project is to further our knowledge of thermoelectric properties in MOFs and find which available self-consistent-charge density functional tight binding (SCC-DFTB) method can best predict (at least trends in) the electronic properties of MOFs at a lower computational cost than standard density functional theory (DFT). In this work, the electronic properties of monolayer, serrated, AA-stacked, and/or AB-stacked Zn3C6O6, Cd3C6O6, Zn-NH-MOF─for which no previous calculations of thermoelectric performance exist─and Ni3(HITP)2 MOFs are modeled with DFT-PBE, DFT-HSE06, GFN1-xTB, GFN2-xTB, and DFTB-3ob/mio. The band structures, density of states, and their relative orbital contributions, as well as the electrical conductivity, Seebeck coefficient, and power factor, are compared across methods and geometries. Our results suggest that GFN-xTB is adequate to predict the MOFs' band structure shape and density of states but not band gap. Our calculations further indicate that Zn3C6O6, Cd3C6O6, and Zn-NH-MOF have higher power factor values than Ni3(HITP)2, one of the highest performing synthesized MOFs, and are therefore promising for thermoelectric applications.

LiDAR-Based Intensity-Aware Outdoor 3D Object Detection.

LiDAR-based 3D object detection and localization are crucial components of autonomous navigation systems, including autonomous vehicles and mobile robots. Most existing LiDAR-based 3D object detection and localization approaches primarily use geometric or structural feature abstractions from LiDAR point clouds. However, these approaches can be susceptible to environmental noise due to adverse weather conditions or the presence of highly scattering media. In this work, we propose an intensity-aware voxel encoder for robust 3D object detection. The proposed voxel encoder generates an intensity histogram that describes the distribution of point intensities within a voxel and is used to enhance the voxel feature set. We integrate this intensity-aware encoder into an efficient single-stage voxel-based detector for 3D object detection. Experimental results obtained using the KITTI dataset show that our method achieves comparable results with respect to the state-of-the-art method for car objects in 3D detection and from a bird's-eye view and superior results for pedestrian and cyclic objects. Furthermore, our model can achieve a detection rate of 40.7 FPS during inference time, which is higher than that of the state-of-the-art methods and incurs a lower computational cost.

Strong convergence of adaptive time-stepping schemes for the stochastic Allen–Cahn equation

Abstract It is known from Beccari et al. (2019) that the standard explicit Euler-type scheme (such as the exponential Euler and the linear-implicit Euler schemes) with a uniform timestep, though computationally efficient, may diverge for the stochastic Allen–Cahn equation. To overcome the divergence, this paper proposes and analyzes adaptive time-stepping schemes, which adapt the timestep at each iteration to control numerical solutions from instability. The a priori estimates in $\mathscr{C}(\mathscr{O})$-norm and $\dot{H}^{\beta }(\mathscr{O})$-norm of numerical solutions are established provided the adaptive timestep function is suitably bounded, which plays a key role in the convergence analysis. We show that the adaptive time-stepping schemes converge strongly with order $\frac{\beta }{2}$ in time and $\frac{\beta }{d}$ in space with $d$ ($d=1,2,3$) being the dimension and $\beta \in (0,2]$. Numerical experiments show that the adaptive time-stepping schemes are simple to implement and at a lower computational cost than a scheme with the uniform timestep.

Fusion inception and transformer network for continuous estimation of finger kinematics from surface electromyography.

Decoding surface electromyography (sEMG) to recognize human movement intentions enables us to achieve stable, natural and consistent control in the field of human computer interaction (HCI). In this paper, we present a novel deep learning (DL) model, named fusion inception and transformer network (FIT), which effectively models both local and global information on sequence data by fully leveraging the capabilities of Inception and Transformer networks. In the publicly available Ninapro dataset, we selected surface EMG signals from six typical hand grasping maneuvers in 10 subjects for predicting the values of the 10 most important joint angles in the hand. Our model's performance, assessed through Pearson's correlation coefficient (PCC), root mean square error (RMSE), and R-squared (R2) metrics, was compared with temporal convolutional network (TCN), long short-term memory network (LSTM), and bidirectional encoder representation from transformers model (BERT). Additionally, we also calculate the training time and the inference time of the models. The results show that FIT is the most performant, with excellent estimation accuracy and low computational cost. Our model contributes to the development of HCI technology and has significant practical value.

Efficient second-order accurate scheme for fluid–surfactant systems on curved surfaces with unconditional energy stability

Accurately simulating the interplay between fluids and surfactants is a challenge, especially when ensuring both mass conservation and guaranteed energy stability. This study proposes a highly accurate numerical scheme for the water–oil–surfactant system coupled with the Navier–Stokes equation. We use the second-order accurate discrete operators on triangular grids representing these surfaces. We use the Crank–Nicolson method to achieve the second-order temporal accuracy. To handle high-order nonlinearities, we employ the consistency-enhanced Scalar Auxiliary Variable method. Additionally, the projection method tackles the Navier–Stokes equations. These techniques guarantee the unconditional stability. Consequently, larger time steps are permissible. Our method achieves the high accuracy with lower computational cost. This efficiency stems from solely utilizing surface information, eliminating the need for calculations across the entire 3D space. The proposed scheme’s accuracy, stability, and robustness can be confirmed through various numerical experiments.

Neural-Integrated Meshfree (NIM) Method: A differentiable programming-based hybrid solver for computational mechanics

While deep learning and data-driven modeling approaches based on deep neural networks (DNNs) have recently attracted increasing attention for solving partial differential equations, their practical applications to real-world scientific and engineering problems remain limited due to the relatively low accuracy and high computational costs. In this study, we present a differentiable programming-based hybrid meshfree approach within the field of computational mechanics. This approach seamlessly integrates neural network-based deep learning architectures with physics-based meshfree discretization techniques, thus referred to as the neural-integrated meshfree (NIM) method. To effectively approximate the solution, NIM employs a hybrid scheme called NeuroPU approximation, which combines continuous DNN representations with partition of unity (PU) basis functions associated with the underlying spatial discretization. This neural-numerical hybridization not only enhances the solution representation through functional space decomposition but also reduces both the size of DNN model and the need of automatic differentiation for spatial gradient computations, leading to a significant improvement in training efficiency. Under the NIM framework, we propose two truly meshfree solvers: the strong form-based NIM (S-NIM) and the local variational form-based NIM (V-NIM). In the S-NIM solver, the strong-form governing equation is directly considered in the loss function, while the V-NIM solver employs a local Petrov–Galerkin approach that allows the construction of variational residuals based on arbitrary overlapping subdomains. This ensures both the satisfaction of underlying physics and the preservation of the meshfree property. We perform extensive numerical experiments on both stationary and transient benchmark problems to assess the effectiveness of the proposed NIM methods in terms of accuracy, scalability, generalizability, and convergence properties. Moreover, comparative analysis with other physics-informed machine learning methods demonstrates that NIM, especially V-NIM, significantly enhances both accuracy and efficiency in end-to-end predictive capabilities.

On the inaccuracies of point-particle approach for char conversion modeling

Char conversion is a complex phenomenon that involves not only heterogeneous reactions but also external and internal heat and mass transfer. Reactor-scale simulations often use a point-particle approach (PP approach) as sub-models for char conversion because of its low computational cost. Despite a number of simplifications involved in the PP approach, there are very few studies that systematically investigate the inaccuracies of the PP approach. This study aims to compare and identify when and why the PP approach deviates from resolved-particle simulations (RP approach). Simulations have been carried out for CO2 gasification of a char particle under zone II conditions (i.e., pore diffusion control) using both PP and RP approaches. Results showed significant deviations between the two approaches for the effectiveness factor, gas compositions, particle temperature, and particle diameter. The most significant sources of inaccuracies in the PP approach are negligence of the non-uniform temperature inside the particle and the inability to accurately model external heat transfer. Under the conditions with low effectiveness factors, the errors of intra-particle processes were dominant while the errors of external processes became dominant when effectiveness factors were close to unity. Because it assumes uniform internal temperature, the models applying the PP approach always predict higher effectiveness factors than the RP approach, despite its accurate estimation of intra-particle mass diffusion effects. As a consequence, the PP approach failed to predict the particle size changes accurately. Meanwhile, no conventional term for external heat transfer could explain the inaccuracy, indicating the importance of other sources of errors such as 2D/3D asymmetry or penetration of external flows inside the particles.

Estimating genomic relationships of metafounders across and within breeds using maximum likelihood, pseudo-expectation–maximization maximum likelihood and increase of relationships

BackgroundThe theory of “metafounders” proposes a unified framework for relationships across base populations within breeds (e.g. unknown parent groups), and base populations across breeds (crosses) together with a sensible compatibility with genomic relationships. Considering metafounders might be advantageous in pedigree best linear unbiased prediction (BLUP) or single-step genomic BLUP. Existing methods to estimate relationships across metafounders Γ\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$${\\varvec{\\Gamma}}$$\\end{document} are not well adapted to highly unbalanced data, genotyped individuals far from base populations, or many unknown parent groups (within breed per year of birth).MethodsWe derive likelihood methods to estimate Γ\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$${\\varvec{\\Gamma}}$$\\end{document}. For a single metafounder, summary statistics of pedigree and genomic relationships allow deriving a cubic equation with the real root being the maximum likelihood (ML) estimate of Γ\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$${\\varvec{\\Gamma}}$$\\end{document}. This equation is tested with Lacaune sheep data. For several metafounders, we split the first derivative of the complete likelihood in a term related to Γ\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$${\\varvec{\\Gamma}}$$\\end{document}, and a second term related to Mendelian sampling variances. Approximating the first derivative by its first term results in a pseudo-EM algorithm that iteratively updates the estimate of Γ\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$${\\varvec{\\Gamma}}$$\\end{document} by the corresponding block of the H-matrix. The method extends to complex situations with groups defined by year of birth, modelling the increase of Γ\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$${\\varvec{\\Gamma}}$$\\end{document} using estimates of the rate of increase of inbreeding (ΔF\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\Delta F$$\\end{document}), resulting in an expanded Γ\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$${\\varvec{\\Gamma}}$$\\end{document} and in a pseudo-EM+ΔF\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\Delta F$$\\end{document} algorithm. We compare these methods with the generalized least squares (GLS) method using simulated data: complex crosses of two breeds in equal or unsymmetrical proportions; and in two breeds, with 10 groups per year of birth within breed. We simulate genotyping in all generations or in the last ones.ResultsFor a single metafounder, the ML estimates of the Lacaune data corresponded to the maximum. For simulated data, when genotypes were spread across all generations, both GLS and pseudo-EM(+ΔF\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\Delta F$$\\end{document}) methods were accurate. With genotypes only available in the most recent generations, the GLS method was biased, whereas the pseudo-EM(+ΔF\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\Delta F$$\\end{document}) approach yielded more accurate and unbiased estimates.ConclusionsWe derived ML, pseudo-EM and pseudo-EM+ΔF\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\Delta F$$\\end{document} methods to estimate Γ\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$${\\varvec{\\Gamma}}$$\\end{document} in many realistic settings. Estimates are accurate in real and simulated data and have a low computational cost.

SnakeNet: An adaptive network for small object and complex background for insulator surface defect detection

Aiming at the current problems of small-target defect detection difficulty, weak resistance to complex background interference, and low detection precision in insulator defect detection, an insulator defect detection model, SnakeNet, is proposed based on the YOLOv8 detection framework. Firstly, the Dynamic Snake Convolutional Neck (DSConv-Neck) is proposed to extract rich fine-grained features and improve the detection ability of small-target defects. Then, the background interference is suppressed by the C2f Efficient Multi-Scale Attention (C2MA) module to enhance the anti-background resistance; finally, the GhostConv VanillaNet (GCVNet) module is proposed to minimize the number of model parameters while ensuring the detection precision. Many experiments are carried out on two public datasets, the China power line insulator dataset (CPLID) and the Insulator Defect Image Dataset (IDID). The results show that the optimal detection precision and recall of SnakeNet are 98.7% and 98.8%, respectively, and the number of model parameters and computation volume are only 2.9M and 6.7G. Compared with the original YOLOv8, the detection precision and recall are improved by 2.4% and 3.5%, respectively, and the number of model parameters and computation volume are reduced by 0.3M and 1.3M, respectively. Compared with other algorithms, such as SSD and RT-DETR, SnakeNet performs well in detecting small targets and adapting to complex backgrounds with lower computational costs.

A once-through artificial neural network approach for used nuclear fuel inverse depletion analysis: A comparative study

This work addresses the challenge of identifying the origins and burnup history of used nuclear fuel, which is crucial for safeguarding and non-proliferation analysis. It presents a comparative evaluation of two distinct computational approaches to bundles at the same in-core position: the first entails building a forward Artificial Neural Networks (ANN) surrogate model for fuel inventory prediction followed by solving the inverse depletion problem using an optimization solver. The second approach is to build an inverse ANN model that relates the used nuclear fuel inventory to its initial and burnup conditions (a once-through approach), eliminating the need for the inverse optimization solver step which results in better accuracy, less computational costs, and renders the application of such approach practical. KENO VI depletion model is utilized to deplete a VVER assembly model generating the necessary datasets for the ANN inverse model training and validation. The inverse model is then employed to estimate the initial composition and the burnup for several test cases. The results are then compared to the known initial and burnup conditions as well as to the estimations made by benchmark results from an existing ANN forward surrogate model combined with the Particle Swarm Optimization (PSO) solver. Comparative analyses across eight scenarios consistently reveal that the once-through ANN inverse model has superior accuracy and consistency in terms of estimating the actual burnup and initial fuel enrichment with a relative root mean squared error (RMSE) of 1.20×10-2 compared to the benchmark approach (ANN-PSO) which yielded a relative RMSE of 2.18×10-2. Moreover, the once-through inverse approach has a significantly lower computational cost as it requires 1 inverse model run compared to the 9,450,000 model runs required by the benchmark approach (ANN-PSO). Overall, the results demonstrate that the proposed once-through approach simplifies and enhances the efficiency of solving the inverse fuel depletion problem while outperforming the benchmark approach in terms of predictive accuracy computational cost.

ST-TransNet: A Spatiotemporal Transformer Network for Uncertainty Estimation from a Single Deterministic Precipitation Forecast

Abstract The forecast uncertainty, particularly for precipitation, serves as a crucial indicator of the reliability of deterministic forecasts. Traditionally, forecast uncertainty is estimated by ensemble forecasting, which is computationally expensive since the forecast model is run multiple times with perturbations. Recently, deep learning methods have been explored to learn the statistical properties of ensemble prediction systems due to their low computational costs. However, accurately and effectively capturing the uncertainty information in precipitation forecasts remains challenging. In this study, we present a novel spatiotemporal transformer network (ST-TransNet) as an alternative approach to estimate uncertainty with ensemble spread and probabilistic forecasts, by learning from historical ensemble forecasts. ST-TransNet features a hierarchical structure for extracting multiscale features and incorporates a spatiotemporal transformer module with window-based attention to capture correlations in both spatial and temporal dimensions. Additionally, window-based attention can not only extract local precipitation patterns but also reduce computational costs. The proposed ST-TransNet is evaluated on the TIGGE ensemble forecast dataset and Global Precipitation Measurement (GPM) precipitation products. Results show that ST-TransNet outperforms both traditional and deep learning methods across various metrics. Case studies further demonstrate its ability to generate reasonable and accurate spread and probability forecasts from a single deterministic precipitation forecast. It demonstrates the capacity and efficiency of neural networks in estimating precipitation forecast uncertainty.