Computational Steps Research Articles

A simple approach for the computation of magnetic characteristics of 4-phase switched reluctance motor

Switched reluctance motor (SRM) is non-linear in nature due to non-uniform air gap between stator and rotor that is produced from doubly salient poles. To control the operation of non-linear motor, a precise modelling of magnetic characteristic is essential. Traditionally magnetic equivalent circuit, finite element method (FEM) and experimental approaches have been used to derive magnetic characteristics of SRM. These methods involve complex computational steps and numerous assumptions for its calculations. To avoid this complexity, a simple approach is suggested in this paper to compute magnetic characteristics of 4-phase SRM. In this approach, SRM phase inductance is measured directly to identify the location of rotor with respect to stator from unaligned to aligned position. Then, single phase of SRM is excited through asymmetric converter and voltage and currents of corresponding phase is recorded with rotor being blocked using indexing head. Later recorded values are used to compute the magnetic characteristics from unaligned to aligned position. The proposed method is carried out using field programmable gate array (FPGA) controlled 8/6, 4-phase SRM. Further, accuracy of obtained magnetic characteristic is verified using FEM. In addition, fidelity of the magnetic characteristic is also validated by developing a dynamic model of SRM in MATLAB/Simulink

MgMViT: Multi-Granularity and Multi-Scale Vision Transformer for Efficient Action Recognition

Nowadays, the field of video-based action recognition is rapidly developing. Although Vision Transformers (ViT) have made great progress in static image processing, they are not yet fully optimized for dynamic video applications. Convolutional Neural Networks (CNN) and related models perform exceptionally well in video action recognition. However, there are still some issues that cannot be ignored, such as high computational costs and large memory consumption. In the face of these issues, current research focuses on finding effective methods to improve model performance and overcome current limits. Therefore, we present a unique Vision Transformer model based on multi-granularity and multi-scale fusion to accomplish efficient action recognition, which is designed for action recognition in videos to effectively reduce computational costs and memory usage. Firstly, we devise a multi-scale, multi-granularity module that integrates with Transformer blocks. Secondly, a hierarchical structure is utilized to manage information at various scales, and we introduce multi-granularity on top of multi-scale, which allows for a selective choice of the number of tokens to enter the next computational step, thereby reducing redundant tokens. Thirdly, a coarse-fine granularity fusion layer is introduced to reduce the sequence length of tokens with lower information content. The above two mechanisms are combined to optimize the allocation of resources in the model, further emphasizing critical information and reducing redundancy, thereby minimizing computational costs. To assess our proposed approach, comprehensive experiments are conducted by using benchmark datasets in the action recognition domain. The experimental results demonstrate that our method has achieved state-of-the-art performance in terms of accuracy and efficiency.

A regularized orthogonal activated inverse-learning neural network for regression and classification with outliers

A novel regularized orthogonal activated inverse-learning (ROAIL) neural network is proposed and investigated for reducing the impact of outliers in regression and classification fields. The proposed ROAIL network does not require extensive iterative computations. Instead, it can achieve the desired results with a single step of computation, allowing for the efficient acquisition of network weights. By extending the Gegenbauer polynomials to a multi-variate version, and integrating the ℓ2 regularization and Welsch loss function into the orthogonal activated inverse-learning framework, two forms of ROAIL are obtained, i.e., ℓ2 norm ROAIL (ℓ2-ROAIL) and Welsch-ROAIL (W-ROAIL). ℓ2-ROAIL neural network is proposed to minimize the empirical and structural risk simultaneously since taking the structural risk as a part of loss function can effectively reduce the complexity of the model and thus improve the generalization ability. W-ROAIL neural network further improves the robustness of the ℓ2-ROAIL neural network by replacing the original two-norm in loss function with Welsch function. The Welsch function can determine the weights of each sample according to its output error, and influence of outliers could be weakened since the weights of outliers would be reduced. Both regression and classification experiments show that W-ROAIL neural network has strong ability to suppress the influence of outliers.

Skeleton-based Human Activity Classification in Sparse Image Sequences

Research results on human activity classification in video are described, based on initial human skeleton estimation in video frames. Both single person actions and two-person interactions are considered. The initial skeleton data is estimated in selected video frames by OpenPose, HRNet or other dedicated library. Important contributions of presented work are computational steps of skeleton tracking and -refinement, and relational feature extraction from pairs of skeleton joints. It is shown, that this feature engineering significantly increases the classification accuracy. Regarding the final neural network encoder-classifier, two different architectures are designed and tested. The first solution is a lightweight MLP network, implementing the idea of a "mixture of pose experts". Several pose classifiers (experts) are trained independently on different time periods (snapshots) of single-person visual actions (or 2-person interactions), while the final classification is a time-related pooling of weighted expert classifications. All pose experts use the same deep encoding network. The second (middle weight) solution is based on a LSTM network.Both solutions are trained and tested on the action set of the well-known NTU RGB+D dataset, although only 2D data are used.Our results show comparable performance with some of the best reported STM- and CNN-based classifiers for this dataset. We conclude that by reducing the noise of skeleton data, highly successful lightweight- and midweight-approaches to visual activity recognition in image sequences can be achieved.

RUBICON: a framework for designing efficient deep learning-based genomic basecallers

Nanopore sequencing generates noisy electrical signals that need to be converted into a standard string of DNA nucleotide bases using a computational step called basecalling. The performance of basecalling has critical implications for all later steps in genome analysis. Therefore, there is a need to reduce the computation and memory cost of basecalling while maintaining accuracy. We present RUBICON, a framework to develop efficient hardware-optimized basecallers. We demonstrate the effectiveness of RUBICON by developing RUBICALL, the first hardware-optimized mixed-precision basecaller that performs efficient basecalling, outperforming the state-of-the-art basecallers. We believe RUBICON offers a promising path to develop future hardware-optimized basecallers.

"Now I Get It!": Eureka Experiences During the Acquisition of Mathematical Concepts.

Many famous scientists have reported anecdotes where a new understanding occurred to them suddenly, in an unexpected flash. Do people generally experience such "Eureka" moments when learning science concepts? And if so, do these episodes truly vehicle sudden insights, or is this impression illusory? To address these questions, we developed a paradigm where participants were taught the mathematical concept of geodesic, which generalizes the common notion of straight line to straight trajectories drawn on curved surfaces. After studying lessons introducing this concept on the sphere, participants (N = 56) were tested on their understanding of geodesics on the sphere and on other surfaces. Our findings indicate that Eureka experiences are common when learning mathematics, with reports by 34 (61%) participants. Moreover, Eureka experiences proved an accurate description of participants' learning, in two respects. First, Eureka experiences were associated with learning and generalization: the participants who reported experiencing Eurekas performed better at identifying counterintuitive geodesics on new surfaces. Second, and in line with the firstperson experience of a sudden insight, our findings suggest that the learning mechanisms responsible for Eureka experiences are inaccessible to reflective introspection. Specifically, reports of Eureka experiences and of participants' confidence in their own understanding were associated with different profiles of performance, indicating that the mechanisms bringing about Eureka experiences and those informing reflective confidence were at least partially dissociated. Learning mathematical concepts thus appears to involve mechanisms that operate unconsciously, except when a key computational step is reached and a sudden insight breaks into consciousness.

Efficiency Improvisation of Large-Scale Knowledge Systems in Feature Determination using Proposed HVGAN Architecture

Knowledge Management (KM) is crucial for efficient information retrieval and forms the backbone of implementing an Information System (IS). Efficiently extracting meaningful trends from headlines aids in visualising demographic situations and addressing emergency handling needs. Understanding Information Resource Management (IRM) and Knowledge Resource Management (KRM) concepts benefits end users. This proposed KRM concept organises knowledge, ensuring global resource control through a neural decision framework. Traditional methods like Structural Equation Modelling (SEM) have higher computational steps, whereas neural network strategies in KM reduce steps and improve data prediction accuracy. The paucity of input data is a common challenge, which can be overcome by using Generative Adversarial Network (GAN) architecture for probabilistic generative processes. This research evaluates acquisition, spreading, vulnerability, and application using GAN-based knowledge systems that mitigate data scarcity through generative methods. The proposed Hierarchical Vanilla Generative Adversarial Network (HVGAN) utilises gradient functions and hierarchical computation algorithms to achieve desired knowledge extraction with lower time complexity and higher accuracy. Implementation improves efficiency in accessing stored knowledge, yielding an AUC score of 96% compared to other GAN architectures.

Time-Efficient Constant-Space-Overhead Fault-Tolerant Quantum Computation

Scaling up quantum computers to attain substantial speedups over classical computing requires fault tolerance. Conventionally, protocols for fault-tolerant quantum computation demand excessive space overheads by using many physical qubits for each logical qubit. A more recent protocol using quantum analogues of low-density parity-check codes needs only a constant space overhead that does not grow with the number of logical qubits. However, the overhead in the processing time required to implement this protocol grows polynomially with the number of computational steps. To address these problems, here we introduce an alternative approach to constant-space-overhead fault-tolerant quantum computing using a concatenation of multiple small-size quantum codes rather than a single large-size quantum low-density parity-check code. We develop techniques for concatenating different quantum Hamming codes with growing size. As a result, we construct a low-overhead protocol to achieve constant space overhead and only quasi-polylogarithmic time overhead simultaneously. Our protocol is fault tolerant even if a decoder has a non-constant runtime, unlike the existing constant-space-overhead protocol. This code concatenation approach will make possible a large class of quantum speedups with feasibly bounded space overhead yet negligibly short time overhead.

A modified RI-IGABEM with only weakly singular integral for viscoelastic analysis

In this paper, an improved isogeometric boundary element method based on radial integration method (RI-IGABEM) is proposed to solve the viscoelastic problems. Prony series are utilized to discretize the stress with the two most recent time-step strains. The boundary integral equation (BIE) for the viscoelastic problems is derived. The domain integral related to stress is decomposed into the linear combination of two sub-integrals by volumetric–deviatoric split. The radial integration method (RIM) is applied to transform all the domain integrals into equivalent boundary integrals. The present method has several advantages. Firstly, the present method eliminates the need for stress calculation at each time step, thus avoiding the error accumulation associated with the traction recovery method (TRM). Secondly, the BIE for the viscoelastic problems contains only weakly singular integrals. In addition, we provide an analytical solution for the radial component of the equivalent boundary integrals, which is transformed using RIM. To sum up, the present method improves the accuracy and efficiency of RI-IGABEM by reducing the required computational steps, eliminating error accumulation, and providing analytical solutions for certain integrals. Numerical examples demonstrate that the proposed RI-IGABEM is an efficient and accurate method suitable for viscoelastic problems.

Fast data fitting scheme for compressive multispectral fluorescence lifetime imaging.

A single-pixel camera combined with compressive sensing techniques is a promising fluorescence microscope scheme for acquiring a multidimensional dataset (space, spectrum, and lifetime) and for reducing the measurement time with respect to conventional microscope schemes. However, upon completing the acquisition, a computational step is necessary for image reconstruction and data analysis, which can be time-consuming, potentially canceling out the beneficial effect of compressive sensing. In this work, we propose and experimentally validate a fast-fit workflow based on global analysis and multiple linear fits, which significantly reduces the computation time from tens of minutes to less than 1 s. Moreover, as the method is interlaced with the measurement flow, it can be applied in parallel with the acquisitions.

Software Component Selection: An Optimized Selection Criterion for Component-based Software Engineering (CBSE)

Component-Based Software Engineering (CBSE) is gaining popularity in software development due to time and cost limitations. As software applications have become integral to people's lives, developing high-quality and user-friendly applications within a reasonable timeframe and budget has become increasingly challenging. Software development firms frequently use Commercial-Off-The-Shelf (COTS) components to address this challenge and reduce development costs and time. However, selecting appropriate components that meet customer requirements and integrate seamlessly with the target system is a complex task requiring considering the entire software system's quality. This study investigates the critical factors that software industry practitioners and experts must consider when selecting software components. First, the author asked practitioners to identify the most important quality criteria for an online bookstore from a list using subjective judgment and evaluation grades. Then, the study employed the Evidential Reasoning (ER) approach to tackle the multi-level evaluations and information uncertainty associated with software component selection issues. The ER approach's primary features, such as weight normalization, probability assessment, uncertainty management, and utility intervals, offer several benefits for COTS selection problems, including cost and time reduction, improved software reliability, effectiveness, and efficiency. This study assessed the quality criteria for an online bookstore using the ER approach and provided analysis results based on the approach's computational steps. Finally, the study ranked the four components according to their weights, evaluation grades, and belief degrees for selection

Secure and Efficient Federated Learning Against Model Poisoning Attacks in Horizontal and Vertical Data Partitioning.

In distributed systems, data may partially overlap in sample and feature spaces, that is, horizontal and vertical data partitioning. By combining horizontal and vertical federated learning (FL), hybrid FL emerges as a promising solution to simultaneously deal with data overlapping in both sample and feature spaces. Due to its decentralized nature, hybrid FL is vulnerable to model poisoning attacks, where malicious devices corrupt the global model by sending crafted model updates to the server. Existing work usually analyzes the statistical characteristics of all updates to resist model poisoning attacks. However, training local models in hybrid FL requires additional communication and computation steps, increasing the detection cost. In addition, due to data diversity in hybrid FL, solutions based on the assumption that malicious models are distinct from honest models may incorrectly classify honest ones as malicious, resulting in low accuracy. To this end, we propose a secure and efficient hybrid FL against model poisoning attacks. Specifically, we first identify two attacks to define how attackers manipulate local models in a harmful yet covert way. Then, we analyze the execution time and energy consumption in hybrid FL. Based on the analysis, we formulate an optimization problem to minimize training costs while guaranteeing accuracy considering the effect of attacks. To solve the formulated problem, we transform it into a Markov decision process and model it as a multiagent reinforcement learning (MARL) problem. Then, we propose a malicious device detection (MDD) method based on MARL to select honest devices to participate in training and improve efficiency. In addition, we propose an alternative poisoned model detection (PMD) method considering model change consistency. This method aims to prevent poisoned models from being used in the model aggregation. Experimental results validate that under the random local model poisoning attack, the proposed MDD method can save over 50% training costs while guaranteeing accuracy. When facing the advanced adaptive local model poisoning (ALMP) attack, utilizing both the proposed MDD and PMD methods achieves the desired accuracy while reducing execution time and energy consumption.

Efficient Deep Spiking Multilayer Perceptrons With Multiplication-Free Inference.

Advancements in adapting deep convolution architectures for spiking neural networks (SNNs) have significantly enhanced image classification performance and reduced computational burdens. However, the inability of multiplication-free inference (MFI) to align with attention and transformer mechanisms, which are critical to superior performance on high-resolution vision tasks, imposes limitations on these gains. To address this, our research explores a new pathway, drawing inspiration from the progress made in multilayer perceptrons (MLPs). We propose an innovative spiking MLP architecture that uses batch normalization (BN) to retain MFI compatibility and introduce a spiking patch encoding (SPE) layer to enhance local feature extraction capabilities. As a result, we establish an efficient multistage spiking MLP network that blends effectively global receptive fields with local feature extraction for comprehensive spike-based computation. Without relying on pretraining or sophisticated SNN training techniques, our network secures a top-one accuracy of 66.39% on the ImageNet-1K dataset, surpassing the directly trained spiking ResNet-34 by 2.67%. Furthermore, we curtail computational costs, model parameters, and simulation steps. An expanded version of our network compares with the performance of the spiking VGG-16 network with a 71.64% top-one accuracy, all while operating with a model capacity 2.1 times smaller. Our findings highlight the potential of our deep SNN architecture in effectively integrating global and local learning abilities. Interestingly, the trained receptive field in our network mirrors the activity patterns of cortical cells.

Applying AI and fuzzy modeling to optimize advertising effectiveness in digital marketing strategies for the energy sector

The purpose of the study is to formulate the main aspects of the use of artificial intelligence in advertising and determine its effectiveness in the energy sector. The intensity of development of artificial intelligence technologies and its application to ensure the effectiveness of advertising activities of companies in the energy sector is emphasized. The ethical and practical aspects of using artificial intelligence in advertising are presented. An analysis of the efficiency factors of the advertising process was carried out based on the classification of tools and methods of advertising automation using artificial intelligence technologies. Practical recommendations for the use of neural networks based on artificial intelligence have been formed. The provisions of the classical and meta-heurastic methods of parametric identification of a fuzzy system, which is a model of the marketing process of companies in the energy sector, are substantiated. The effectiveness of advertising was determined using the methodology of modeling fuzzy systems and the sequence of computational steps of a fuzzy neural system. The practical significance of the results lies in the fact that metaheuristic methods make it possible to assess the effectiveness of advertising based on the use of artificial intelligence tools by companies in the energy sector.

NVS-GAN: Benefit of generative adversarial network on novel view synthesis

The methodology to generate new views for an object from provided input object view is called Novel View Synthesis (NVS). Humans imagine novel views through prior knowledge gathered through their lifetime. NVS-GAN predicts the novel views through computation. Literature survey reveals that there are limited NVS models with low Trainable Parameter Count (TPC) and low model size. Also, a study on the effect of different loss functions on NVS models was lacking. Lowering the TPC indicates less computational steps for the model to predict the output, therefore desirable. Combined with a low model size, the proposed model will become more suitable for deployment in diverse devices having limited resources for computation. Application of right combination of loss functions yield better accuracy. To address these research gaps, NVS-GAN is proposed. NVS-GAN is a Generative Adversarial Network (GAN) approach which yields NVS-Generator which performs NVS. NVS-Generator incorporates identity skip connections, bilinear sampling module, Depthwise Separable Convolution (DSC) as design features and results in low TPC, model size. In addition to discriminator loss, NVS-GAN is trained with different combinations of loss functions i.e. Mean Absolute Error (MAE) loss, Structural Similarity Index Measure (SSIM) loss, Huber loss on chair and car objects of ShapeNet dataset. The performance of NVS-Generator on test set measured in terms of MAE and SSIM is tabulated and analysed. The performance is compared with existing NVS models. The proposed NVS-GAN experiment recorded reduction in NVS-Generator TPC in 37 %–54.6 % range and reduction in model size between 37.2 % and 47.6 % range. NVS-Generator reduced MAE upto 55 % and improved SSIM upto 4 % than existing models. Summarily, NVS-GAN increased model performance and made the model “lightweight”.

Economical quasi-Newton unitary optimization of electronic orbitals.

We present an efficient quasi-Newton orbital solver optimized to reduce the number of gradient evaluations and other computational steps of comparable cost. The solver optimizes orthogonal orbitals by sequences of unitary rotations generated by the (preconditioned) limited-memory Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) algorithm equipped with trust-region step restriction. The low-rank structure of the L-BFGS inverse Hessian is exploited when solving the trust-region problem. The efficiency of the proposed "Quasi-Newton Unitary Optimization with Trust-Region" (QUOTR) solver is compared to that of the standard Roothaan-Hall approach accelerated by the Direct Inversion of Iterative Subspace (DIIS), and other exact and approximate Newton solvers for mean-field (Hartree-Fock and Kohn-Sham) problems.

AmiR-P3: An AI-based microRNA prediction pipeline in plants.

MicroRNAs (miRNAs) are small noncoding RNAs that play important post-transcriptional regulatory roles in animals and plants. Despite the importance of plant miRNAs, the inherent complexity of miRNA biogenesis in plants hampers the application of standard miRNA prediction tools, which are often optimized for animal sequences. Therefore, computational approaches to predict putative miRNAs (merely) from genomic sequences, regardless of their expression levels or tissue specificity, are of great interest. Here, we present AmiR-P3, a novel ab initio plant miRNA prediction pipeline that leverages the strengths of various utilities for its key computational steps. Users can readily adjust the prediction criteria based on the state-of-the-art biological knowledge of plant miRNA properties. The pipeline starts with finding the potential homologs of the known plant miRNAs in the input sequence(s) and ensures that they do not overlap with protein-coding regions. Then, by computing the secondary structure of the presumed RNA sequence based on the minimum free energy, a deep learning classification model is employed to predict potential pre-miRNA structures. Finally, a set of criteria is used to select the most likely miRNAs from the set of predicted miRNAs. We show that our method yields acceptable predictions in a variety of plant species. AmiR-P3 does not (necessarily) require sequencing reads and/or assembled reference genomes, enabling it to identify conserved and novel putative miRNAs from any genomic or transcriptomic sequence. Therefore, AmiR-P3 is suitable for miRNA prediction even in less-studied plants, as it does not require any prior knowledge of the miRNA repertoire of the organism. AmiR-P3 is provided as a docker container, which is a portable and self-contained software package that can be readily installed and run on any platform and is freely available for non-commercial use from: https://hub.docker.com/r/micrornaproject/amir-p3.

Detection of poikilocytosis cells in anemia using (ANN) & embedded systems

In the bloodstream, erythrocytes stand as vital carriers of oxygen and carbon dioxide, facilitated by the presence of hemoglobin within their structures. However, deviations in erythrocyte size can lead to the formation of Poikilocyte cells, a characteristic feature of conditions like Iron Deficiency Anemia. Variants of Poikilocytoses, such as Degmacyte, Dacrocyte, Schistocyte, and Elliptocyte, denote distinct alterations in erythrocyte morphology, often associated with diminished iron levels crucial for haemoglobin synthesis. In a recent study, the differentiation between normal RBCs and Poikilocyte cells has been addressed through the application of Artificial Neural Network (ANN) algorithms, leveraging extracted features from digital images of blood smears. This approach offers a more precise means of identifying blood disorders compared to traditional visual inspection, utilizing image analysis techniques to detect deviations in color, size, and statistical parameters. The methodology involves a series of computational steps including preprocessing, segmentation, morphological operations, feature extraction, and classification, all executed within the Matlab environment. Furthermore, to enhance diagnostic capabilities, the system integrates glucose level measurement alongside erythrocyte analysis, transmitting data to a controller which relays results via GSM signal as SMS and LCD display. This comprehensive approach not only automates cell identification and classification but also ensures efficient and accurate analysis, including the automated separation of overlapped cells.

An Adomian decomposition method with some orthogonal polynomials to solve nonhomogeneous fractional differential equations (FDEs)

<p>The present study introduced modifications to the standard Adomian decomposition method (ADM) by combining the Taylor series with orthogonal polynomials, such as Legendre polynomials and the first and second kinds of Chebyshev polynomials. These improvements can be applied to solve fractional differential equations with initial-value problems in the Caputo sense. The approaches are based on the use of orthogonal polynomials, which are essential components in approximation theories. The study carefully analyzed their respective absolute error differences, highlighting the computational benefits of the proposed modifications, which offer improved accuracy and require fewer computational steps. The effectiveness and accuracy of the approach were validated through numerical examples, confirming its efficiency and reliability.</p>

Grey relational analysis-driven optimization of neutrosophic MAGDM for cloud SaaS credibility evaluation

In the basic cloud service delivery model, since SaaS (Software as a Service) is located at the application layer, if a relatively complete cloud service credibility comprehensive evaluation framework can be established from the perspective of SaaS cloud service consumers and combined with their trust context, it can not only effectively solve the trust problem between cloud service consumers (CSC) and cloud service provider (CSP), but also has very important practical significance for the popularization and promotion of SaaS service applications. The credibility evaluation of SaaS services in cloud computing environment is often considered as a multi-attribute group decision making (MAGDM) problem. In this paper, the grey relational analysis (GRA) method is extended to the single-valued neutrosophic sets (SVNSs) setting to deal with MAGDM with incomplete weight information. First, the SVNSs are reviewed. In addition, the single-valued neutrosophic number GRA (SVNN-GRA) is established for MAGDM with incomplete weight information, and the computational steps for all designs are listed. Finally, the credibility evaluation of SaaS services in cloud computing environment is given to demonstrate the SVNN-GRA model and some comparative analysis is done to demonstrate the SVNN-GRA.