Efficient Computational Model Research Articles

Classification of Diabetic Retinopathy Based on Efficient Computational Modeling

Convolutional neural networks (CNN) and Vision Transformers (ViT) have long been the main backbone networks for visual classification in the field of deep learning. Although ViT has recently received more attention than CNN due to its excellent fitting ability, their scalability is largely limited by the quadratic complexity of attention computation. For the determination of diabetic retinopathy, the fundus lesions as well as the width, angle, and branching pattern of retinal blood vessels are characterized, inspired by the ability of Mamba and VMamba to efficiently model long sequences, VMamba-m is proposed in this paper. This is a generalized visual skeleton model designed to reduce computational complexity to linear while retaining the advantageous features of ViTs. By modifying the cross-entropy loss function, we enhance the model’s attention to rare categories, especially in large-scale multi-category classification tasks. In order to enhance the adaptability of the VMamba-m model in processing visual data, we introduce the se channel attention mechanism, which enables the model to learn features in the channel dimension and form the importance of each channel. Finally, different weights are assigned to each channel through the incentive part. In addition to this, this paper further improves the implementation details and architectural design by introducing a novel attention mechanism implemented based on the local windowing method, which aims to optimize the model’s ability in processing long sequence data to enhance the performance of VMamba-m and improve its inference speed. Extensive experimental results show that VMamba-m performs well in the retinopathy V classification task, and it has significant advantages in terms of accuracy and computation time over existing benchmark models.

Efficient computational model for particle capture by a heterogeneous filter

Efficient computational model for particle capture by a heterogeneous filter

Study on the 2D equivalent nonlinear dynamics simulation model related to high speed precision bearing and spindle

The stiffness and contact stress of the bearing spindle system are the key factors affecting its machining accuracy and service life under the condition of high-speed service, which will be affected by different structural parameters and service conditions. It is essential to establish an efficient and accurate computational simulation model to understand the influence of complex factors on the nonlinear dynamic characteristics of the bearing spindle system. Firstly, in the paper, a 2D axisymmetric finite element model of the bearing, aims at the relationship between stiffness and contact stress of the bearing under high-speed service, has been built based on a classical dynamic analysis model and the reversed method of material parameters of equivalent rolling balls. Additionally, for the BT30 spindle, the 2D axisymmetric finite element model of the bearing spindle system also has been built and applied in mechanical analysis of spindle under different conditions of assembly and service, based on the bearing model. The results show that the axial force of bearings decreases as the rotational speed increases, and an augmentation in speed will result in a reduction in the axial stiffness of the BT30 spindle. In addition, the maximum contact stress exhibits a slight decline as the rotational speed increases. Furthermore, with an escalating preload, the stiffness and contact stress of the spindle undergo substantial increments, however, these parameters will cease to alter once a certain threshold is reached.

STNMDA: A Novel Model for Predicting Potential Microbe-Drug Associations with Structure-Aware Transformer

Introduction: Microbes are intimately involved in the physiological and pathological processes of numerous diseases. There is a critical need for new drugs to combat microbe-induced diseases in clinical settings. Predicting potential microbe-drug associations is, therefore, essential for both disease treatment and novel drug discovery. However, it is costly and time-consuming to verify these relationships through traditional wet lab approaches. Methods: We proposed an efficient computational model, STNMDA, that integrated a StructureAware Transformer (SAT) with a Deep Neural Network (DNN) classifier to infer latent microbedrug associations. The STNMDA began with a “random walk with a restart” approach to construct a heterogeneous network using Gaussian kernel similarity and functional similarity measures for microorganisms and drugs. This heterogeneous network was then fed into the SAT to extract attribute features and graph structures for each drug and microbe node. Finally, the DNN classifier calculated the probability of associations between microbes and drugs. Results: Extensive experimental results showed that STNMDA surpassed existing state-of-the-art models in performance on the MDAD and aBiofilm databases. In addition, the feasibility of STNMDA in confirming associations between microbes and drugs was demonstrated through case validations. Conclusion: Hence, STNMDA showed promise as a valuable tool for future prediction of microbedrug associations.

Computational design of mechanical metamaterials.

In the past few years, design of mechanical metamaterials has been empowered by computational tools that have allowed the community to overcome limitations of human intuition. By leveraging efficient optimization algorithms and computational physics models, it is now possible to explore vast design spaces, achieving new material functionalities with unprecedented performance. Here, we present our viewpoint on the state of the art of computational metamaterials design, discussing recent advances in topology optimization and machine learning design with respect to challenges in additive manufacturing.

GraphPro: An interpretable graph neural network-based model for identifying promoters in multiple species

Promoters are DNA sequences that bind with RNA polymerase to initiate transcription, regulating this process through interactions with transcription factors. Accurate identification of promoters is crucial for understanding gene expression regulation mechanisms and developing therapeutic approaches for various diseases. However, experimental techniques for promoter identification are often expensive, time-consuming, and inefficient, necessitating the development of accurate and efficient computational models for this task. Enhancing the model's ability to recognize promoters across multiple species and improving its interpretability pose significant challenges. In this study, we introduce a novel interpretable model based on graph neural networks, named GraphPro, for multi-species promoter identification. Initially, we encode the sequences using k-tuple nucleotide frequency pattern, dinucleotide physicochemical properties, and dna2vec. Subsequently, we construct two feature extraction modules based on convolutional neural networks and graph neural networks. These modules aim to extract specific motifs from the promoters, learn their dependencies, and capture the underlying structural features of the promoters, providing a more comprehensive representation. Finally, a fully connected neural network predicts whether the input sequence is a promoter. We conducted extensive experiments on promoter datasets from eight species, including Human, Mouse, and Escherichia coli. The experimental results show that the average Sn, Sp, Acc and MCC values of GraphPro are 0.9123, 0.9482, 0.8840 and 0.7984, respectively. Compared with previous promoter identification methods, GraphPro not only achieves better recognition accuracy on multiple species, but also outperforms all previous methods in cross-species prediction ability. Furthermore, by visualizing GraphPro's decision process and analyzing the sequences matching the transcription factor binding motifs captured by the model, we validate its significant advantages in biological interpretability. The source code for GraphPro is available at https://github.com/liuliwei1980/GraphPro.

An efficient computational model for unequal-area dynamic facility layout problems considering input/output locations under algorithms case

An efficient computational model for unequal-area dynamic facility layout problems considering input/output locations under algorithms case

Review and Comparative Analysis of Methods and Advancements in Predicting Protein Complex Structure.

Protein complexes perform diverse biological functions, and obtaining their three-dimensional structure is critical to understanding and grasping their functions. In many cases, it's not just two proteins interacting to form a dimer; instead, multiple proteins interact to form a multimer. Experimentally resolving protein complex structures can be quite challenging. Recently, there have been efforts and methods that build upon prior predictions of dimer structures to attempt to predict multimer structures. However, in comparison to monomeric protein structure prediction, the accuracy of protein complex structure prediction remains relatively low. This paper provides an overview of recent advancements in efficient computational models for predicting protein complex structures. We introduce protein-protein docking methods in detail and summarize their main ideas, applicable modes, and related information. To enhance prediction accuracy, other critical protein-related information is also integrated, such as predicting interchain residue contact, utilizing experimental data like cryo-EM experiments, and considering protein interactions and non-interactions. In addition, we comprehensively reviewcomputational approaches for end-to-end prediction ofprotein complex structuresbased on artificial intelligence (AI) technologyand describe commonly used datasets and representative evaluation metrics in protein complexes. Finally, we analyze the formidable challenges faced in current protein complex structure prediction tasks, including the structure prediction of heteromeric complex, disordered regions in complex, antibody-antigen complex, and RNA-related complex, as well as the evaluation metrics for complex assessment. We hope that this work will provide comprehensive knowledge of complex structure predictions to contribute to future advanced predictions.

Efficient computational model of the in-flow capturing of magnetic nanoparticles by a cylindrical magnet for cancer nanomedicine.

Magnetic nanoparticles have emerged as a promising approach to improving cancer treatment. However, many nanoparticle designs fail in clinical trials due to a lack of understanding of how to overcome the in vivo transport barriers. To address this shortcoming, we develop a computational model aimed at the study of magnetic nanoparticles in vitro and in vivo. In this paper, we present an important building block for this overall goal, namely an efficient computational model of the in-flow capture of magnetic nanoparticles by a cylindrical permanent magnet in an idealized test setup. We use a continuum approach based on the Smoluchowski advection-diffusion equation, combined with a simple approach to consider the capture at an impenetrable boundary, and derive an analytical expression for the magnetic force of a cylindrical magnet of finite length on the nanoparticles. This provides a simple and numerically efficient way to study different magnet configurations and their influence on the nanoparticle distribution in three dimensions. Such an in silico model can increase insight into the underlying physics, help to design prototypes, and serve as a precursor to more complex systems in vivo and in silico.

Bioplausible Unsupervised Delay Learning for Extracting Spatiotemporal Features in Spiking Neural Networks.

The plasticity of the conduction delay between neurons plays a fundamental role in learning temporal features that are essential for processing videos, speech, and many high-level functions. However, the exact underlying mechanisms in the brain for this modulation are still under investigation. Devising a rule for precisely adjusting the synaptic delays could eventually help in developing more efficient and powerful brain-inspired computational models. In this article, we propose an unsupervised bioplausible learning rule for adjusting the synaptic delays in spiking neural networks. We also provide the mathematical proofs to show the convergence of our rule in learning spatiotemporal patterns. Furthermore, to show the effectiveness of our learning rule, we conducted several experiments on random dot kinematogram and a subset of DVS128 Gesture data sets. The experimental results indicate the efficiency of applying our proposed delay learning rule in extracting spatiotemporal features in an STDP-based spiking neural network.

BJLD-CMI: a predictive circRNA-miRNA interactions model combining multi-angle feature information.

Increasing research findings suggest that circular RNA (circRNA) exerts a crucial function in the pathogenesis of complex human diseases by binding to miRNA. Identifying their potential interactions is of paramount importance for the diagnosis and treatment of diseases. However, long cycles, small scales, and time-consuming processes characterize previous biological wet experiments. Consequently, the use of an efficient computational model to forecast the interactions between circRNA and miRNA is gradually becoming mainstream. In this study, we present a new prediction model named BJLD-CMI. The model extracts circRNA sequence features and miRNA sequence features by applying Jaccard and Bert's method and organically integrates them to obtain CMI attribute features, and then uses the graph embedding method Line to extract CMI behavioral features based on the known circRNA-miRNA correlation graph information. And then we predict the potential circRNA-miRNA interactions by fusing the multi-angle feature information such as attribute and behavior through Autoencoder in Autoencoder Networks. BJLD-CMI attained 94.95% and 90.69% of the area under the ROC curve on the CMI-9589 and CMI-9905 datasets. When compared with existing models, the results indicate that BJLD-CMI exhibits the best overall competence. During the case study experiment, we conducted a PubMed literature search to confirm that out of the top 10 predicted CMIs, seven pairs did indeed exist. These results suggest that BJLD-CMI is an effective method for predicting interactions between circRNAs and miRNAs. It provides a valuable candidate for biological wet experiments and can reduce the burden of researchers.

An intelligent model for prediction of abiotic stress-responsive microRNAs in plants using statistical moments based features and ensemble approaches

This study proposed an intelligent model for predicting abiotic stress-responsive microRNAs in plants. MicroRNAs (miRNAs) are short RNA molecules regulates the stress in genes. Experimental methods are costly and time-consuming, as compare to in-silico prediction. Addressing this gap, the study seeks to develop an efficient computational model for plant stress response prediction. The two benchmark datasets for MiRNA and Pre-MiRNA dataset have been acquired in this study. Four ensemble approaches such as bagging, boosting, stacking, and blending have been employed. Classifiers such as Random Forest (RF), Extra Trees (ET), Ada Boost (ADB), Light Gradient Boosting Machine (LGBM), and Support Vector Machine (SVM). Stacking and Blending employed all stated classifiers as base learners and Logistic Regression (LR) as Meta Classifier. There have been a total of four types of testing used, including independent set, self-consistency, cross-validation with 5 and 10 folds, and jackknife. This study has utilized evaluation metrics such as accuracy score, specificity, sensitivity, Mathew's correlation coefficient (MCC), and AUC. Our proposed methodology has outperformed existing state of the art study in both datasets based on independent set testing. The SVM-based approach has exhibited accuracy score of 0.659 for the MiRNA dataset, which is better than the previous study. The ET classifier has surpassed the accuracy of Pre-MiRNA dataset as compared to the existing benchmark study, achieving an impressive score of 0.67. The proposed method can be used in future research to predict abiotic stresses in plants.

Efficient computational modeling of electronic stopping power of organic polymers for proton therapy optimization

This comprehensive study delves into the intricate interplay between protons and organic polymers, offering insights into proton therapy in cancer treatment. Focusing on the influence of the spatial electron density distribution on stopping power estimates, we employed real-time time-dependent density functional theory coupled with the Penn method. Surprisingly, the assumption of electron density homogeneity in polymers is fundamentally flawed, resulting in an overestimation of stopping power values at energies below 2 MeV. Moreover, the Bragg rule application in specific compounds exhibited significant deviations from experimental data around the stopping maximum, challenging established norms.

On the Efficient and Accurate Non-linear Computational Modeling of Multilayered Bending Plates. State of the Art and a Novel Proposal: The 2D+\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$2\ ext {D}+$$\\end{document} Multiscale Approach

After conducting a comprehensive historical review of presently established methods for computational modeling of multilayered bending plates, the present work introduces a novel 2D multiscale strategy, termed the 2D+ approach. The proposed approach is based on the computational homogenization formalism and is envisaged to serve as an appealing alternative to current methodologies for modeling multilayered plates in bending-dominated situations. Such structural elements involve modern and relevant materials, such as laminated composites characterized by the heterogeneous distribution of low-aspect-ratio layers showing substantial non-linear mechanical behavior across their thickness.Within this proposed approach, the 2D plate mid-plane constitutes the macroscopic scale, while a 1D filament-like Representative Volume Element (RVE), orthogonal to the plate mid-plane and spanning the plate thickness, represents the mesoscopic scale. Such RVE, in turn, is capturing the non-linear mechanical behavior throughout the plate thickness at each integration point of the 2D plate-midplane finite element mesh. The chosen kinematics and discretization at the considered scales are particularly selected to (1) effectively capture relevant aspects of non-linear mechanical behavior in multilayered plates under bending-dominated scenarios, (2) achieve affordable computational times (computational efficiency), and (3) provide accurate stress distributions compared to the corresponding high-fidelity 3D simulations (computational accuracy).The proposed strategy aligns with the standard, first-order, hierarchical multiscale setting, involving the linearization of the macro-scale displacement field along the thickness. It employs an additional fluctuating displacement field in the RVE to capture higher-order behavior, which is computed through a local 1D finite element solution of a Boundary Value Problem (BVP) at the RVE. A notable feature of the presented 2D+ approach is the application of the Hill–Mandel principle, grounded in the well-established physical assumption imposing mechanical energy equivalence in the macro and meso scales. This links the 2D macroscopic plate and the set of 1D mesoscopic filaments, in a weakly-coupled manner, and yields remarkable computational savings in comparison with standard 3D modeling. Additionally, solving the resulting RVE problem in terms of the fluctuating displacement field allows the enforcement of an additional condition: fulfillment of linear momentum balance (equilibrium equations). This results in a physically meaningful 2D-like computational setting, in the considered structural object (multilayered plates in bending-dominated situations), which provides accurate stress distributions, typical of full 3D models, at the computational cost of 2D models.

Optimal Molecular Dynamics System Size for Increased Precision and Efficiency for Epoxy Materials.

Molecular dynamics (MD) simulation is an important tool for predicting thermo-mechanical properties of polymer resins at the nanometer length scale, which is particularly important for efficient computationally driven design of advanced composite materials and structures. Because of the statistical nature of modeling amorphous materials on the nanometer length scale, multiple MD models (replicates) are typically built and simulated for statistical sampling of predicted properties. Larger replicates generally provide higher precision in the predictions but result in higher simulation times. Unfortunately, there is insufficient information in the literature to establish guidelines between MD model size and the resulting precision in predicted thermo-mechanical properties. The objective of this study was to determine the optimal MD model size of epoxy resin to balance efficiency and precision. The results show that an MD model size of 15,000 atoms provides for the fastest simulations without sacrificing precision in the prediction of mass density, elastic properties, strength, and thermal properties of epoxy. The results of this study are important for efficient computational process modeling and integrated computational materials engineering (ICME) for the design of next-generation composite materials for demanding applications.

Ultrasonic Guided Wave Scattering Spectra by Hybrid Global–Local Modeling for Nondestructive Evaluation in Composites With Varying Defect Features

Abstract The nondestructive characterization and monitoring of defects in composite structures by ultrasonic guided waves are hindered by the complexity of the interaction. Efficient computational modeling of propagating ultrasonic guided waves in waveguides with discontinuities is necessary to properly tune inspections and correctly interpret the scattering behavior. This paper studies the influence of different defect types in composite plates on ultrasonic guided wave scattering. Employing a hybrid global–local method, the energy spectra of transmitted and reflected waves in defected composite plates are numerically analyzed over a broad frequency range. Through parametric analyses of defect dimensions and locations, the resulting trends of the effect of specific defect features on the scattering behavior, as well as the importance of observing scattering on a broad frequency range, are accurately computed and highlighted. The presented findings indicate the potential for effective defect characterization and monitoring.

Importance of Inter-residue Contacts for Understanding Protein Folding and Unfolding Rates, Remote Homology, and Drug Design.

Inter-residue interactions in protein structures provide valuable insights into protein folding and stability. Understanding these interactions can be helpful in many crucial applications, including rational design of therapeutic small molecules and biologics, locating functional protein sites, and predicting protein-protein and protein-ligand interactions. The process of developing machine learning models incorporating inter-residue interactions has been improved recently. This review highlights the theoretical models incorporating inter-residue interactions in predicting folding and unfolding rates of proteins. Utilizing contact maps to depict inter-residue interactions aids researchers in developing computer models for detecting remote homologs and interface residues within protein-protein complexes which, in turn, enhances our knowledge of the relationship between sequence and structure of proteins. Further, the application of contact maps derived from inter-residue interactions is highlighted in the field of drug discovery. Overall, this review presents an extensive assessment of the significant models that use inter-residue interactions to investigate folding rates, unfolding rates, remote homology, and drug development, providing potential future advancements in constructing efficient computational models in structural biology.

Improved efficient physics-based computational modeling of regional wave-driven coastal flooding for reef-lined coastlines

Coastal flooding affects low-lying communities worldwide and is expected to increase with climate change, especially along reef-lined coasts, where wave-driven flooding is particularly prevalent. However, current regional modeling approaches are either insufficient or too computationally expensive to accurately assess risks in these complex environments. This study introduces and validates an improved computationally efficient and physics-based approach to compute dynamic wave-driven regional flooding on reef-lined coasts. We coupled a simplified-physics flood model (SFINCS) with a one-dimensional wave transformation model (XBeach-1D). To assess the performance of the proposed approach, we compared its results with results from a fully resolving two-dimensional wave transformation model (XBeach-2D). We applied this approach for a range of storms and sea-level rise scenarios for two contrasting reef-lined coastal geomorphologies: one low relief area and one high relief area. Our findings reveal that SFINCS coupled with XBeach-1D generates flood extents comparable to those produced by XBeach-2D, with a hit rate of 92%. However, this method tends to underpredict the flood extent of weaker, high-frequency storms and overpredict stronger, low-frequency storms. Across scenarios, our approach overpredicted the mean flood water depth, with a positive bias of 7 cm and root mean square difference of 15 cm. Offering approximately 100 times greater computational efficiency than its two-dimensional XBeach counterpart, this flood modeling technique is recommended for wave-driven flood modeling in scenarios with high computational demands, such as modeling numerous scenarios or undertaking detailed regional-scale modeling.

Parametric analysis of additively manufactured polymer nanocomposites: A experimental and multiscale study

Fused Filament Fabrication (FFF) or Fused Deposition Modeling (FDM) is a widely utilized additive manufacturing technology in a variety of sectors. However, voids, poor bonding between layers, and FDM parameters commonly affect the FDM-printed objects, altering their strength. Carbon nanotube (CNT) composites for FDM printing have been researched by researchers to improve their characteristics. This paper proposes an efficient three-scale computational model for predicting mechanical properties as well as a unique water-quenching process for preparing CNT-infused filaments. The influence of FDM process parameters on mechanical properties is revealed by extensive parametric analysis. When compared to pure ABS, the CNT-infused composite demonstrated better bonding and modulus. The experimental study showed that an increase in the layer height deteriorates the elastic modulus by 21.03 % and 27.92 %, for ABS and ABS-CNT, respectively. The infill density increases the modulus by 49.3 % and 69.6 % for pure ABS, from 100 % to 75 % and 50 %, respectively. For parts printed in 0–00 and 0–900 orientations, a 2.11 % and a 1.7 % decrease were found for pure ABS and nanocomposite, respectively. The computational results were in good agreement with the experimental findings, with the difference varying from 10.15 % to 5.5 % for 0.1 mm and 0.2 mm layer heights. For other parameters like the raster orientation, the difference for 0–00 and 0–900 was 5.3 % and 6.9 %, respectively. The computational results agree with the experimental results, making it a useful tool for optimizing FDM printing and exploiting CNTs to improve part performance.

A multi-scale numerical approach to study monoclonal antibodies in solution.

Developing efficient and robust computational models is essential to improve our understanding of protein solution behavior. This becomes particularly important to tackle the high-concentration regime. In this context, the main challenge is to put forward coarse-grained descriptions able to reduce the level of detail, while retaining key features and relevant information. In this work, we develop an efficient strategy that can be used to investigate and gain insight into monoclonal antibody solutions under different conditions. We use a multi-scale numerical approach, which connects information obtained at all-atom and amino-acid levels to bead models. The latter has the advantage of reproducing the properties of interest while being computationally much faster. Indeed, these models allow us to perform many-protein simulations with a large number of molecules. We can, thus, explore conditions not easily accessible with more detailed descriptions, perform effective comparisons with experimental data up to very high protein concentrations, and efficiently investigate protein-protein interactions and their role in phase behavior and protein self-assembly. Here, a particular emphasis is given to the effects of charges at different ionic strengths.