Eugene A. Durenard, Jovana Džunic ́ A, Beny Neta B et al. · 2013
A general way to construct multipoint methods for solving nonlinear equations by using inverse interpolation is presented. The proposed methods belong to the class of multipoint methods with memory. In particular, a new two-point method with memory with the order 5 ffiffiffiffiffiffi 17 p =2 % 4:562 is derived. Computational efficiency of the presented methods is analyzed and their comparison with existing methods with and without memory is performed on numerical examples. It is shown that a special choice of initial approximations provides a considerably great accuracy of root approximations obtained by the proposed interpolatory iterative methods.
Cited by 1Open access
Tshilidzi Marwala · Elsevier eBooks · 2026
Cited by 0
Mohanad Faeq Ali (21354273), Mohammed Shakir Mohmood (21354276), Ban Salman Shukur (21354279) et al. · Figshare · 2025
<div> The rapid development and integration of interconnected healthcare devices and communication networks within the Internet of Medical Things (IoMT) have significantly enhanced healthcare services. However, this growth has also introduced new vulnerabilities, increasing the risk of cybersecurity attacks. These attacks threaten the confidentiality, integrity, and availability of sensitive healthcare data, raising concerns about the reliability of IoMT infrastructure. Addressing these challenges requires advanced cybersecurity measures to protect the dynamic IoMT ecosystem from evolving threats. This research focuses on enhancing cyberattack prediction and prevention in IoMT environments through innovative Machine-learning techniques to improve healthcare data security and resilience. However, the existing model’s efficiency depends on the diversity of data, which leads to computational complexity issues. In addition, the conventional model faces overfitting issues in training data, which causes prediction inaccuracies. Thus, the research introduces the hybridized cyber attack prediction model (HCAP) and analyzes various IoMT data source information to address the limitations of dataset availability issues. The gathered information is processed with the help of Principal Component-Recursive Feature Elimination (PC-RFE), which eliminates the irrelevant features. The extracted features are fed into the lion-optimization technique to fine-tune the hyperparameters of the recurrent neural networks, enhancing the model’s ability to efficiently predict cybersecurity threats with a maximum recognition rate in IoMT environments. The recurrent networks, specifically Long Short Term Memory (LSTM), process data from healthcare devices, identifying abnormal patterns that indicate potential cyberattacks over time. The created system was implemented using Python, and various metrics, including false positive and false negative rates, accuracy, precision, recall, and computational efficiency, were used for evaluation. The results demonstrated that the proposed HCAP model achieved 98% accuracy in detecting cyberattacks and outperformed existing models, reducing the false positive rate by 25%. The false negative rate by 20% and a 30% improvement in computational efficiency enhances the reliability of IoMT threat detection in healthcare applications. </div>
Cited by 0Open access
Ran Duan (821258), Lin Gao (344965), Yong Gao (29929) et al. · Figshare · 2021
<p>(A) Running time of different datasets. Here, we only included the results for the BRCA datasets in Dataset #1 and the Pan-cancer datasets in Dataset #3. The difference between the running times of these methods is similar to other cancer datasets but less significant as those datasets are much smaller than BRCA datasets. The data combinations are placed along the x-coordinate in ascending order of the number of participating data types and data scales. (B) Total running time. (C) Fold changes of running time on different datasets. (D) Computational efficiency rank table. The rank items listed under the table include the datasets and cancer which are connected by underlines. “Com” and “Sig” represent complete and significant datasets, respectively.</p>
Cited by 0Open access
Yi Wei Tye (22208735), XinYing Chew (8690334), Umi Kalsom Yusof (22208738) et al. · Figshare · 2025
<div> Advances in data collection have resulted in an exponential growth of high-dimensional microarray datasets for binary classification in bioinformatics and medical diagnostics. These datasets generally possess many features but relatively few samples, resulting in challenges associated with the “curse of dimensionality”, such as feature redundancy and an elevated risk of overfitting. While traditional feature selection approaches, such as filter-based and wrapper-based approaches, can help to reduce dimensionality, they often struggle to capture feature interactions while adequately preserving model generalization. Therefore, this paper introduces the Adaptive Cluster-Guided Simple, Fast, and Efficient (ACG-SFE) feature selection, a hybrid approach designed to address the challenges of high-dimensional microarray data in binary classification. ACG-SFE enhances the Simple, Fast, and Efficient (SFE) evolutionary feature selection model by integrating hierarchical clustering to dynamically group correlated features based on the optimal number of clusters determined by the Silhouette index, Davies-Bouldin score, and the feature-to-observation ratio while adaptively selecting representative features within clusters using mutual information and adjusting the selection threshold through a progress factor. This hybrid filter-wrapper approach improves feature interactions, effectively minimizing redundancy and overfitting while enhancing classification performance. The proposed model is assessed against four state-of-the-art evolutionary feature selection models on 11 high-dimensional microarray datasets. Experimental results indicate that ACG-SFE effectively selects a small yet pertinent feature subset, minimizing redundancy while attaining enhanced classification accuracy and F-measure. Furthermore, its reduced RMSE between train and test accuracy substantiates its capability to reduce overfitting, outperforming comparative models. These findings establish ACG-SFE as an effective feature selection model for handling high-dimensional microarray data in binary classification, enhancing classification accuracy while selecting minimal relevant features to reduce unnecessary complexity and the risk of overfitting. </div>
Cited by 0Open access
Ruizhi Yuan (16864629), Ziqi Rong (22339789), Haoran Hu (6032510) et al. · Figshare · 2025
<div> Large-scale single-cell projects generate rapidly growing datasets, but downstream analysis is often confounded by data sources, requiring data integration methods to do correction. Existing data integration methods typically require data centralization, raising privacy and security concerns. Here, we introduce Federated Harmony, a novel method combining properties of federated learning with Harmony algorithm to integrate decentralized omics data. This approach preserves privacy by avoiding raw data sharing while maintaining integration performance comparable to Harmony. Experiments on various types of single-cell data showcase superior results, highlighting a novel data integration approach for distributed multi-omics data without compromising data privacy or analytical performance. </div>
Cited by 0Open access
Li Liu (75607) · Figshare · 2025
<div> Campus safety is an essential concern as schools, colleges, and universities work to create secure environments for students, staff, and visitors. Many existing security systems are not fully effective at detecting unusual behaviors or sending fast alerts, which can delay responses to potential threats. To improve this, the research introduces DeepCARE(Deep-learning-based Campus Anomaly & Risk Evaluation), a deep learning-based framework designed to enhance behavior recognition and early warning for campus security. DeepCARE combines convolutional neural networks (CNNs) with long short-term memory (LSTM) networks to process real-time video footage and detect abnormal activities such as aggression, unauthorized access, and people staying too long in restricted areas. The system’s main feature is its hybrid model, where CNNs extract key visual features from surveillance footage while LSTM networks analyze these features over time to recognize behavior patterns. DeepCARE also includes an anomaly detection module using autoencoders, which helps improve the system’s accuracy and reduces false alarms. This makes DeepCARE a flexible and scalable solution, suitable not only for educational campuses but also for public spaces, transport hubs, and smart cities. By applying deep learning, DeepCARE supports early risk detection and faster response times, helping security teams create safer spaces. Experimental results show that DeepCARE achieves a behavior recognition accuracy of 94.5%, performs 8% better than traditional methods, and shortens emergency response times by 30%. </div>
Cited by 0Open access
Wenhao Ren (2561731), Zuowei Zhong (21300382) · Figshare · 2025
<div> Developing an efficient and accurate algorithm for detecting building cracks, especially micro-cracks, is essential for ensuring structural integrity and safety. The identification and precise localization of cracks remain challenging due to varying crack sizes and the inconsistency in available datasets. To address these issues, this study introduces an innovative crack detection model based on YOLOv8n. The proposed method incorporates two novel components: AC-LayeringNetV2, a hierarchical backbone network that optimizes feature extraction by integrating local, peripheral, and global contextual information, and RAK-Conv, a convolutional module that combines an attention mechanism with irregular convolution operations to enhance the model’s ability to handle complex backgrounds. These innovations significantly improve semantic segmentation accuracy while reducing computational overhead. Experimental results on a benchmark dataset demonstrate a 2.20% improvement in precision, a 3.50% increase in recall, and a 1.90% rise in mAP@50 compared to the baseline model. Additionally, the model achieves a 6.55% reduction in size and a 0.03% decrease in computational complexity. These results highlight the practical applicability and efficiency of the proposed approach for automatic crack detection in building structures, emphasizing the novel integration of feature fusion and attention mechanisms to address challenges in real-time and high-accuracy detection of micro-cracks in complex environments. </div>
Cited by 0Open access
Bin Wu (138543), Caifen Wang (1501207), Hailong Yao (8693943) · Figshare · 2020
Cited by 0Open access
Tao Geng (32392), Yuxuan Yang (8088098) · Figshare · 2025
<div> Interior design is a complex and creative discipline involving aesthetics, functionality, ergonomics, and materials science. Effective solutions must meet diverse requirements, typically producing multiple deliverables such as renderings and design drawings from various perspectives. Consequently, interior design processes are often inefficient and demand significant creativity. With advances in machine learning, generative models have emerged as a promising means of improving efficiency by creating designs from text descriptions or sketches. However, few generative works focus on interior design, leading to substantial discrepancies between outputs and practical needs, such as differences in size, spatial scope, and the lack of controllable generation quality. To address these challenges, we propose DiffDesign, a controllable diffusion model with meta priors for efficient interior design generation. Specifically, we utilize the generative priors of a 2D diffusion model pre-trained on a large image dataset as our rendering backbone. We further guide the denoising process by disentangling cross-attention control over design attributes, such as appearance, pose, and size, and introduce an optimal transfer-based alignment module to enforce view consistency. Simultaneously, we construct an interior design-specific dataset, DesignHelper, consisting of over 400 solutions across more than 15 spatial types and 15 design styles. This dataset helps fine-tune DiffDesign. Extensive experiments conducted on various benchmark datasets demonstrate the effectiveness and robustness of DiffDesign. </div>
Cited by 0Open access
Ran Duan (821258), Lin Gao (344965), Yong Gao (29929) et al. · Figshare · 2021
Cited by 0Open access
Jun Wang (5906), Ning Huang (623370), Houzhong Zhang (22521907) et al. · Figshare · 2025
<div> The rapid evolution of cyber threats poses significant challenges to the adaptability and performance of anomaly detection systems. This study presents an innovative hybrid deep learning framework that integrates Convolutional Neural Networks (CNN), Long Short-Term Memory networks (LSTM), and Transformer models with a novel self-learning mechanism to enhance network traffic anomaly detection. Our key contributions include: (1) a synergistic two-stage model fusion architecture that captures both spatial and temporal traffic patterns; (2) an adaptive learning mechanism with multi-metric drift detection that autonomously responds to evolving threats; and (3) a knowledge preservation strategy that maintains detection capabilities while adapting to new attack patterns. The proposed CNN-LSTM model achieves F1-scores of 0.9778 and 0.9695 on the UNSW-NB15 and CICIDS2017 datasets respectively for binary classification of normal vs. anomalous traffic. The LSTM-Transformer model further classifies specific anomaly types with accuracies of 0.9632 and 0.9528 on these datasets, representing significant improvements over recent methods. Experiments demonstrate the framework’s robustness, maintaining an average accuracy of 0.955 ( 0.005) over a 15-day simulated period with multiple induced concept drifts. The self-learning mechanism, with multi-metric drift detection and an efficient model update strategy, enables the system to detect drifts and recover performance within 23.4 ± 0.20 hours post-drift, while achieving a 92.8% detection rate for zero-day attacks. The proposed framework offers a promising direction for developing efficient and autonomous cybersecurity systems capable of handling dynamic and evolving threat landscapes. </div>
Cited by 0Open access
Huijun Zhu (317891), Dong Xie (16842), Haseeb Ahmad (4971733) et al. · Figshare · 2021
Cited by 0Open access
Huijun Zhu (317891), Dong Xie (16842), Haseeb Ahmad (4971733) et al. · Figshare · 2021
Cited by 0Open access
Huijun Zhu (317891), Dong Xie (16842), Haseeb Ahmad (4971733) et al. · Figshare · 2021
Cited by 0Open access
Mehmet Zahid Genc (22683526), Yaser Dalveren (22683529), Ali Kara (690960) et al. · Figshare · 2025
<div> Accurate measurement of spleen volume is essential for the diagnosis of splenomegaly. While Computed Tomography (CT) is among the most reliable imaging modalities for this task, manual segmentation of the spleen is labor-intensive and impractical for routine clinical workflows. Automatic segmentation methods provide a more viable alternative for clinical deployment. In recent years, 3D Convolutional Neural Network (CNN) models have been widely used for this purpose due to their high segmentation accuracy. However, their computational and memory demands make them less suitable for real-time applications on edge devices with limited processing capabilities. To address these limitations, we introduce SVNC-Net (Spleen Volume and Neighborhood Convolutional Network) for efficient 3D spleen segmentation from CT scans. Rather than developing an entirely new architecture from scratch, SVNC-Net builds upon the U-Net framework with targeted architectural optimizations for efficiency. In SVNC-Net, each CT slice is processed independently using 2D convolutions. In its architecture, depthwise separable convolution is used to significantly reduce computational complexity and memory usage. To evaluate its performance and efficiency, a comparative analysis was conducted against well-known CNN-based models, including UPerNet, EMANet, CCNet, SegNet, and ShuffleNet. This evaluation was performed on two publicly available datasets used together for the first time in the literature. The promising results achieved from the comparative analysis verified that SVNC-Net is highly suitable for real-time applications and resource-constrained environments. Additionally, we explore post-training compression techniques such as pruning and quantization, which further enhance the model’s compactness and inference speed. These findings contribute to the ongoing efforts to develop efficient 2D deep learning models for 3D organ segmentation, particularly in resource-constrained clinical scenarios. </div>
Cited by 0Open access
Eric León Olivares, Luis Carlos Márquez Strociak, Mayra Lorena González-Mosqueda et al. · International Journal of Combinatorial Optimization Problems and Informatics. · 2025
The implementation of mathematical algorithms plays a fundamental role in computational efficiency. Sequential programming, which processes instructions in a linear manner, often struggles with large data volumes due to its inherent limitations. In contrast, parallel programming distributes tasks across multiple cores, significantly reducing processing times and improving overall performance. This paper presents a comparative analysis of both approaches and their relevance in Systems Engineering, where computational optimization is critical. To this end, we implement and evaluate the Sobel algorithm—commonly used for edge detection in images—in both sequential and parallel modes. The implementation is carried out in Python, leveraging the NumPy, OpenCV, and Multiprocessing libraries. This study analyzes the conditions under which parallelization enhances performance and identifies scenarios where process overhead may negate its benefits, thus establishing fundamental criteria for applying these techniques to solve mathematical problems in engineering. The source code is available on GitHub at: [GitHub Repository].
Cited by 0Open access
Bird, Thomas · UCL Discovery (University College London) · 2023
In this thesis we seek to make advances towards the goal of effective learned compression. This entails using machine learning models as the core constituent of compression algorithms, rather than hand-crafted components. \nTo that end, we first describe a new method for lossless compression. This method \nallows a class of existing machine learning models – latent variable models – to be \nturned into lossless compressors. Thus many future advancements in the field of \nlatent variable modelling can be leveraged in the field of lossless compression. We \ndemonstrate a proof-of-concept of this method on image compression. Further, we \nshow that it can scale to very large models, and image compression problems which \nclosely resemble the real-world use cases that we seek to tackle. \nThe use of the above compression method relies on executing a latent variable \nmodel. Since these models can be large in size and slow to run, we consider how \nto mitigate these computational costs. We show that by implementing much of the \nmodels using binary precision parameters, rather than floating-point precision, we \ncan still achieve reasonable modelling performance but requiring a fraction of the \nstorage space and execution time. \nLastly, we consider how learned compression can be applied to 3D scene data - a \ndata medium increasing in prevalence, and which can require a significant amount of \nspace. A recently developed class of machine learning models - scene representation \nfunctions - has demonstrated good results on modelling such 3D scene data. We \nshow that by compressing these representation functions themselves we can achieve \ngood scene reconstruction with a very small model size.
Cited by 0Open access
Shahar Dobzinski · arXiv (Cornell University) · 2016
We characterize the communication complexity of truthful mechanisms. Our departure point is the well known taxation principle. The taxation principle asserts that every truthful mechanism can be interpreted as follows: every player is presented with a menu that consists of a price for each bundle (the prices depend only on the valuations of the other players). Each player is allocated a bundle that maximizes his profit according to this menu. We define the taxation complexity of a truthful mechanism to be the logarithm of the maximum number of menus that may be presented to a player. Our main finding is that in general the taxation complexity essentially equals the communication complexity. The proof consists of two main steps. First, we prove that for rich enough domains the taxation complexity is at most the communication complexity. We then show that the taxation complexity is much smaller than the communication complexity only in "pathological" cases and provide a formal description of these extreme cases. Next, we study mechanisms that access the valuations via value queries only. In this setting we establish that the menu complexity - a notion that was already studied in several different contexts - characterizes the number of value queries that the mechanism makes in exactly the same way that the taxation complexity characterizes the communication complexity. Our approach yields several applications, including strengthening the solution concept with low communication overhead, fast computation of prices, and hardness of approximation by computationally efficient truthful mechanisms.
Cited by 3Open access
Ding, Yihang · OpenGrey (Institut de l'Information Scientifique et Technique) · 2009
Cited by 0