Your search

Completely Automated Public Turing tests to tell Computers and Humans Apart (CAPTCHAs) are widely used mechanisms for protecting websites against automated bot attacks. In this paper, we investigate the trade-off between usability and security in CAPTCHA design using a Stackelberg game-theoretic framework. Based on empirical evaluation data on CAPTCHA solving time and accuracy, we incorporate expected solving time into the defender's utility function to jointly capture usability and security. We propose two models: a discrete-strategy game, in which different CAPTCHA types and difficulty levels form the strategy space, and a continuous-strategy game, in which CAPTCHA hardness is represented as a continuous decision variable. For the discrete model, we derive a stepwise solution characterizing transitions in the defender's optimal strategy. For the continuous model, we analyze a threshold-based attacker response and obtain equilibrium solutions under deterrence and accommodation regimes. The proposed framework is adaptable to other CAPTCHA evaluation datasets.

Cardiovascular disease (CVD) risk stratification often relies on anthropometric measures that do not fully reflect central adiposity or differences across populations. The Body Roundness Index (BRI) is a body-shape indicator that may improve assessment of cardiometabolic risk, but evidence from large national cohorts remains limited. This study analyzed harmonized data from the China Health and Retirement Longitudinal Study (CHARLS, 2015) and the U.S. National Health and Nutrition Examination Survey (NHANES, 2005–2016) to examine the relationship between BRI and prevalent CVD. Participants were grouped into quartiles of BRI, with the highest quartile serving as the reference. Weighted multivariable logistic regression models were used to estimate odds ratios (ORs) and 95% confidence intervals (CIs), with sequential adjustment for demographic and clinical covariates. Subgroup analyses were performed to evaluate heterogeneity across population strata, and an extreme gradient boosting (XGBoost) model was applied to assess the relative importance of BRI and other predictors. In both cohorts, individuals with CVD were older and had higher prevalences of hypertension and diabetes, whereas BRI distributions differed between cohorts. Lower BRI quartiles were associated with greater odds of CVD. Compared with Q4, Q1 was associated with increased CVD risk in CHARLS (OR 1.73, 95% CI 1.14-2.68; p=0.012) and NHANES (OR 1.82, 95% CI 1.44-2.30; p<0.001). These findings suggest that lower BRI is associated with higher CVD risk and may complement conventional anthropometric measures in population-based cardiovascular risk assessment.

This study evaluates the viability of deep learningbased super-resolution (SR) for enhancing real-time video streaming. We compare a traditional HEVC-compressed streaming pipeline against an AI-assisted framework that transmits lowresolution video and reconstructs high-resolution output at the client using the Efficient Sub-Pixel Convolutional Neural Network (ESPCN) algorithm. The evaluation spans multiple resolutions (1080p to 4K), motion types, and processing methodologies (subprocess vs. frame-by-frame). Experimental results indicate that while AI-assisted streaming significantly reduces transmission bandwidth, it introduces substantial computational overhead, with completion times increasing by over 300% compared to traditional methods. Furthermore, quality assessments using PSNR and VMAF metrics reveal that most AI-assisted streams failed to meet the target quality range (30–50 dB PSNR), with high-resolution 4K streams experiencing up to 24.58% frame loss during real-time inference. These findings highlight critical performance barriers in utilizing current SR algorithms for standard consumer-grade real-time video delivery.

Capacitive swipe gestures have recently emerged as a promising behavioral biometric modality for user authentication and identification on smartphones. While prior studies have demonstrated their effectiveness using classical machine learning techniques, the potential of deep learning in this domain remains largely unexplored, particularly under limited training data conditions. In this paper, we investigate the use of deep learning for capacitive swipe–based user identification and address the challenge of data scarcity through an image fusion strategy. Experiments on a dataset of 30 users show that the proposed "deep learning plus image fusion" approach significantly improves performance over a standalone deep learning baseline and achieves results comparable to state-of-the-art classical machine learning methods. These findings highlight the importance of data size and demonstrate the potential of image fusion for enhancing deep learning–based user identification.

This research presents a comparative performance analysis of three prominent hypervisor architectures—KVM, Xen, and VMware Workstation—to evaluate their scalability and resource efficiency under increasing virtual machine (VM) densities. Utilizing a standardized benchmarking framework, this study quantifies performance regressing across CPU throughput (GFLOPS), memory bandwidth, disk I/O, and initialization latency. The experiment results reveal that architectural design significantly dictates scalability limits: KVM’s kernel-integrated Type-1 model achieved the highest efficiency, maintaining a near-native CPU baseline and superior I/O resilience at peak density. In contrast, Xen’s microkernel architecture demonstrated stable but lower throughput due to Domain 0 (Dom0) management overhead, while VMware’s Type-2 hosted model exhibited the most acute performance degradation. These results demonstrate that while hosted hypervisors are suitable for low-density personal use, kernel-level integration is a mechanical necessity for maintaining performance in high-concurrency server environments. This work provides a reproducible methodology and critical data for system architects and researchers seeking to optimize resource orchestration in distributed and parallel systems.

Arabic extractive text summarization remains challenging due to rich morphology and limited lexical consistency, which weaken traditional surface-based sentence ranking methods. Existing approaches typically rely either on graph centrality to capture document structure or on transformer-based embeddings to estimate semantic relevance, yet the relationship between these two ranking signals has not been systematically examined. This paper investigates the complementary roles of semantic similarity and structural centrality in Arabic summarization. A controlled hybrid ranking formulation is implemented that combines contextual sentence embeddings with PageRank-based graph scoring and evaluate it using multi-reference ROUGE on the Extended Arabic Summaries Corpus (EASC). The study analyzes how each component influences sentence salience, coverage, and redundancy. Experimental results show that semantic embeddings improve the identification of informative sentences, while graph centrality enhances coverage and reduces repetition; their integration consistently yields stronger summaries, achieving ROUGE-1 = 0.605, ROUGE-2 = 0.497, and ROUGE-L = 0.514. These findings provide empirical evidence that semantic and structural centrality capture complementary aspects of importance in morphologically rich languages, offering guidance for designing more robust extractive summarization systems.

Cardiovascular diseases (CVDs) are still one of the foremost causes of death globally. An urgent need for an accurate premature diagnosis is essential for efficient treatment and avoidance. The goal of this research is to introduce our initial idea for developing a hybrid intelligent diagnosis model that integrates the Extreme Learning Machine (ELM) with Particle Swarm Optimization (PSO) to enhance diagnosis accuracy for heart disease. To address the problems of random initialization, the PSO algorithm is used to optimize the ELM’s input biases and weights. Experiments using the Cleveland Heart Disease dataset evaluated the performance of traditional ELM and the proposed PSO-ELM model. Results indicate that integrating PSO for parameter tuning yields significant improvements across evaluation metrics. The traditional ELM recorded an accuracy of 0.8595 on the training set and 0.8525 on the test set, while the PSO-ELM model achieved up to 0.9174 in training and 0.8852 in testing. Enhancements were also noted in precision, recall, and F1-score, with the F1-score increasing from 0.8768 for traditional ELM to over 0.92 for several PSO configurations.

Gold serves as a safe-haven asset, an inflation hedge, and a store of value, among other economic purposes. History has a wealth of information about its connections to important macroeconomic and financial variables. We compare three supervised methods for one-day GLD (SPDR Gold Shares) prediction using cross-Asset drivers: SPX (S&P 500 Index), USO (United States Oil Fund), SLV (iShares Silver Trust), and EUR/USD (Euro-U.S. Dollar exchange rate). To fill this gap: (i) a linear regression baseline, (ii) a feedforward neural network (FFNN) trained by a Genetic Algorithm (GA), and (iii) an FFNN optimized through Particle Swarm Optimization (PSO). We use standard scaling, a held-out train/test split, and R2, Variance Accounted For (VAF), MSE, RMSE, MAE, and GA/PSO convergence curves to check how the optimizer works with daily data from 2008 to 2018. Empirically, PSO-FFNN does better than both the linear baseline and GA-FFNN. It has the best generalization (for example, Test MSE 38.99, R 0.929). This implies that (1) GLD possesses exploitable nonlinear structure concerning these predictors, and (2) PSO traverses the FFNN search space more effectively than GA in our context. The findings validate the application of evolutionary training for reliable and accurate gold price forecasts, with implications for risk management and strategic asset allocation. © 2025 IEEE.

Traditional physical biometrics - such as fingerprints, facial recognition, and iris scans - have long been utilized for user identification in areas like border control, military operations, law enforcement, and public safety. However, the rise of smartphone technology has introduced new avenues for security research. One emerging area is behavioral biometrics, particularly the use of touchscreen interaction data as a more accessible and user-friendly identification method. In this context, our study focuses on a novel form of touchscreen input: capacitive swipe gestures for user identification. We compiled a comprehensive dataset of capacitive swipe gestures collected over multiple sessions from 30 participants. To evaluate this modality, we conducted thorough experiments using established machine learning algorithms, including Support Vector Machine, Random Forest, and XGBoost. Additionally, we developed a new preprocessing algorithm tailored for capacitive swipe data. Our findings reveal that this algorithm significantly enhances identification performance compared to existing methods. Overall, our results highlight the strong potential of capacitive swipe gestures as a viable biometric modality for user identification. © 2025 IEEE.

Tomato plant diseases pose a significant threat to agricultural productivity, resulting in substantial economic losses. Early and accurate diagnosis is crucial for effective disease management. This paper describes the design and implementation of expert systems for tomato disease detection using the CLIPS (C Language Integrated Production System) platform. The tool is designed to help farmers and agronomists accurately identify diseases affecting tomato crops by simulating knowledge from professional experts. We carefully developed a set of rules to distinguish leaf blight symptoms from those of other tomato diseases and provided recommendations to minimize crop losses and maximize yields. The expert system was developed using a forward-chaining inference engine, and its performance was evaluated through a set of real-world test cases, demonstrating a high level of accuracy and consistency in decision-making. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2026.

With the increasing interest in natural language processing, text summarization has become essential for condensing large volumes of data into concise and meaningful summaries. Extractive summarization, which involves selecting key sentences based on textual features, has gained attention due to its efficiency and effectiveness. This research explores extractive summarization using multiple machine learning classifiers, including Support Vector Machines (SVM), Logistic Regression (LR), Decision Trees (DT), K-Nearest Neighbors (KNN), and Random Forest (RF). Our findings indicate that the Random Forest model achieved the highest accuracy, reaching 80% in classifying sentences for summary generation. Additionally, we evaluated text classification on the same BBC dataset using ChatGPT, which attained an accuracy of 62%. Furthermore, comparisons with results from prior research confirm the competitive performance of our approach, reinforcing the potential of machine learning models in extractive summarization. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2026.

Coronary heart disease (CHD) is the leading global cause of death, making early detection essential. While coronary angiography is the diagnostic gold standard, its invasive nature poses risks, and non-invasive symptom-based methods often lack accuracy. Machine learning-powered computer-aided diagnostic systems can effectively address challenges in clinical decisionmaking. This work presents an Evolutionary Strategy-optimized Support Vector Machine (ES-SVM) model for classifying CHD based on non-invasive test results and patient characteristics. Using the Coronary Heart Disease dataset, the proposed ESSVM demonstrated significant precision and F1-scores, as well as the accuracy of the proposed model. The results indicate that SVM performance can be significantly enhanced through evolutionary hyperparameter tuning, resulting in a reliable, noninvasive diagnostic tool for initial CAD screening and supporting early intervention techniques. © 2025 IEEE.

Modeling lipase activity aids researchers in optimizing features such as temperature, pH, and substrate concentration to boost enzyme performance. This is essential in biotechnology for progressing the productivity and yield of processes such as fermentation, biodiesel production, and bioremediation. Fermentation is a highly complex, multivariable, and non-linear biotechnological process that produces bioactive materials. This study leverages artificial neural networks (ANN) to predict lipase activity in batch fermentation processes, addressing the inherent challenges in weight learning optimization often encountered with traditional algorithms like Backpropagation (BP). Several metaheuristic algorithms were employed to optimize the Multilayered Perceptron (MLP) structure and weights, including moth-frequency optimization (MFO), Particle Swarm Optimization (PSO), Dandelion Optimizer Algorithm (DO), Crow Search Algorithm (CSA), and Salp Swarm Algorithm (SSA) to overcome these limitations. Among the tested algorithms, MFO emerged as the most effective approach, achieving superior performance in weight learning with the best fitness value (i.e., mean square error (MSE)) of 0.6006. MFO-optimized ANN models deliver the most accurate predictions for lipase activity, highlighting their potential as a powerful tool for advancing industrial fermentation process optimization. © 2025 IEEE.

Traditional brain tumor diagnosis and classification are time-consuming and heavily reliant on radiologist expertise. The ever-growing patient population generates vast data, rendering existing methods expensive and inefficient. Deep Learning (DL) is a promising approach for developing automated systems to diagnose or segment brain tumors with high accuracy in less time. Within Deep Learning, Convolutional Neural Networks (CNNs) are potent tools for image classification tasks. This is achieved through a series of specialized layers, including convolution layers that identify patterns within images, pooling layers that summarize these patterns, fully connected layers that ultimately classify the image, and a feedforward layer to produce the output class. This study employed a CNN to classify brain tumors in T1-weighted contrast-enhanced images with various image resolutions, including 30×30, 50×50, 70×70, 100×100, and 150×150 pixels. The model successfully distinguished between three tumor types: glioma, meningioma, and pituitary. The CNN's impressive accuracy on training data reached up to 86.38% for image resolution (30×30) and 94.64% for higher resolution (150×150). This indicates its potential as a valuable tool in real-world brain tumor classification tasks. © 2025 IEEE.

The λ-fold complete 3-uniform hypergraph on v vertices has the edge multiset consisting of λ copies of each 3-element subset of its vertex set. A tight 6-cycle, denoted TC6, is a hypergraph with vertex set {a,b,c,d,e,f} and edge set {{a,b,c},{b,c,d},{c,d,e},{d,e,f},{e,f,a},{f,a,b}}. We give necessary and sufficient conditions on v for the existence of a TC6-decomposition of the λ-fold complete 3-uniform hypergraph on v vertices for any positive integer λ. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

Intensity interferometry, also known as the Hanbury Brown and Twiss effect, has seen significant interest in astronomy in recent years. The method involves recording timing correlations between photons received at two or more telescopes in order to derive extremely high spatial resolution information about an astronomical object, potentially including imaging stellar surfaces and other objects at unprecedented scales. This paper will briefly review the technique, discuss the performance characteristics of the of photon counters used in modern intensity interferometers, and describe opportunities for the future. As an example of photon counting with a working instrument, observing experiences with the Southern Connecticut Stellar Interferometer (SCSI), a three-station instrument using single-photon avalanche diode (SPAD) detectors, will be described. The recent lessons learned with this and other instruments in use today give a clear picture of the next steps needed to upgrade efficiency and successfully observe fainter objects. If successful, these improvements would provide a strong argument for creating situations where intensity interferometers can have baselines of one to several kilometers, which would unlock the spatial detail needed to address several exciting astrophysical questions.

The exponential growth of big data, driven by AI and machine learning technologies, underscores the need for an ethical and sustainable approach to data utilization. Using problematization methodology, we consider the assumptions underpinning Big Data and AI and reconsider them from a sensemaking perspective. Big data represents an enactment rather than an objective reality, and organizations play an active role in its adoption and use. Strategizing is driven by plausibility rather than accuracy, and big data generates a retrospection of the past rather than a prediction of the future. A sensemaking perspective serves as reality check for managers, emphasizing the necessity of long-term sustainability and societal well-being. By cultivating experiments for learning communities and incubating innovation, organizations can effectively leverage big data in marketing, fostering transparent, collaborative, ethical, and sustainable data practices. © 2025 IEEE Computer Society. All rights reserved.

Lung Adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC) are the two main histology subtypes of non-small cell lung cancer (NSCLC) with 70% of total Lung Cancer. In this article we proposed an ensemble-based model for the identification of subtypes of NSCLC using methylation data. Proposed Random Forest-based model along with out of bag (OOB) error based feature selection technique identified the top ten most important CpG sites that are highly differentiator between LUSC and LUAD subtypes of NSCLC with an accuracy, precision and F1 Score of \(97\%\) . The proposed model outperformed the other existing models for the same purpose with huge margin of 12%. Pathway analysis of the proposed 10 CpG sites revealed different pathways for LUAD and LUSC associated genes, LUAD-associated genes primarily participated in TP53, PTEN, GLP-1, Incretin regulation, and apoptosis. Conversely, LUSC-associated genes were predominantly involved in pathways for platelet degranulation, serine biosynthesis, and Nephrin family interaction.