Information doi: 10.3390/info17060533

Authors: Amine Ziane Houda El Bouhissi Thomas Hanne

Diabetes mellitus is a highly prevalent chronic disease; early diagnosis reduces severe complications. This work presents a diabetes prediction pipeline that combines metaheuristic feature selection with machine learning classification. We propose a hybrid Particle Swarm Optimization and Grey Wolf Optimizer (PSO-GWO) with alternating collaboration and an adaptive fitness function that adjusts to class balance, sample size, and dimensionality. Selected features are evaluated with random forest (primary), support vector machines, k-nearest neighbors, and logistic regression. The approach is assessed on three clinical datasets (Pima Indians, Frankfurt Hospital, Iraq) using stratified five-fold cross-validation. At the feature selection stage, the hybrid selector reaches 83.36% mean cross-validation accuracy while retaining about 74% of features on average. At the final classification stage, after random forest hyperparameter optimization on the selected features, the optimized random forest achieves 84.74% mean accuracy. Feature count is reduced by about 26% on average without loss of performance, improving interpretability and prospects for clinical use.

Information doi: 10.3390/info17060532

Authors: Abbas Albarq Amal K. Suleiman Ahmed Mohamed Hasanein Azzam Albutayh

The rapid expansion of digital healthcare technologies has accelerated the adoption of smartwatches and mobile health applications; however, empirical evidence explaining patient adoption behavior in rapidly digitalizing healthcare systems such as Saudi Arabia remains limited. This study examines the determinants influencing patients’ intention to use smartwatches and healthcare mobile applications by applying an extended Technology Acceptance Model (TAM). A cross-sectional survey was conducted among 418 participants with prior experience using wearable or mobile health technologies, and the data were analyzed using structural equation modeling. The results reveal that perceived usefulness, perceived ease of use, social influence, and facilitating conditions significantly and positively influence users’ attitudes toward digital healthcare technologies, while attitude toward use strongly predicts behavioral intention. The findings extend TAM by demonstrating the combined role of individual perceptions and contextual support factors in shaping digital health adoption in an emerging national digital health ecosystem. These results provide actionable implications for healthcare policymakers, system developers, and service providers seeking to accelerate the adoption of wearable and mobile health technologies and support national digital health transformation initiatives.

Information doi: 10.3390/info17060531

Authors: Andrea Maugeri Martina Barchitta Antonella Agodi

Web-based repeated-measures cohorts enable remote, scalable, and temporally structured monitoring of health-related outcomes in naturalistic settings. This paper presents the DiCoBENE study, a web-based cohort of healthcare-track university students, and reports evidence-informed instrument selection together with protocol features and pilot baseline findings. A structured review was used to inform the web-based administration of patient-reported outcome measures (PROMs) covering sleep quality, perceived stress, anxiety symptoms, depressive symptoms, and quality of life. In the pilot baseline sample, 442 students constituted the analytic dataset and 370–372 completed the core PROM battery, depending on the instrument. Poor sleep quality, anxiety symptoms, depressive symptoms, and perceived stress were common. Internal consistency was good to excellent for the Generalized Anxiety Disorder 7-item scale (GAD-7), the Patient Health Questionnaire 9-item depression module (PHQ-9), and the 10-item Perceived Stress Scale (PSS-10), and moderate for the Pittsburgh Sleep Quality Index (PSQI). Exploratory multivariate analyses, including latent profile analysis, principal component analysis, and partial-correlation network analysis, suggested that baseline heterogeneity was more parsimoniously summarized as a graded multidimensional burden continuum than as sharply separated phenotypes. Taken together, these findings position DiCoBENE as a methodologically explicit framework for web-based repeated outcome assessment in student well-being research.

Information doi: 10.3390/info17060530

Authors: Dan-Yang Yi Xiao-Dong Sun Jun-Hui Wang

While virtual tourism (VT) has emerged as a disruptive force in destination marketing, the mechanism by which virtual immersion translates into physical visitation remains debated. Addressing the “virtual-to-real” conversion gap, this study proposes an integrated theoretical framework combining the Stimulus–Organism–Response (SOR) model with the Technology Acceptance Model (TAM). Unlike traditional studies, we position Perceived Usefulness (PU) and Perceived Ease of Use (PEOU) as boundary conditions rather than direct antecedents. Empirical data were collected from 476 tourists with virtual experiences of Zhangjiajie National Forest Park and analyzed using Structural Equation Modeling (SEM). The results indicate that virtual experiences not only directly trigger visit intention but also indirectly foster it by enhancing destination attitude. Crucially, a novel “asymmetric moderation” effect was revealed: while technical attributes (PU and PEOU) do not influence the affective formation of attitude, they significantly moderate the translation of attitude and experience into behavioral intention. These findings suggest that while immersion drives “liking,” technical utility drives “going.” This study offers strategic insights for Destination Marketing Organizations (DMOs) to optimize VT platforms by balancing hedonic experience with functional utility to maximize actual visitor conversion.

Information doi: 10.3390/info17060529

Authors: Alejandro Carrillo-Gómez Daniela Moctezuma Enrique Camacho-Pérez

The early detection and spatial characterization of crop damage are critical for improving decision-making in precision agriculture, particularly in regions where traditional monitoring methods are limited in scalability and objectivity. This study presents an integrated information processing framework that couples UAV-based image acquisition, instance segmentation, slicing-aided inference of large orthomosaics, and georeferenced spatial analysis into a single reproducible pipeline for the detection and mapping of crop damage. The framework is applied to maize cultivated under traditional milpa systems in Yucatán, Mexico, a region characterized by intercropping, irregular plant spacing, and complex backgrounds rarely represented in mainstream agricultural deep learning benchmarks. High-resolution RGB images were systematically acquired over maize fields in Yucatán, Mexico, and curated into specialized datasets representing parcels, individual plants, and damaged vegetation. Instance segmentation models based on the YOLOv11 architecture were trained and evaluated to extract visual information related to crop condition, while the Slicing-Aided Hyper Inference (SAHI) method was integrated to enable efficient processing of large orthomosaic images. The proposed framework achieved high performance in detecting maize plants, with a precision of 92.9% and an mAP50 of 94.2%, and demonstrated reliable identification of damage patterns associated with Spodoptera frugiperda, reaching a precision of 79.2% and an mAP50 of 71.7%. The resulting georeferenced outputs provide spatially explicit information that supports quantitative analysis of crop health and damage distribution. The results indicate that the proposed framework constitutes a scalable and reproducible approach for UAV-based visual information extraction, with potential applicability to broader agricultural monitoring and data-driven decision support systems.

Information doi: 10.3390/info17060528

Authors: Marwa Abdellah Aya Hesham Ahmad Salah Gamal M. Behery

Federated Learning (FL) is a machine learning technique that preserves data privacy and security by training models directly on decentralized edge network devices. This generates substantial communication overhead due to the repeated exchange of model updates across numerous edge network devices. Quantization has tackled this challenge by reducing communication overhead and computational costs by quantizing model updates. Although selecting the most suitable quantization level to balance communication efficiency and model accuracy is challenging, failing to achieve this balance results in excessive compression, leading to accuracy degradation due to the lossy nature of the quantization technique. This challenge was tackled in this paper via a Quantile-based lossless compression method named Pcodec, which implements lossless compression in the FL context. Pcodec is a Quantile-based lossless compression algorithm designed for numerical data that utilizes mode identification with delta encoding and binning, where binning groups similar values into entropy-coded bins and stores the exact offset within each bin, thus achieving high compression ratios and efficient processing speed. Using MNIST and CIFAR-10 datasets and models such as CNN and ResNet18, we demonstrate that Pcodec achieves up to 58.19% size reduction with no accuracy loss compared to standard quantization methods. The experiments showed that the proposed Quantile-based compression approach in FL reduces up to 2.81× the communication overhead between each server and edge network device while maintaining the accuracy. In comparison to quantization, the Quantile approach reduced the communication overhead by 2.74×, tackling the main challenge of FL context by reducing communication overhead with a remarkably high compression ratio while maintaining the model’s accuracy.

Information doi: 10.3390/info17060527

Authors: Yabin Peng Chenyu Zhou Yuchen Liu Kunlin Li Fan Zhang Shaoxun Liu

Graph neural networks (GNNs) are highly effective on relational data, yet their performance degrades sharply when graph topology is poisoned before training. Existing defenses usually assume a fixed attack pattern and a fixed graph structure, which makes them brittle when the poisoned graph changes across attacks, perturbation budgets, or deployment conditions. We propose GADD, a game-inspired adversarial distillation framework for robust graph defense. GADD first constructs multiple positive and negative graph views through a homophily-aware graph sampling scheme, allowing the model to learn from both purified and high-risk subgraphs. It then trains a heterogeneous group of student GNNs online, where each student receives global class-distribution knowledge from its peers and local structural knowledge through an adversarial cyclic distillation objective. Finally, GADD replaces uniform ensembling with an entropy-regularized adaptive aggregation rule that assigns graph-adaptive weights according to confidence and inter-model agreement. On Cora, CiteSeer, and PubMed, GADD consistently improves robustness against both Meta and Nettack attacks while preserving clean accuracy. Under the strongest Meta and Nettack settings in the main benchmark, GADD improves the best competing baseline by up to 2.99 and 3.42 percentage points, respectively. Additional ablations show that graph sampling, adversarial distillation, and adaptive aggregation all contribute materially to the final robustness gains.

Information doi: 10.3390/info17060526

Authors: Bing Qi Chunyang Ye Yuliang Ding

The widespread availability of image editing tools and generative AI has made image forgery more accessible and deceptive, demanding more advanced localization techniques. Existing CNN-based methods are limited by local receptive fields, struggling with long-range dependencies, while Transformers suffer from the quadratic complexity of self-attention, hindering practical deployment. Moreover, effectively utilizing multi-scale features remains challenging. To address these challenges, we propose a Multi-scale Wavelet-enhanced U-Mamba network (MWEU-Mamba). The proposed framework employs a Mamba-based state space model as the backbone to achieve global contextual modeling with linear complexity. A wavelet enhancement module is introduced to integrate spatial–frequency representations, improving sensitivity to subtle manipulation traces across scales, while a channel attention mechanism further amplifies forgery-relevant feature responses. Extensive experiments on six public benchmark datasets (e.g., CASIA and Coverage) demonstrate that the proposed method achieves state-of-the-art performance on multiple datasets in terms of pixel-level F1-score.

Information doi: 10.3390/info17060525

Authors: Alex Bulzan Robert Botez Virgil Dobrota

Background: Conventional blockchain ledgers are immutable by design, which conflicts with the right to erasure mandated by the European Union General Data Protection Regulation (GDPR, Article 17), particularly for identity and access management (IAM) workloads that store personal data on-chain. Methods: We integrated the National Institute of Standards and Technology (NIST) data block matrix (DBM) into Hyperledger Fabric, developed an IAM chaincode with erasure-aware role-based and attribute-based access control, formally modeled the deletion protocol in TLA+ at a bounded scope (up to 4 organizations, 3 assets, quorum =3), and evaluated the design on a local 3-VM testbed and a 20-organization cross-region Google Kubernetes Engine (GKE) deployment. Results: Hash-operation counts from 102 to 106 blocks track the theoretical O(N) bound; chaincode-level throughput reaches 78–102 TPS for individual operations and 1967–2465 TPS for batch writes, with a 305 TPS sustained measurement across 20 organizations over an 80–100 ms cross-region link. Conclusions: The design supports technical alignment with GDPR Article 17 for on-chain data, but a deterministic timing oracle on the deletion reject path and a rate-limiter gap on the request path remain open; these are disclosed as scope limitations and prioritized for structural mitigation in future work.

Information doi: 10.3390/info17060524

Authors: Elham Aldousari

Most TikTok users have low electronic health (eHealth) literacy. Yet, evidence on the availability, content, and engagement metrics of TikTok videos on eHealth literacy is scanty. This study analyzes the content and engagement metrics of recent TikTok videos on health literacy to identify the unmet eHealth literacy needs of TikTok users. A convergent mixed-methods study was conducted. The hashtag #eHealthliteracy was searched for videos published between January 2025 and February 2026. Engagement and content data were retrieved and analyzed statistically and thematically, respectively. Sixty of the 69 retrieved TikTok videos were shorter than 288 s. Favorite:like ratio was the only high engagement ratio at 15%; the rest were less than 4%. Only four of the 69 videos were generated using artificial intelligence, and their engagement metrics were the highest. Video lengths were negatively associated with all engagement metrics (β = −0.004 to −0.008, p < 0.001). Favorites, shares, and likes were significantly higher for educational videos compared to promotional videos (p = 0.0071, 0.0252, and 0.0413), with medium effect sizes of ε2 = 0.1072, 0.0743, and 0.0617, respectively. Only six of the 69 videos were directly on eHealth literacy; the rest were on health literacy. The increasing availability of health information is not accompanied by eHealth literacy on TikTok. Future eHealth literacy videos should be short, institutional, AI-optimized, and embedded with health topics for better reach and engagement.

Information doi: 10.3390/info17060523

Authors: Mario Caterino Tammaro Ucciero Riccardo Emanuele Landi Luca Mozzillo Antonio Negro Federico Saponaro Domenico Soriano Ivo Surano Salvatore Miranda Roberto Macchiaroli Marcello Fera

Predictive maintenance (PdM) plays a key role in improving the reliability and efficiency of industrial systems by anticipating abnormal operating conditions through data-driven approaches. In this context, many PdM solutions rely on temporal windowing strategies applied to multivariate time series; however, the choice of observation and prediction horizons is often fixed a priori and rarely analyzed in a systematic way. This paper investigates the impact of temporal window design by formulating alarm prediction as a window-based classification problem, explicitly distinguishing between the observation horizon and the prediction horizon. A comprehensive experimental study is conducted on three real-world industrial datasets characterized by different process dynamics. Multiple machine learning and deep learning models are evaluated across a wide range of combinations of observation and prediction horizons. The results show that predictive performance is strongly influenced by the joint configuration of observation and prediction horizons, rather than by their individual selection. While increasing the prediction horizon generally reduces accuracy, the magnitude of this effect depends on system dynamics. Similarly, longer observation horizons substantially improve performance in processes with slow dynamics, while providing limited benefits in faster-evolving systems. Overall, the study highlights the importance of explicitly modeling temporal window parameters in predictive maintenance design, providing both methodological insights and practical guidelines for selecting appropriate temporal configurations in industrial alarm prediction tasks.

Information doi: 10.3390/info17060522

Authors: Patrizia Piersigilli Rocco Citroni Fabio Mangini Fabrizio Frezza

Modern oil paintings are characterized by the extensive use of industrial pigments, synthetic binders, and chemical additives introduced during the late nineteenth and twentieth centuries. While these innovations enabled significant artistic experimentation, they also introduced new conservation challenges due to the chemical instability of many modern paint formulations. As a consequence, modern oil paintings frequently exhibit degradation phenomena such as efflorescence, yellowing, blistering, peeling and cracking, and high sensitivity to water and organic solvents. A comprehensive understanding of the materials used in modern oil paintings—including pigments, binders, and additives—is therefore essential for developing effective conservation strategies. In this context, electromagnetic (EM) diagnostic techniques represent powerful tools for the noninvasive or minimally invasive investigation of artworks. These techniques allow researchers to characterize the chemical composition, morphology, and degradation processes affecting paint layers and substrates. This paper provides an overview of the EM techniques most commonly used in the conservation of modern oil paintings. Particular attention is devoted to spectroscopic and imaging methods such as scanning electron microscopy with energy-dispersive X-ray spectroscopy (SEM-EDX), Fourier-transform infrared (FTIR) spectroscopy, Raman spectroscopy, UV-Vis spectroscopy, and X-ray-based techniques, as well as to the laser technique for the delicate cleaning process. Through selected case studies reported in the literature, this review highlights the role of these techniques in pigment identification, degradation analysis, and the development of more effective conservation strategies for modern oil paintings.

Information doi: 10.3390/info17060521

Authors: Xingxuan Li Ting Guo Zhen Tian

For the challenging scenario in which only normal node labels are available in semi-supervised graph anomaly detection, existing generative methods usually synthesize abnormal nodes through random perturbation or feature interpolation. However, these methods fail to consider node abnormality comprehensively from both structural and attribute perspectives, resulting in generated pseudo-anomalies of limited quality and insufficient reliability. In order to address this problem, we propose DGAM (dual-guided anomaly mining), a framework for selecting pseudo-anomaly nodes based on the dual-index measurement of topological anomaly and feature consistency. The core of the framework is the joint anomaly evaluation module, which quantifies node anomaly through two computable metrics. The topological boundary score (TBS) measures the boundary of a node’s topological position based on the proportion of connections between a node and labeled normal nodes in its K-hop neighborhood. The feature deviation score (FDS) evaluates the consistency of a node’s local features by calculating the average cosine similarity between its features and those of its K-hop neighbors. The module selects a fixed set of nodes with higher comprehensive anomaly scores from the labeled normal nodes as pseudo-anomalies, so as to construct a training set containing explicit supervision signals. The model adopts a shared encoder architecture and jointly optimizes the classification loss based on pseudo-labels and the embedding regularization loss of the graph nodes to learn a more discriminative node representation. Experimental results on multiple real-world graph datasets show that DGAM can stably improve anomaly detection performance, effectively verifying the effectiveness of the proposed screening mechanism and joint training strategy.

Information doi: 10.3390/info17060519

Authors: Yu-Cheng Wang

Visual explainable artificial intelligence (XAI) is an important mechanism for connecting analytically complex decision models with practitioners who must interpret and act upon their outputs in industrial supply chains. In facility localization problems, wafer foundries and other capital-intensive manufacturers must evaluate geographically dispersed candidate sites against multiple uncertain criteria. The ability to communicate fuzzy group decision-making (FGDM) outcomes in a transparent, interpretable form has direct operational relevance. The literature has introduced hanging gradient bar charts, gradient bidirectional scatterplots, and traceable aggregation charts as visual XAI instruments for semiconductor supply chain localization that show substantial reductions in interpretation error versus conventional plots. However, the quantitative assessment of explanation quality itself remains underdeveloped. To address such a gap, this research proposes a quantitative explainability quality index (XQI) that formalizes visual explanation quality in FGDM as a composite measurable construct. XQI integrates two complementary layers: (1) An objective explainability layer (OEI), consisting of normalized fuzzy interpretation deviation, response time, ranking fidelity, and interpretation accuracy, and (2) a subjective explainability layer (SEI), consisting of perceived understanding, perceived transparency, decision confidence, and cognitive load. Trust, acceptance, and decision quality are downstream outcome constructs rather than components of the index. A weighted linear combination of OEI and SEI produces a single index for systematic, reproducible comparison across competing visualization designs. A structural equation model is specified as a planned validation mechanism for examining how explanation quality may relate to trust, acceptance, and downstream decision quality. The proposed validation framework includes a semiconductor facility localization scenario, three visualization conditions, and a planned participant pool of 150–240 supply chain managers, engineers, and graduate students. The XQI framework transforms visual XAI from a descriptive communication aid into a testable decision-support construct, thereby addressing a key evaluation gap in the FGDM visualization literature.

Information doi: 10.3390/info17060520

Authors: Jiahao Li Yilin Zhai Jun Ma

The rapidly aging global population is placing immense pressure on healthcare systems, which are struggling to meet the needs of older adults. Information and communication technologies (ICTs) are considered a key driver in supporting the development of age-friendly healthcare models. This scoping review aims to map and structure the multifaceted applications of ICTs in age-friendly healthcare, focusing on their design, benefits, challenges, and implementation in different contexts. We followed the PRISMA-ScR guidelines and conducted a systematic search of five major databases (Web of Science, Scopus, PubMed, ScienceDirect, and IEEE Xplore), supplemented with backward citation chaining to improve the robustness of literature identification. The results show that ICTs can help older adults by improving their access to healthcare information, enhancing their care coordination, supporting their independent living, and personalizing their health management. Key challenges include user experience issues for older adults, data privacy and security concerns, and implementation barriers related to resources and professional support. Effective implementation of ICTs requires greater emphasis on age-centered design, robust data governance, and scalable integration with existing healthcare systems. We further propose a Technology Design–Scenario Application–Effect Evaluation (TD-SA-EE) analytical framework for ICT application in age-friendly healthcare; the framework is grounded in sociotechnical systems theory to provide explanatory insights beyond descriptive classification. This research provides insights into optimizing age-friendly healthcare through ICTs and contributes to fully leveraging ICTs in building sustainable and equitable age-friendly healthcare systems.

Information doi: 10.3390/info17060518

Authors: Rong Huang Qiqiang Wu Mingwei Yang Yanhua Liu Baokang Zhao

Anomaly detection performance in sensor data is highly sensitive to model hyperparameters, which is central to reliable monitoring in mobile Internet security and industrial IoT (IIoT) scenarios. We propose an IWOA-LightGBM-based anomaly detection method for sensor data. For machine learning-based anomaly detection methods, hyperparameter selection often determines model performance, so we propose an Improved Whale Optimization Algorithm (IWOA) and further use it to optimize the hyperparameters of the LightGBM algorithm. To avoid falling into local optima and accelerate algorithm convergence, the WOA is improved by integrating nonlinear convergence factor, adaptive inertia weight factor and stochastic differential mutation strategy. Experimental results show that during hyperparameter optimization for LightGBM model training, the IWOA achieves faster convergence and higher computational efficiency compared to the Whale Optimization Algorithm (WOA), with anomaly detection accuracy exceeding 90%.

Information doi: 10.3390/info17060517

Authors: Dimitrios Galiatsatos Panagiota Galiatsatou

Predictive analytics has become increasingly important in educational decision-making, supporting at-risk identification and adaptive tutoring. The accurate early prediction of school achievement can enable timely intervention. Using the Math Students dataset, which contains data on students from two Portuguese secondary schools, we model three categorical outcomes derived from the students’ final grade, namely the final grade level (low, medium, high), its qualitative evaluation (fail, satisfactory, good, excellent), and the final pass/fail outcome. After preprocessing, three filter methods—Correlation-Based Feature Subset Selection (CFS), Correlation Attribute Evaluation (CorrEval), and Information Gain (InfoGain)—are applied to reduce the dimensionality of the datasets. Nine classifiers (Naive Bayes, Logistic, MLP, SMO, IBk, Bagging, J48, Random Forest, Random Tree) are evaluated using ten-fold cross-validation in the Waikato Environment for Knowledge Analysis (Weka) platform. Random Forest with InfoGain achieves 90.7% accuracy on the three-band task, while Bagging with InfoGain achieves 92.5% on the binary pass/fail outcome, outperforming benchmarks in prior Educational Data Mining (EDM) studies. Results confirm that prior academic performance indicators (first- and second-period grades) and failure history dominate predictive power and contribute substantially to the success of ensemble models, particularly when paired with feature selection methods that reduce noise and highlight relevant attributes.

Information doi: 10.3390/info17060516

Authors: Qiong Hong Lailatul Qadri Zakaria Sabrina Tiun

Automated depression assessment is critical for scalable mental healthcare but faces dual challenges: the lack of clinical interpretability in “black-box” deep learning models and the excessive computational cost of large-scale fusion architectures. To bridge this gap, we propose NS-Dep-KAN, a novel neuro-symbolic framework that harmonizes DSM-5-guided reasoning with Kolmogorov–Arnold Networks (KANs). Our approach leverages a Large Language Model (LLM) to extract symbolic symptom evidence aligned with diagnostic criteria, which then guides the aggregation of multimodal features from frozen pretrained encoders (WavLM and Qwen). Unlike traditional Multi-Layer Perceptrons, the proposed KAN prediction head employs learnable B-spline activation functions to capture complex nonlinear symptom–severity mappings with extreme parameter efficiency. Evaluations on the DAIC-WOZ benchmark demonstrate that NS-Dep-KAN achieves state-of-the-art performance among audio-text models (MAE 2.69, 13.5% improvement over the three-modality baseline MSGAF at MAE 3.11), with only ∼4.9 K trainable parameters. Moreover, the framework offers inherent interpretability, revealing granular symptom contribution profiles that align with clinical intuition. This work establishes a path toward explainable trustworthy AI for mental health screening.

Information doi: 10.3390/info17060515

Authors: Muhammad Shaban Qabil Hafiza Zarafshan Mukhtiar Ghulam Mustafa Muhammad Tanvir Afzal Isabel De la Torre Díez Elizabeth Caro Montero Mirtha Silvana Garat de Marin

Researcher evaluation underpins critical academic decisions, yet traditional bibliometric indicators lack predictive capability and cross-domain generalizability, while most predictive approaches offer limited interpretability and narrow domain validation. This study proposes a SHAP interpretable, multi-domain supervised learning framework for predicting academic award recognition using thirty two publication count-based bibliometric indices. A balanced dataset was constructed across four disciplines, namely Computer Science, Neuroscience, Mathematics, and Civil Engineering, comprising verified awardees from recognized professional societies and matched non-awardee researchers. Eight classifiers were evaluated under stratified five fold cross validation, assessed via accuracy, precision, recall, F1-score, and ROC AUC. The framework achieved domain-specific F1-scores of 0.70 in Computer Science, 0.73 in Neuroscience, 0.72 in Civil Engineering, and 0.78 in Mathematics, with SVM and XGBoost demonstrating the strongest cross-domain robustness across disciplines. SHAP analysis consistently identified normalized h index, h2 family, q2 index, and g index as dominant cross-domain predictors, while domain-specific indicators, including Rm and w indices in Neuroscience and P index in Civil Engineering, reflected disciplinary recognition patterns. By unifying publication-based feature engineering, multi-domain classification, and SHAP explainability within a single reproducible pipeline, this framework offers a scalable, transparent, and evidence-based tool for institutional researcher evaluation.

Information doi: 10.3390/info17060513

Authors: Ziguo Wang Mei Nian Yaling Jing Jun Zhang

To address the rapid growth of software vulnerabilities, the latency of manual expert classification, and the limitations of existing methods restricted to fixed categories, this paper proposes V2W-LLM, an automated vulnerability-to-weakness mapping model based on Large Language Models (LLMs). First, a dataset of CVE-CWE description pairs is constructed based on established expert correlations from MITRE. Subsequently, the LLM is instruction-tuned on this dataset to leverage its reasoning capabilities in generating CWE-style descriptive text for newly disclosed, unmapped vulnerabilities. Finally, using a BAAI-based embedding model, the semantic representations of the generated text and official CWE descriptions are computed to identify the optimal mapping via cosine similarity (Top-1). Experimental results indicate that V2W-LLM achieves an accuracy of 90.18% and a Macro-F1 of 87.64% in common categories. Furthermore, on the public ChatGPT-VDMEval and the latest 2024 NVD datasets, the model attains F1 scores of 86.02% and 94.02% respectively, validating its effectiveness in automating the vulnerability-to-weakness mapping process.

Information doi: 10.3390/info17050514

Authors: Tip-aroon Kiawkaew Thanaruk Theeramunkong

Dense retrieval pipelines often rely on large candidate pools before reranking, making candidate generation and downstream scoring a practical bottleneck. This paper studies training-free binary projection filtering as a lightweight pre-filter for reducing the candidate set before dense reranking. Rather than presenting it as a universally superior retrieval method or a validated speedup technique, we ask a narrower practical question: how far can the candidate pool be reduced before average top-rank quality, retained relevance, and query-level reliability begin to break down? We evaluate the approach on five BEIR datasets: SciFact, NFCorpus, FiQA, ArguAna, and TREC-COVID. The revised evaluation compares exhaustive Dense retrieval, FAISS-HNSW, FAISS-IVF-Flat, and Binary+Dense retrieval, and includes projection-dimension ablations over Db∈{128,256,512,1000}, candidate-budget ablations over K∈{50,100,200,500}, five-seed robustness analysis, and typo-perturbed queries. In addition to MRR@10, nDCG@10, and Recall@100, we report filter-stage metrics including Retained@K, catastrophic failure rate, and Best Relevant Survival. Across datasets, Binary+Dense often remains close to exhaustive Dense retrieval in average top-rank metrics at representative operating points, but the filter-stage behavior is strongly collection-dependent. Larger Db and K generally improve retained relevance and reduce catastrophic failures, but they also increase filtering cost or reduce the degree of pruning. The latency results show that structural candidate reduction does not translate into consistent end-to-end wall-clock speedup in the current Python 3.16/NumPy implementation. Taken together, the results suggest that training-free binary projection filtering is best understood as a calibration-sensitive pre-filter and failure risk analysis mechanism rather than as a replacement for Dense or ANN retrieval.

Information doi: 10.3390/info17050512

Authors: Weining Li Songsong Li Xingshuo Yue Xu Wang Yuhang Zhu Xiaoming Chen

This paper introduces CMA-YOLO, a network that integrates multi-scale features with dual attention mechanisms to address weak feature representation, low detection accuracy, and loss of fine-grained details in deep networks for wind turbine blade surface defect detection. First, we construct the C2MSA module by designing a Multi-scale Feature-enhanced Attention Convolution Mix (MS-ACmix) based on ACmix and embedding it into the C2PSA block. This lets the network capture local and global contextual features, strengthening multi-scale target recognition and lowering missed detections. Second, we devise a Monte Carlo Dual Attention (MCDA) mechanism combining random sampling with dual attention. This approach retains the regularization benefits of the Monte Carlo method while leveraging dual attention selection, enabling improved detection accuracy with low computational cost. Finally, we substitute the original downsampling layers in the backbone and neck with the ADown module. This lightweight design, together with efficient feature extraction and fusion, reduces fine-grained detail loss and improves defect detection capability. Quantitative results reveal that, compared to YOLO11n, CMA-YOLO yields improvements of 3.4% in mAP@0.5, 6.1% in mAP@0.5:0.95, and 8.8% in recall, with a 0.7 GFLOPs reduction in computational cost, thus validating the proposed algorithm. Overall, CMA-YOLO provides a lightweight and effective approach for inspecting blade surface defects on wind turbines operating in resource-limited settings.

Information doi: 10.3390/info17050511

Authors: Yunyixuan Zhang Haoran Wang Binrong Zhu Xiaozhi Li Siyu Xia

Archery is a fine-grained skill sport in which small posture deviations can markedly affect performance, motivating the need for reliable automated technique assessment. However, most existing methods still focus on large-amplitude sports and cannot match coach-level nuance. To overcome these limitations, we introduce SEMA (Semantic Evidence-Driven Multimodal Assessment), a large language model (LLM)-based end-to-end system for fine-grained archery action quality assessment. Beyond score prediction and evaluation-text generation, SEMA further supports knowledge-grounded question answering and feedback generation through a hierarchical multi-source knowledge framework that integrates assessment outputs, structured coaching guidance, and general archery knowledge. Experimental results show that SEMA achieves strong performance on the novel AAV dataset, outperforming general-purpose VLMs and adapted prior AQA methods. In addition, we introduce the AAV (Archery Action Video) dataset, the first multimodal, fine-grained action quality assessment (AQA) dataset dedicated to archery, and release it publicly to the community. This dataset addresses a critical gap in current benchmarks for assessing archery action quality and intelligent archery training.

Information doi: 10.3390/info17050509

Authors: Zhida Fu Nobuo Funabiki Zihao Zhu Yue Zhang Wen-Chung Kao Yi-Fang Lee Pi-Kuang Tseng

Nowadays, Python is very popular as the first programming language for novices, including high school students, to learn due to its short code features with rich libraries. Thus, it is important to provide a learning environment supporting studies starting from the fundamentals, since students have no knowledge on how a program runs on a computer. Previously, we have developed a web-based programming learning assistant system (PLAS) to allow the self-study of major programming languages, including Python, by university students. It offers several types of exercise problems that have different learning goals and levels for step-by-step study. Any student answer is automatically marked at the answer interface for quick feedback. However, PLAS has not implemented functions to assist the learning needs of high school-level students. In this paper, we propose a novice-friendly answer interface for a Python programming learning assistant system (PyPLAS) that introduces a code behavior visualization and an AI assistant with learning logs. The visualization allows learners to observe the changes in variable states and the control flow. The assistant provides multi-level hints during learning and reflective feedback after it by analyzing the logs based on engagement, reasoning strategies, learning pace, and tool usage. For evaluation, we implemented the proposed interface using Python Flask for the web platform and Ollama as a locally deployed AI model. A pilot application was conducted with high school students solving introductory Python exercises in PyPLAS. The results showed high task completion, positive questionnaire responses toward embedded visualization and interface usability, and teacher-observed usefulness of the four-dimensional learning analytics for interpreting learner behaviors. These findings provide preliminary evidence for the feasibility and practical value of the proposed interface, while larger controlled studies are required to validate its instructional effectiveness.

Information doi: 10.3390/info17050510

Authors: Pengzhuo Wu Yao Wang Guodong Li

In the digital economy era, digital inclusive finance represents a paradigmatic reconstruction of key economic information flows. This study integrates multi-source panel data of 27 cities in the Yangtze River Delta from 2011 to 2023. By constructing an economic geography composite spatial weight matrix and a nonlinear spatial panel model, this study analyzes the impact of the diffusion of digital inclusive financial information on industrial structure upgrading. The results show that: (1) digital financial inclusion exerts a significant direct effect and spatial spillover effect on industrial structure; (2) the local effect exhibits a “U-shaped” curve with an accelerating characteristic on the right side; the spatial spillover effect demonstrates an “inverted U-shaped” curve, revealing the transformation law and threshold effect of the diffusion and aggregation of digital financial information benefits; (3) digital payment and digital credit constitute the core information flows driving the coordinated upgrading of industries; and (4) entrepreneurial activity exerts a partial mediating effects, and exhibits a spatial mediating effect, while the technological innovation only demonstrates a significant local mediating effect. The findings provide quantitative evidence to support the optimization of the digital financial information ecosystem and the realization of coordinated industrial upgrading in the Yangtze River Delta.

Information doi: 10.3390/info17050508

Authors: Orkun Demirbağ Halil İbrahim Kaymak Hale Alan Ferhan Akdeniz

In today’s dynamic business environment, organizations are increasingly investing in strategies to protect themselves against information system violations. While these technologies offer remarkable benefits—boosting efficiency, productivity, and overall performance—they also bring significant risks, particularly regarding information security breaches. This study delves into the critical connections between technostress, techno-strain, and the violation of information security policies. Our research aims to shed light on how technostress, which is commonly experienced by engineers working in technology-intensive environments within the IT sector, drives information security violations. Importantly, we will also explore how techno-strain mediates this relationship. By focusing on engineers who are consistently engaged with advanced technology, we seek to answer essential questions about their experiences. It is worth noting that the requirements for information security technology can widely vary based on factors such as industry type, organizational structure, departmental roles, and cultural norms. Therefore, this study examines how technostress increases the likelihood of information security policy violations and how techno-strain mediates this relationship. Looking ahead, future research should consider both the broader institutional contexts and the individual characteristics that may shape the relationship between information security violations and technostress. Furthermore, understanding the repercussions of information security violations—stemming from technostress—on a company’s financial health is vital for organizations aiming to safeguard their assets and maintain a competitive edge. Emphasizing these insights can lead to more effective strategies for managing both technology and talent in the workplace.

Information doi: 10.3390/info17050507

Authors: Yidi Zhang Xuqiu Wei

(1) Background: Knowledge diffusion of academic publications has become a crucial indicator of research impact in the context of open science, yet its influencing factors and underlying mechanisms remain insufficiently studied. (2) Methods: Based on literature research, this study constructed a multi-dimensional factor system consisting of 17 factors. The DEMATEL method was used to identify the key influencing factors. The mean and standard deviation of the comprehensive influence matrix were taken as the threshold (λ = 0.87) to filter important relationships and establish an adjacency matrix. The ISM method was used to explore the hierarchical relationship of the influencing factors, and finally, strategies were proposed. (3) Results: The results show that the journal characteristics have the highest centrality (30.292), the author characteristics have the strongest causal effect (0.686), and the language characteristics are at the bottom of the influence factor hierarchical structure model, possessing certain driving force and being closely related to multiple factors. (4) Conclusions: Effective enhancement of knowledge diffusion requires coordinated optimization across language expression, core research elements, methodological execution, dissemination channels, and presentation formats, providing theoretical and practical implications for academic evaluation and research management.

Information doi: 10.3390/info17050506

Authors: Mario Alviano Paola Guarasci

DTDL (Digital Twins Definition Language) provides no mechanism for logical reasoning or constraint checking over digital twin models. We integrate DTDL with ASP Chef, a web-based Answer Set Programming (ASP) platform, via a structured DTDL-to-ASP mapping and three dedicated operations: @DTDL/Parse for fact generation, @DTDL/Analysis for structural metrics, and @DTDL/Debug for symbolic validation with LLM-guided repair. The key design decision is that error detection is symbolic and deterministic within the implemented set of constraint classes; a language model is invoked only after the ASP layer has produced a concrete, grounded diagnostic, keeping the correctness boundary with the symbolic layer. Soundness and completeness guarantees are scoped to these constraint classes; a formal proof is left as future work. We illustrate the framework on two agricultural use cases and report a proof-of-concept assessment on 99 diagnostics spanning 21 error classes across four domains. Three binary metrics are used: json_valid and entity_recall are computed mechanically; fix quality (judge_correct) is assessed by an independent LLM judge (Claude Sonnet 4.6). The complete grounded workflow achieves 90% judge_correct and 86% json_valid; a fair ablation baseline—same LLM and system message, but error type and entity name in natural language without structured diagnostics—achieves 77% and 75%, respectively. The gap is consistent across three independent judges and statistically significant (McNemar p<0.01), but the inter-judge reliability of judge_correct is limited (κ ranging from 0.00 to 0.44), so results should be read as directional evidence rather than precise effect estimates. Excluding the dominant isolated_interface class (n=28, ceiling score), the conservative estimate is 87% vs. 79% on the remaining 71 diagnostics. These results constitute a preliminary proof-of-concept limited to a small number of models, a few application domains, and a single LLM configuration; results do not generalize beyond this specific setting. The judge_correct metric is assessed by LLM-as-judge and does not carry a perfect inter-annotator agreement.

Information doi: 10.3390/info17050505

Authors: Grace Byun Swati Rajwal Jinho D. Choi

Large Language Models (LLMs) are increasingly explored for educational tasks such as grading, yet their alignment with human evaluation in real classrooms remains underexamined. In this study, we investigate the feasibility of using OpenAI GPT-4o to evaluate short-answer quizzes and project reports in an undergraduate Computational Linguistics course. We collect responses from approximately 50 students across five quizzes and receive project reports from 14 teams. LLM-generated scores are compared against human evaluations conducted independently by the course teaching assistants (TAs). Our results show that GPT-4o achieves strong correlation with human graders (up to 0.98) and exact score agreement in 55% of quiz cases. For project reports, it also shows strong overall alignment with human grading, while exhibiting some variability in scoring technical, open-ended responses. We release all code and sample data to support further research on LLMs in educational assessment. This work highlights both the potential and limitations of LLM-based grading systems and contributes to advancing automated grading in real-world academic settings.

Information doi: 10.3390/info17050504

Authors: Angel A. Juan Antoni Guerrero Marc Escoto Javier Panadero Alvaro Garcia-Sanchez Mauricio G. C. Resende

This paper proposes the concept of biased-randomized transformers, a novel methodology that combines biased-randomized techniques and transformer-based deep learning for ‘agile’ optimization (i.e., real-time optimization that is carried out iteratively in dynamic systems). On the one hand, biased-randomization techniques have been used in the past to inject controlled randomness into greedy heuristics, thus converting them into probabilistic algorithms capable of generating thousands of good-quality solutions while preserving heuristic logic. On the other hand, transformer models can capture complex patterns across thousands of variables. Once trained, these models can be seen as data-driven heuristics able to provide fast solutions to new instances and adapt to changing inputs. The combination of biased-randomization techniques with trained transformers allows for a fast exploration and selection of the high-quality solutions to NP-hard combinatorial optimization problems. The paper includes two case studies that illustrate the potential of these biased-randomized transformers.

Information doi: 10.3390/info17050503

Authors: Joaquim Jose Carvalho Proença Nelson Jesús Campos Rosendo Soratna Veronica Navas Gotopo

This study aims to provide a comprehensive cross-regional bibliometric analysis of artificial intelligence (AI) research in East Asia and Latin America from 2020 to 2025. By quantifying publication trends, authorship, institutional productivity, collaboration networks, and citation impact, the research seeks to identify regional leaders, thematic clusters, and disparities in visibility and impact between these two regions. Design/methodology/approach; Scopus-indexed publications containing the phrases “artificial intelligence research” or “artificial intelligence innovation” in their title, abstract, or keywords were retrieved for the period 2020–2025. Inclusion criteria required at least one author’s affiliation in any of the fourteen specified countries across East Asia or Latin America. All document types (articles, reviews, conference papers, book chapters) were considered. Metadata were manually extracted from Scopus database ranking to identify the top-cited papers, most prolific authors, leading institutions, thematic and subject-area concentrations, and crossnational collaboration patterns. Findings; this bibliometric review clarifies the dynamic trajectory of AI research in East Asia and Latin America, revealing significant disparities in productivity, visibility, and thematic focus. The findings underscore the need for targeted investments in research capacity building, strategic international partnerships, and thematic realignment particularly for Latin America to enhance global visibility and align with emerging AI trends. Originality; by contrasting two understudied regions (East Asia vs. Latin America), we capture shifts in the AI landscape—specifically, the generative AI boom across subfields and regions that no single region or pre 2022 study can. By highlighting structural disparities in productivity, citation impact, and institutional support, it offers policymakers, funding agencies, and academic leaders novel insights.

Information doi: 10.3390/info17050501

Authors: Maximilian Nawrath Andrea Merlina Jemmima Knight Sam A. Welch Mahla Rashidian Isabel Seifert-Dähnn

Literature screening is a major bottleneck in systematic reviews, yet Large Language Models (LLMs) can substantially reduce workloads. However, performance varies across models and is sensitive to evaluation metrics, particularly in low-prevalence screening contexts. We validated five LLMs (GPT-4.1, Claude 3.5 Sonnet, Gemini 2.0 Flash, DeepSeek V3, and Mistral Large) against a 500-record gold-standard dataset (8 inclusions; 1.6% prevalence) using a conservative zero-shot prompt aligned with standard systematic review workflows. Performance was assessed through classification metrics (sensitivity, specificity, precision), logistic regression (GLM; Firth-penalised where separation occurred), and agreement indices (Cohen’s κ, MCC, PABAK, Gwet’s AC1). Gemini 2.0 Flash and Mistral Large showed no false negatives (1.00) but differed in specificity (0.858 vs. 0.697) and accuracy (0.860 vs. 0.702). GPT-4.1 and Claude 3.5 Sonnet performed identically (sensitivity 0.875; specificity 0.876; accuracy 0.876). In contrast, DeepSeek V3 maximised specificity (0.980) and accuracy (0.970) but demonstrated lower sensitivity (0.375). Regression analyses confirmed strong positive associations with human decisions (OR 28.9–49.5). Agreement indices revealed the expected low-prevalence artefact, with Cohen’s κ low despite high concordance while MCC, PABAK, and AC1 indicated substantially stronger agreement. Our results highlight a fundamental sensitivity-specificity trade-off, with conclusions dependent on the evaluation framework chosen. LLMs may meaningfully support title-abstract screening as decision-support tools, provided that human oversight is maintained and validation is transparent and reproducible.

Information doi: 10.3390/info17050502

Authors: Xi Zhao Jungang Yang

The digital transformation of high-end equipment is not only a critical means to enhance national core competitiveness, but also a necessary requirement within the framework of national development strategy. Major stakeholders in this transformation include local governments, high-end equipment manufacturers, financial institutions, and industrial technology platforms, all of whose interactions significantly influence the transformation process. This paper constructs a four-party evolutionary game model involving local governments, high-end equipment manufacturers, financial support institutions, and industrial technology platforms. Numerical simulations are conducted to analyze the stable strategies and evolutionary trends of these four players under various parameters, while also exploring the long-term mechanisms for the digital transformation of high-end equipment facilitated by government subsidies. The results indicate that in the initial stage of digital transformation, the government assumes a leading role by implementing high-subsidy policies to encourage participation from manufacturers, financial institutions, and technology platforms. As the transformation progresses into a stable promotion phase, the government gradually reduces subsidies to a normal level and increasingly relies on market mechanisms to foster active engagement. Both models represent ideal scenarios for the digital transformation of high-end equipment. Finally, this paper offers relevant policy recommendations aimed at enhancing policy guidance, stimulating the motivation of market entities, and improving the benefit linkage mechanism among all four stakeholders.

Information doi: 10.3390/info17050500

Authors: Chin-Wen Liao Nguyen Van Thanh Yi-Hsin Tai

As Industry 4.0 transforms manufacturing operations, maintenance organizations face a group decision-making problem: how to consolidate diverse expert judgments into a defensible, transparent ranking of the competencies that maintenance personnel most need. This paper applies an integrated Delphi–AHP framework—with explicit notation, operators, and diagnostics—to prioritize maintenance competencies in advanced-manufacturing settings. The Delphi stage consolidates expert-generated items under median–interquartile-range consensus and round-to-round stability rules, while the Analytic Hierarchy Process (AHP) transforms validated pairwise comparisons into ratio-scale priority weights through geometric-mean Aggregation of Individual Judgments (AIJ) and eigenvector derivation. Consistency screening (CI/CR), inter-rater agreement (Kendall’s W), and perturbation-based sensitivity analysis accompany the resulting weight vector. A bounded AI-assisted consistency-check step supports terminology harmonization during Delphi statement consolidation, subject to explicit human-validation constraints. A panel of fifteen industry experts participated in the study; five competency dimensions and twenty-nine indicators were retained through three Delphi rounds. AHP weighting identified Basic Knowledge and Skills as the highest-priority dimension, followed by Safety and Regulation Awareness and Problem-Solving Ability. Aggregated pairwise comparison matrices, local and global weights, and sensitivity results are reported to support reproducibility. The study contributes a rigorously specified application of combined Delphi–AHP to a domain—Industry 4.0 maintenance asset management—where multi-criteria decision analysis has seen limited formal application, and closes common specification gaps in published Delphi–AHP implementations.

Information doi: 10.3390/info17050499

Authors: Sangheethaa Sukumaran Arun Korath Gowri Arun Menon

Machine learning models increasingly rely on continuously generated sensor data for automated decision-making in Internet of Things (IoT) environments. The distributed and often insecure nature of IoT infrastructures introduces risks related to data manipulation, lack of traceability, and unverifiable model evolution. Existing solutions typically address isolated aspects such as data security or access control but do not provide end-to-end provenance across the machine learning lifecycle. This paper proposes a tamper-evident data and model provenance framework for IoT-based machine learning that integrates blockchain with off-chain storage. The framework records cryptographic hashes and metadata of data, preprocessing outputs, and trained models on-chain while maintaining large artifacts off-chain to ensure scalability. Smart contracts establish verifiable linkage among lifecycle artifacts and automate provenance registration. The framework is evaluated in a simulated IoT–ML pipeline under integrity attack scenarios including data manipulation, model tampering, and metadata modification. Experimental results demonstrate reliable detection of unauthorized modifications with low verification latency and constant on-chain storage per record under controlled conditions. These findings indicate the feasibility of hybrid blockchain architectures for tamper-evident provenance in IoT-based machine learning systems, while highlighting the need for further validation in real-world deployments.

Information doi: 10.3390/info17050498

Authors: Raghav Mantri Saun Chen Yixuan Wang Duygu Ataman

As large language models (LLMs) scale to cover more languages, their potential to support low-resource settings becomes increasingly promising. However, the mechanisms underlying cross-lingual transfer and the factors that facilitate it remain insufficiently understood. Prior work has highlighted the role of linguistic similarity—particularly syntactic structure—in enabling transfer across languages. In this study, we present a broad empirical analysis of how multilingual LLMs encode and relate structural information across languages with varying typological properties. We combine multiple complementary methods, including hidden-state similarity analysis, typological correlation, probing for syntactic features, and attention-based structural comparisons, across four multilingual models and thirteen languages. Our findings show consistent correlations between representational similarity and syntactic relatedness, suggesting that structural properties of language influence how information is organized and shared across languages. We further observe that attention-derived structures exhibit partial alignment with gold-standard syntax, though this alignment should be interpreted as heuristic rather than direct evidence of syntactic encoding. Overall, our results provide a comparative empirical perspective on cross-lingual structural bias in multilingual LLMs and highlight the importance of careful methodological interpretation when linking representation geometry to linguistic structure.

Information doi: 10.3390/info17050497

Authors: Praveen Kumar Reddy Kudumula P. Balachennaiah

The growing integration of solar, wind and battery energy storage (BES) of the microgrids (MGs) has increased the necessity of real-time energy management, especially in the multi-microgrid (multi-MG) setting, where the generation and the load change stochastically. This paper presents a Java Agent DEvelopment (JADE)-based Multi-Agent System (MAS) for real-time energy management of a low-voltage hybrid multi-MG system incorporating solar photovoltaic (PV), wind generation, and battery energy storage (BES). The proposed framework’s novelty lies in its physical campus-scale hardware deployment—validated across four operating scenarios (single MG off-grid, single MG on-grid, dual MG off-grid, and dual MG on-grid)—combined with autonomous inter-MG power sharing, which distinguishes it from existing simulation-only MAS-based microgrid studies. The suggested framework facilitates decentralized communication between interconnected MGs and the utility AC grid to facilitate the proper management of power flow, its exchange, and the reliability of the system. The intelligent agents are used to coordinate solar, wind, BES, and load changes in order to adjust to changing demand conditions. The system is physically implemented on a campus rooftop with two 1 kW solar PV arrays and two 1.5 kW wind turbine generators, each paired with a 24 V, 150 Ah battery bank, operating on a 24 V DC bus. Results across 24 h real operational profiles demonstrate effective power balance maintenance, renewable energy maximization, and constraint-compliant battery operation (SOC is bounded within 20–90%). A direct comparison with a conventional centralized JavaScript-based EMS confirms equivalent dispatch accuracy while demonstrating superior scalability, fault tolerance, and modularity of the proposed JADE MAS architecture.

Information doi: 10.3390/info17050496

Authors: Muhammad Azhar Naureen Riaz Waqar Azeem Deshinta Arrova Dewi Adeen Amjad Muhammad Arman

The ability to identify emotions based on written text is one of the core areas of Natural Language Processing (NLP) and has many applications in areas such as mental health monitoring, sentiment analysis, and dialogue systems. This study proposes an explainable emotion recognition (EER) framework built on a fine-tuned RoBERTa-base model trained on the Emotions for NLP dataset with an accuracy of 92.4% and a weighted F1 score of 92.5%. To interpret the decision process of the EER model, we systematically applied four complementary explainable artificial intelligence (XAI) techniques to provide explanations and insights into how the model makes its predictions: SHAP for global token-level feature attribution, LIME for local instance-level explanations, multi-head attention visualization for structural interpretability, and integrated gradients via Captum for axiom-satisfying gradient-based attribution. Each of these four methods provides complementary multi-perspective views of EER model behavior, which can help increase model transparency, identify potential biases, and enable the responsible use of transformer-based models in critical environments (e.g., those requiring formal clinical documentation). Our experiments consistently show that the EER model identifies tokens as having the highest emotional expression level as the strongest predictive feature across methodological perspectives, with strong evidence of cross-methodological agreement regarding the semantic coherence of learned representations. Our findings have direct implications for the responsible implementation of AI-based emotion recognition systems in mental health support systems, where model user-interface transparency, bias mitigation, and clinical trust are necessary to ensure quality patient care.

Information doi: 10.3390/info17050495

Authors: Jingxue Li Youping Teng Weijia Wang

As carriers of traditional culture, Chinese idiom allusions contain rich semantic and emotional content. High-quality illustrations of these idioms hold significant potential for applications in cultural communication and education. Although generative artificial intelligence has achieved substantial progress in general image synthesis, it remains challenging to produce idiom illustrations in culture-intensive scenarios that simultaneously preserve cultural symbols, maintain affective ontology, and exhibit high visual aesthetic quality. To address this gap, we propose a three-dimensional evaluation framework—Zhen-Shan-Mei (Truth-Goodness-Beauty)—for idiom illustrations. The ‘Truth’ module uses Chinese vision–language models to quantify cultural symbols; the ‘Goodness’ module applies cross-modal affective analysis to assess affective ontology; and the ‘Beauty’ module computes quantitative aesthetic metrics (composition balance, color harmony, and line expressiveness). Based on this system, an AI-idiom prototype system is constructed to realize closed-loop iteration of generation-evaluation-regeneration and threshold screening. Experiments show that the proportion of illustrations selected by subjects after the “Truth-Goodness-Beauty” screening reaches 78.1%. The results suggest that the proposed method is effective in maintaining cultural symbols, strengthening affective ontology, and improving visual aesthetics and offers a potentially interpretable and reproducible evaluation and optimization framework for culture-intensive image generation tasks.

Information doi: 10.3390/info17050494

Authors: Hui Zhu Bin Xiao Yun Li Xiaolong Li

Aiming at the problems of low information management levels and low reuse rate of massive heterogeneous geological data of the Jiangxi Geological Bureau, a Jiangxi geological big data platform based on cloud service and big data technology was designed to realize the integration and sharing of Jiangxi geological big data. Firstly, the architecture of the Jiangxi geological big data platform is designed based on hierarchical thinking, including the infrastructure layer, data layer, platform service layer, application layer and user layer from the bottom up. Secondly, the key technologies for building a Jiangxi big data platform are described, including multi-layer service aggregation, geographic information service bus, geocoding service, Spark big data technology and elastic scaling technology. Finally, the main functions of the Jiangxi geological big data platform are introduced, including a platform portal website, a mobile portal system, a geological big data comprehensive analysis system and a geological 3D modeling system. The operation results of the platform show that the Jiangxi geological big data platform can effectively manage the massive heterogeneous geological data of the Jiangxi Geological Bureau and mine the value of the data.

Information doi: 10.3390/info17050493

Authors: Leila Rzayeva Madi Shayakhmetov Olzhas Konakbayev Gul Gabdulualitovna Jussupova Igor Seniushin Anara Tasbolat

Analog digital video recorders (DVRs) are still extensively used in small-to-medium business and home security systems, but there are special problems when it comes to forensic recovery of video evidence in these systems that are not covered by tools or methodology. Compared to the IP-based network video recorders, analog DVRs packetize video frames of several coaxial-connected cameras into a single interleaved binary stream on disk, necessitating channel demultiplexing before single camera footage can be reassembled. In this paper, we discuss a multi-channel analog Dahua DVR system utilizing the DHAV frame format, with a focus on the forensic recovery approach. Three significant contributions are presented in the methodology: (1) a channel demultiplexing algorithm that separates interleaved frames with up to 32 cameras on the basis of embedded channel identifiers and temporal coherence analysis; (2) a frame sequence stitching mechanism to reassemble continuous video segments on the basis of non-contiguous disk fragments using adaptive frame number tolerance (±3 frames) and temporal validation (≤1 second difference); and (3) a native C implementation with Win32 GUI providing significant performance improvements over interpreted alternatives. The system was tested on 14 analog Dahua DVR hard drives of various models, with a 92.3% recovery rate (97.1% on hard drives with no hardware damage), 91.3% temporal accuracy, 97.5% channel separation accuracy and a 1.8% false positive rate. The methodology fills an important gap in the literature of surveillance forensics, where current studies have only concentrated on IP-based digital systems, and analog DVRs form an estimated 35–40% of operational surveillance systems across emerging markets. The channel demultiplexing capability, which is not found in any current commercial or academic tool, enables automated per-camera organization of interleaved streams, converting what was previously a manual multi-day process into an automated one.

Information doi: 10.3390/info17050492

Authors: Yusaku Kato Shoko Imaizumi

In this paper, we propose a block-distortion-free reversible data hiding method for encryption-then-compression (EtC) images that supports fully flexible access privileges without constraints on the restoration order. The proposed approach redesigns the pre-processing strategy of previous work to ensure a clear separation of processing roles between the image owner and the data hider. It also introduces a pixel-value modification process that divides the target range into two regions to mitigate the influence of negative–positive inversion during restoration. As a result, block distortion in marked images is eliminated while preserving role separation between the image owner and the data hider. The proposed method offers four key advantages: flexible access privileges, elimination of block distortion, explicit role separation, and competitive hiding capacity comparable to existing methods with flexible restoration capabilities. Experimental results demonstrate that the proposed method achieves a high marked-image quality and competitive hiding capacity while maintaining the compression performance of marked EtC images. Furthermore, security analysis confirms the robustness of the generated EtC images against a representative ciphertext-only attack.

Information doi: 10.3390/info17050491

Authors: Sikun Wang Jing-Wein Wang Cunwei Lu

With the growing demand for intelligent monitoring in land-based aquaculture, rapid and accurate fish counting from visual data has become important for stocking density regulation, feeding management, and production decisions. To address the challenges in above-water fish images, including scale variation, severe occlusion and adhesion, blurred boundaries, and frequent switching between low- and high-density scenes, this study proposes MultiTask-Fish, a shared backbone multitask counting method. The network uses ResNet34 as the backbone and integrates a feature pyramid network and channel attention to learn unified feature representations. It jointly predicts detection heatmaps, foreground masks, separation boundaries, density maps, density gating, and global count regression, allowing the model to combine local localization cues, structural information, and global statistics. Based on existing polygon annotations, heatmap, mask, boundary, and density supervision are automatically generated for integrated multitask training. Experiments on 495 fish images, including 346 training and 149 validation images, showed that the proposed method achieved an MAE of 5.875, an RMSE of 11.839, and an MAPE of 0.152 on the validation set, while reducing the MAE on the high-density subset from 16.717 to 13.895. These results demonstrate its effectiveness for fish counting in complex above-water aquaculture scenes.

Information doi: 10.3390/info17050490

Authors: Swetha Pandithahalli Mahadevaswamy Prasanna Bantaganahalli Thimmappa

WSNs are a core technology that enables real-time sensing and data collection in most applications; however, because of the uncontrollable nature of their open deployment environments, they are susceptible to severe security risks. The node clone attacks are the most dangerous: a malicious individual physically captures a legitimate sensor and steals its stored credentials and introduces several replica nodes into the network. These clones have legitimate identities, and hence, the clones act as legitimate members and can disrupt data streams, disrupt routing and affect general network reliability. Addressing this menace is not easy since sensor equipment has limited resources. Carefully, detection algorithms have to be energy efficient, friendly to memory, and usable in a large network. In the given paper, it is suggested to implement a detection framework that consists of a combination of Spatial Distributive Clustering (SDC) and a Block Ensemble Neural Network (BENN). SDC clusters node features based on spatial layout and behavioral patterns, which minimizes redundancy of data and enhances the quality of information that is inputted by the classifier. BENN then undergoes an ensemble-based classification to be able to differentiate cloned and legitimate nodes. Validation of the experimental results of the SDC-BENN framework with conventional classification metrics indicates that it can be used to ensure a high detection rate with minimal communication overhead, which is of high benefit in terms of enhancing the security of WSNs.

Information doi: 10.3390/info17050489

Authors: Satoshi Aoki Hisashi Koga

The performance of pattern classification is affected significantly by feature selection. However, for security applications, selecting proper features is difficult, as malicious software continuously changes its characteristics. Thus, compression-based pattern recognition has attracted much attention because it does not require explicit feature selection to design proper distance measures. LZJD (Lempel–Ziv Jaccard Distance), in particular, has been useful for malware classification, as it computes compression distances without actually compressing objects and is suitable for handling large files like malware. LZJD extracts a compression dictionary for every object in advance and estimates a similarity between two objects by comparing their compression dictionaries. However, LZJD ignores the similarity between words in a compression dictionary. As a result, even if the dictionary has many similar words, they are simply processed as different words. To exploit the similarity between words, we propose to remove the last characters of words in the dictionary and to unify similar words that share the same prefix. This unification of words turns the compression dictionary into a multiset of words. Hence, our compression distance is named MLZJD (Multiset LZJD). In addition, the unification of words in MLZJD decreases the number of word kinds in compression dictionaries and contributes to speeding up the distance computation. We experimentally show that MLZJD halves the execution time as compared with LZJD, while hardly damaging the classification accuracy. Even on condition that the compression distances are approximated with Min-Hash, MLZJD achieves a much shorter running time than LZJD, while retaining almost the same classification accuracy as LZJD.

Information doi: 10.3390/info17050488

Authors: Rosen Ivanov

This article examines how MongoDB optimizes aggregation pipeline queries, focusing on two mechanisms: a trial-based plan selection process that runs candidate execution plans in parallel and picks the one returning the most results for the least work, and rule-based operator rewriting by the Pipeline Optimizer. The study tests nine aggregation query types on a synthetic e-commerce dataset with 50K documents, using MongoDB versions 6.0.3 and 8.2.5 under identical conditions. For each query, all valid operator orderings are evaluated together with the physical execution plan and the Pipeline Optimizer output. Each test runs 20 times with the plan cache cleared before every run. The study also tests scalability with datasets of 150K and 250K documents. Three cases are identified where the rule-based optimizer falls short: IXSCAN preference bias at low selectivity, where the suboptimal plan is up to nine times slower than the optimal (80 ms vs. 699 ms at 250K under MongoDB 8.2.5), unbounded document multiplication after $unwind, and failure to account for $group output cardinality. MongoDB 8.2.5 improves performance in most cases compared to version 6.0.3. $match + $group queries run up to 28% faster. Queries that rely on IXSCAN improve by up to 18%. Unbounded projection operations run slower in MongoDB 8.2.5 at all tested sizes. The slowdown is +23% at 50K, +3% at 150K, and +14% at 250K, pointing to a change in the projection execution path between versions.

Information doi: 10.3390/info17050487

Authors: Xiyu Zhang Xu Chen Yang Wang Yifeng Hong Yuntian Bai

Source-free domain adaptation (SFDA) for cross-modality prostate MRI segmentation is challenging because source data are unavailable and pseudo-labels on target ADC images are often noisy. To address this problem, we propose Curriculum Anatomy-guided Learning Method with Population Template Priors (CALM), a source-free adaptation framework for this task. CALM constructs a population template prior from target predictions using top-k consensus aggregation and cross-round exponential moving average, then combines this prior with instance-level predictions through Soft-AND fusion. A high-confidence background constraint is further introduced to provide reliable negative supervision, and a coverage-driven curriculum is used to expand training from easy to hard cases based on pseudo-label/template agreement. This design forms an iterative process in which prior refinement and sample-reliability refinement reinforce each other during adaptation. Experiments on the PI-CAI dataset under the T2W-to-ADC setting show that CALM achieves an average Dice score of 73.63% and outperforms representative SFDA baselines in both segmentation accuracy and boundary quality. Ablation and model analyses support the contribution of each component. These results suggest that population-level anatomical priors can provide practical structural guidance for source-free cross-modality adaptation.

Information doi: 10.3390/info17050485

Authors: Feruza Malikova Valery Lakhno Zhuldyz Alimseitova Myroslav Lakhno Kuljan Togzhanova Gulzhanat Beketova

Advances in quantum computing have created a serious threat to modern asymmetric cryptosystems protecting heterogeneous critical information infrastructures (CIIs). During this transition period, the primary threat is the “Harvest Now, Decrypt Later” (HNDL) temporal strategy of attackers, which requires the forced migration of CIIs to post-quantum cryptography (PQC) algorithms. However, such migration is associated with nonlinear “technological friction.” This will manifest as a drop in the performance of legacy systems, such as SCADA. In the context of deep cross-industry integration, this can trigger avalanche-like cascading CII failures. This article presents a model of a zero-sum differential game between a CII defender and an attacker (APT group). Using Pontryagin’s maximum principle and the Forward–Backward Sweep Method (FBSM) iterative algorithm, a saddle point was found that determines the equilibrium trajectories of limited resource allocation over a given planning horizon for the CII transition to PQC. The results of the computational experiment demonstrated that isolated sectoral migration is ineffective. It is shown that optimal control requires cross-sector synchronization to prevent cascading degradation of the CII. The proposed mathematical framework provides a practical toolkit for strategic IT budget planning and national security risk management in anticipation of quantum supremacy (Q-Day).

Information doi: 10.3390/info17050486

Authors: Shiwen Li Ruifeng Shi

In network intrusion detection, class imbalance, the scarcity of minority-class attack samples, high feature dimensionality, and substantial feature redundancy are prevalent issues that limit the detection capability of intrusion detection models. To address these issues, this paper proposes a network traffic anomaly detection method based on a Variational Autoencoder and a Conditional Wasserstein Generative Adversarial Network (VAE-CWGAN). First, a feature selection strategy that combines ANOVA and mutual information is employed to select informative network traffic features, thereby improving the discriminative capability of the input features. Second, a minority-class sample generation model that integrates VAE and CWGAN is constructed. The VAE is used to learn the latent distribution characteristics of minority-class attack samples, while class-conditional constraints and the Wasserstein distance are introduced to generate high-quality synthetic minority-class samples, thereby alleviating class imbalance in the training dataset. Finally, Random Forest (RF), a representative machine learning classifier, is adopted for the classification experiments. Experimental results on the NSL-KDD dataset demonstrate that the proposed method performs well in minority-class attack detection, achieving Precision, Recall, and F1-score values of 95.89%, 75.18%, and 84.28% for the R2L class and 77.08%, 55.22%, and 64.35% for the U2R class, respectively.

Information doi: 10.3390/info17050484

Authors: Marco Parrillo Elia Santoro Luigi Laura Valerio Rughetti

Facial expression recognition (FER) is a well-established task in computer vision, yet its application to non-photorealistic domains, such as anime and manga, remains largely underexplored. The stylized, exaggerated, and often non-proportional facial features of illustrated characters present unique challenges for deep learning models trained predominantly on realistic imagery. In this work, we construct a balanced dataset of 3000 manga and anime face images spanning six emotion categories (Angry, Embarrassed, Happy, Manic–Euphoric, Sad, Scared) and conduct a systematic comparison of two major deep learning paradigms: Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs). Specifically, we evaluate ResNet-18, ResNet-50, ViT-B/16, and ViT-S/16 under four fine-tuning strategies: linear probing, partial fine-tuning, full fine-tuning, and progressive unfreezing, enabling a controlled comparison of both architectural families and transfer learning depth. Our results show that fine-tuning strategy significantly impacts performance: the best configuration (ViT-B/16 with progressive unfreezing) achieves 81.33% test accuracy (single run, seed 42), compared to 61.33% for the weakest linear probe baseline (ViT-S/16), a gap of 20.00 percentage points. To isolate architectural differences from strategy effects, we note that under full fine-tuning, the only strategy applied identically to all four models, ViT-S/16 (76.00%) outperforms ResNet-18 (74.44%) by 1.56 percentage points and ViT-B/16 (74.22%) by 1.78 percentage points, confirming a modest but consistent architectural advantage for Transformers once backbone adaptation is permitted. Vision Transformers benefit disproportionately from fine-tuning, and the relative ranking of architectures changes across fine-tuning regimes. Confusion matrix analysis reveals persistent cross-class confusion between visually similar emotions (e.g., Happy vs. Embarrassed), while the highly distinctive Manic–Euphoric category is consistently well recognized across all architectures. To the best of our knowledge, this is the first work to conduct a controlled multi-architecture, multi-strategy transfer learning benchmark specifically for FER in anime and manga, revealing findings that are not predictable from photographic FER literature and that carry direct practical implications for model selection in non-photorealistic visual recognition tasks. The anime and manga domain provides a uniquely controlled testbed for studying transfer learning under deliberate stylization, where the domain gap from realistic imagery is not an artifact of image degradation or environmental noise but a principled artistic choice with codified visual conventions; observing that fine-tuning depth dominates architectural choice in this domain suggests the same conclusion likely holds in other non-photorealistic transfer scenarios such as medical illustrations, architectural drawings, and synthetic training data.

Information doi: 10.3390/info17050483

Authors: Iuliana Malina Grigore Azin Moradbeikie Allegra Francesca Rosso Alan Del Piccolo Dario Campagna Sylvio Barbon Junior

Railway information systems routinely register the displacement of trains across the network as sequences of station passages and segment traversals. This paper proposes a design-oriented framework that systematically transforms such train displacements into event logs to enable established process mining analyses. Here, design-oriented means that the event log is not assumed to be readily available, but is explicitly constructed from railway records through modelling choices grounded in operational semantics. The framework comprises: (i) an eventization pipeline that maps displacements to semantically precise events with explicit lifecycle and case notions; (ii) construction of a timetable-derived reference model representing planned control flow; and (iii) a structural comparison and variant analysis stage that identifies execution-level deviations from the timetable-derived reference and organizes them into recurrent behavioural patterns. The paper contributes design principles for mapping train displacements into process-mining events, a timetable-derived representation of expected control flow, and an empirical demonstration on real-world railway data showing how this framework supports operational process analysis.

Information doi: 10.3390/info17050482

Authors: Yinghui Meng Jiaoshuai Song Wen Feng Duan Li Jiaofen Nan Fubao Zhu Changxiang Yuan

The unilateral limb motor imagery paradigm can effectively address the cognitive dissociation problem among multiple limbs and provide strong technical support for extending the functionality of external devices. However, feature mining and accurate decoding of unilateral limb movements remain challenging. In this study, we propose a feature mining method that combines automatic frequency band boundary localization with regularized common spatial pattern (AFBBL-RCSP), and employ a pinball-loss-based twin support vector machine (Pin-UTSVM) to decode EEG signals corresponding to reaching, turning, and grasping movements. First, multiple optimal frequency band boundaries were identified for each subject using AFBBL. Then, regularized spatial features were extracted from each sub-band, and all features were reduced using Fisher’s discriminant analysis. Finally, the Pin-UTSVM classifier was used to categorize the three types of movement data. The results show that, compared with CSP and RCSP feature mining methods using the fixed 8–30 Hz band, the proposed method improves decoding accuracy by 9.52% and 3.89%, respectively. Compared with fixed single-band feature mining methods based on the α band, β band, and α + β band, the proposed method improves accuracy by 5.56%, 3.89%, and 3.73%, respectively. In addition, compared with existing unilateral limb decoding methods based on temporal-spatial features, temporal-frequency features, and temporal-spatial-temporal-frequency fusion CNN features, the proposed method improves decoding accuracy by 34.93%, 34.09%, and 28.11%, respectively. These results suggest that the proposed AFBBL-RCSP method is effective for unilateral limb motor imagery EEG decoding.

Information doi: 10.3390/info17050481

Authors: Sara Behnamian Zeinab Shahbazi Zahra Shahbazi Sadiqa Jafari

The spread of unverified health claims about drugs, dietary supplements, and alternative remedies on social media poses a growing public health concern. In this study, we present a retrieval-augmented generation (RAG) pipeline that uses large language models (LLMs) grounded in biomedical evidence from PubMed, openFDA adverse event reports, and NIH/NCCIH dietary supplement fact sheets to detect and classify health product misinformation. A total of 3493 health-related posts were collected from Reddit (948 posts across 12 subreddits) and YouTube (2545 video descriptions and comments), from which 8250 structured claims were extracted using Claude Haiku. Each claim was matched to biomedical evidence from three authoritative sources, achieving 79.4% evidence coverage, and classified into one of five veracity categories: supported (7.0%), unsupported (59.9%), exaggerated (22.4%), contradicted (2.0%), or dangerous (8.6%), together with an associated risk tier. Overall, 13.5% of claims were assigned high or critical risk. Cross-platform analysis showed that YouTube contained higher proportions of dangerous (11.3% vs. 2.9%) and exaggerated (27.0% vs. 12.4%) claims than Reddit. Compared with keyword-based and zero-shot transformer baselines, the LLM+RAG pipeline produced a more balanced and fine-grained classification of unsupported, exaggerated, contradicted, and dangerous claims. The most frequently implicated products were ashwagandha, kratom, black seed oil, turmeric, and ivermectin, with disease cure claims showing the highest dangerous classification rate (30.1%). These model-assigned results suggest that evidence-grounded LLM pipelines can support health misinformation surveillance, while also highlighting the need for expert validation and broader cross-platform evaluation.

Information doi: 10.3390/info17050480

Authors: Hao Tao Tao Chen

The spread of multimodal misinformation demands verification that is both accurate and fast while keeping knowledge current. Large language models are powerful but costly and slow, and their static knowledge can lag behind events. We introduce EC-MFR, a hierarchical framework that divides work between edge and the cloud. The system first optionally decomposes the claim into a few targeted sub-claims to guide retrieval, retrieves text and image evidence, and then compresses it into a small set of question–answer items using a lightweight, quantized multimodal language model deployed at the edge. A compact verifier on the edge predicts a label with calibrated confidence. If confidence is high, the decision is returned immediately. If confidence is low, the claim is sent to the cloud where retrieval can be expanded and the reasoning can be redone by a stronger verifier. This design offers three core benefits. It makes reasoning explicit through question–answer items, which shortens prompts and improves auditability. It improves retrieval recall via a light decomposition step that produces targeted sub-queries. Finally, it lets most easy claims finish on the edge to reduce cost and latency while preserving accuracy on difficult claims by allowing the cloud to broaden evidence and refine reasoning. Experiments on MOCHEG and AVERITEC validate the approach. Notably, EC-MFR achieves highly competitive accuracy of 54.10% on the multimodal MOCHEG dataset, and reaches 68.80% on AVERITEC under realistic retrieval settings, outperforming the GPT-4o cloud-only baseline by 6.6 percentage points. Furthermore, system-level profiling on edge hardware demonstrates that EC-MFR reduces processing costs by 51.8% and accelerates inference latency by 2.4× for edge-resolved claims, confirming a highly favorable accuracy–efficiency trade-off compared to existing multimodal fact-checking systems. We also formalize routing and efficiency and analyze calibration and retrieval.

Information doi: 10.3390/info17050479

Authors: Huafeng Qu Shafrida Sahrani Fariza Fauzi Xiacheng Song Yanfeng Zhao

Employment recommendation services are increasingly used to support graduate job search. However, limited research has examined how graduating computer science students perceive a proposed employment recommendation approach that combines static profile-based matching with dynamic interactive functions. Drawing primarily on the Technology Acceptance Model (TAM), with selected dimensions of the Information System (IS) Success Model used as supplement, this study conducted an exploratory questionnaire-based survey of 386 graduating students. The respondents evaluated existing employment recommendation systems and provided open-ended comments, and the findings show that only 38.3% of respondents reported willingness to use existing employment recommendation systems for job hunting. The main reported problems were delayed matching to individual qualifications (71.0%), information lag (55.4%), and jobs not matching students’ majors (54.1%). In contrast, respondents expressed relatively favorable attitudes toward the proposed static-dynamic approach: 67.6% indicated willingness to use it and 59.6% indicated willingness to recommend it to others. Exploratory subgroup analyses further suggested that positive evaluations of the proposed approach were higher among students from emerging computing fields and those with more active job-seeking engagement (p < 0.05). Overall, the findings provide exploratory evidence that graduating computer science students may respond more positively to employment recommendation concepts that integrate profile-based matching with dynamic interaction. However, it is a proposed design concept, not an implemented system, evaluated by the respondents. Therefore, the results should be interpreted as perceptions and stated intentions, instead of evidence of actual adoption or real-world system effectiveness.

Information doi: 10.3390/info17050478

Authors: Olga Ussatova Vladislav Karyukin Yenlik Begimbayeva Galimkair Mutanov Yerlan Kistaubayev Medet Turdaliyev

This article discusses the practical implementation of a prototype academic transcript storage system based on blockchain technology and smart contracts. The digital transformation of higher education requires reliable mechanisms for ensuring the integrity and verifiability of academic documents. It presents the design and experimental validation of a blockchain-based system for storing and verifying academic transcripts within the higher education system of the Republic of Kazakhstan. The proposed solution is based on an Ethereum Virtual Machine-compatible smart contract implemented in Solidity and deployed on a test network. The testnet was used as the experimental environment, and transaction monitoring was performed using the BlockScout v11.0.3 explorer. The architecture of the TranscriptStorage smart contract is presented, including a role-based access model, a data indexing mechanism using keccak-256, and storage of transcripts in a mapping structure (bytes32 => Transcript[ ]). The experimental results confirm the successful recording of the Transcript in the distributed ledger, event recording (Logs), and the correctness of the ABI encoding of input parameters (Raw Input), as well as a change in state (State Changes) reflecting the fee payment. The use of events is shown to enable cost-effective third-party data verification without the need to store the entire text in the contract state. The comparative results showed that the proposed system reduced gas consumption by 804.5% compared to Blockcerts, 48.8% compared to ECertChain, 82.5% compared to ShikkhaChain, and 43.5% compared to zkEVM. These improvements were achieved while maintaining high scalability, robust privacy features, and security, making it a practical solution for Kazakhstan’s educational system.

Information doi: 10.3390/info17050477

Authors: Alexandros Miteloudis Ioannis Hatzilygeroudis

Decision tree ensembles, such as Random Forests and Gradient Boosting Machines, achieve high predictive accuracy but often suffer from limited transparency due to their structural complexity. Due to this lack, interpretability challenges arise in domains where model understanding, accountability, and trust are essential. So, many interpretability/explainability techniques have been proposed for tree-based ensembles. However, although there are enough surveys or overviews concerning interpretability/explainability in artificial intelligence or machine learning in general, there are very few surveys and overviews on interpretability/explainability for tree-based ensembles. This paper provides an overview of recent approaches to interpretability and explainability in decision tree ensembles. We present two categorizations: one based on the kind of technique/architecture used and the second based on the level of scope. The former is a unified taxonomy of acquired (or post hoc) and inherent methods further analyzed in two more levels. The latter concerns the distinction between local (or instance-related) and global (or model-related) methods. We additionally provide a survey of the interpretability/explainability methods/techniques used in various domain applications, like healthcare, finance, law, and privacy preservation. This overview clarifies the current landscape of interpretable/explainable ensemble learning, explicitly addressing emerging challenges. Ultimately, it aims to support researchers and practitioners in selecting and developing ensemble models that move beyond the traditional accuracy–interpretability trade-off, aligning predictive power with strict regulatory, operational, and domain-specific transparency requirements.

Information doi: 10.3390/info17050476

Authors: Firdaous Khemlichi Youness Idrissi Khamlichi Safae Elhaj Ben Ali

Financial markets are characterized by non-stationarity, regime shifts, and complex cross-asset interactions, which challenge traditional portfolio optimization and motivate reinforcement learning (RL) for adaptive decision-making. However, many RL-based approaches remain predominantly return-centric, with risk, uncertainty, and human oversight only weakly integrated, limiting robustness and practical applicability. This review provides a critical synthesis of risk-aware and uncertainty-sensitive reinforcement learning for portfolio optimization from a human–AI collaboration perspective. We analyze major architectural paradigms—including single-agent, hierarchical, multi-agent, and modular systems—together with risk modeling strategies (e.g., reward shaping, constraint-based optimization, and downside risk measures such as CVaR) and probabilistic approaches to uncertainty estimation (e.g., Bayesian neural networks, Monte Carlo dropout, and ensembles). A structured analysis of 57 fully assessed studies reveals that only 5 (9%) explicitly couple uncertainty estimation with risk constraint mechanisms, while 38 (69%) treat risk and uncertainty as structurally independent components. We identify a central structural limitation: risk objectives are rarely conditioned on epistemic uncertainty, while uncertainty estimates seldom influence constraint mechanisms or capital allocation. This decoupling leads to fragmented frameworks that remain difficult to deploy in real financial environments. By integrating architectural design, risk modeling, uncertainty estimation, and evaluation practices, this review proposes a unified, deployment-oriented perspective for developing governance-aligned portfolio decision-support systems.

Information doi: 10.3390/info17050475

Authors: Alessandro Naim Sara Naim Daniele Saverino

Background: The expanding interest in chatbots within the medical domain underscores the imperative for a comprehensive understanding of their capabilities and limitations, particularly in the context of anatomical education. Chatbots possess the potential to comprehend intricate anatomical concepts, deliver both advanced and contextually relevant information, and serve as a valuable resource for medical students and educators. This study aimed to evaluate the proficiency and constraints of chatbots in the domain of neuroanatomy. Methods: We developed 30 questions and administered them to ChatGPT-4, Google Gemini, Microsoft Copilot, and Perplexity.ai in their open versions. Questions were collaboratively constructed by the research team, selected through a semi-randomized process within the domain of neuroanatomy. Chatbots’ responses were evaluated in a blinded manner for validity and appropriateness, utilizing a 5-point Likert scale. Results: The highest observed performance among the evaluated chatbots was exhibited by ChatGPT-4 and Perplexity.ai, which achieved scores of 4.6 ± 0.5 and 4.5 ± 0.5, respectively. Microsoft Copilot (4.4 ± 0.5) and Google Gemini (4.1 ± 1.0) followed. The least successful performance was observed in the task of generating a neuroanatomical structure: only Microsoft Copilot attempted to fulfil the request, albeit with a dramatically flawed outcome. Conversely, Google Gemini and Perplexity.ai provided web links to anatomical illustrations. Conclusions: Despite technological advancements, AI models have not yet reached a level of sophistication sufficient to entirely supplant the role of educators or facilitators in a neuroanatomy course; however, they can serve as valuable adjunct tools for medical educators and students when utilized with careful consideration.

Information doi: 10.3390/info17050474

Authors: Xu Chen Yuntian Bai Yifeng Hong

Unsupervised medical image translation remains challenging because model development often relies on unpaired training, whereas reliable evaluation requires well-matched reference images. PD-weighted and T2-weighted brain MR images provide a useful testbed for this problem because they are closely matched anatomically while still exhibiting distinct contrast characteristics. Existing methods often align only high-level features, overlooking low-level texture details that are important for structural fidelity. In this work, we propose the Multi-Scale Feature Regularization and Patch Mixup (MSFRPM) framework based on an encoder–decoder architecture. It aligns cross-domain features across multiple scales to preserve local details and employs a patch-based mixup strategy to augment training data. The framework was evaluated using an unsupervised learning protocol with strict data partitioning. Experimental results demonstrate that MSFRPM achieves strong performance relative to eight state-of-the-art methods. Our approach achieved improvements in MAE (6.26 ± 0.86), PSNR (23.53 ± 0.92), SSIM (0.83 ± 0.03), and GMSD (0.100 ± 0.010). Qualitative assessments confirmed improved structural fidelity, and t-SNE visualization validated enhanced cross-domain feature alignment. Overall, MSFRPM provides a useful approach for unsupervised PD-to-T2 image translation under the current experimental setting.

Information doi: 10.3390/info17050473

Authors: Jun Li Shenghan Li Huachang Zeng Shengda Zhuo

User behaviors and social interactions on online platforms are intricately intertwined, naturally forming complex graph structures. Leveraging this structure, Graph Neural Networks (GNNs) efficiently aggregate neighborhood information and have become a prevailing paradigm for social recommendation. However, existing methods often overemphasize social modeling while overlooking the joint effects of preference-guided relation filtering and user/item biases, rendering them vulnerable to noise from redundant ties. To address these limitations, we propose PDDSR, a Preference-Guided Debiasing and Denoising Social Recommendation framework. Specifically, for debiasing, PDDSR explicitly models user rating bias and item popularity bias as learnable vectors, integrating them into embedding learning to mitigate bias drift at the embedding level. Simultaneously, for denoising, the model employs a social relation confidence mechanism guided by user preferences and adopts an adaptive graph denoising strategy to retain highly informative connections, effectively capturing social influence while filtering out noise. Extensive experiments on the Ciao and Epinions datasets demonstrate that PDDSR consistently outperforms state-of-the-art methods, and notably on the Ciao dataset, the MAE and RMSE are improved by 1.90% and 1.87%, respectively. These results validate the effectiveness and robustness of the joint debiasing and denoising mechanism in complex social recommendation scenarios.

Information doi: 10.3390/info17050472

Authors: Fotios Pastrakis Markos Konstantakis George Caridakis

Cultural heritage experiences present unique challenges for user experience (UX) evaluation due to their diversity, contextual variability, and the growing need to balance methodological rigor with low cognitive effort. Traditional UX frameworks often assume a one-size-fits-all approach, which fails to address the complexity of cultural heritage environments. This paper introduces a flexible taxonomy of UX evaluation methodologies designed as a decision-support tool for researchers and practitioners. The taxonomy is built on 11 core dimensions: study type, research phase, research objective, evaluation timing, data nature, facilitation setup, observation setup, research environment, participant profile, cognitive burden, and evaluation standards and instruments. Rather than prescribing a single method, the taxonomy enables the selection and combination of qualitative and quantitative approaches tailored to the context and phase of each cultural heritage project. Representative examples illustrate its application in guiding mixed-methods strategies for measuring cultural resonance. By promoting adaptability and methodological diversity, this work advances human-centered UX evaluation practices for cultural heritage and beyond.

Information doi: 10.3390/info17050471

Authors: Adnan Ahmed Abi Sen Adel Ben Mnaouer Omar Tayan Abdullah M. Basahel Nour Mahmoud Bahbouh Sanaa Askool

Protecting the privacy of users’ data while maintaining reliability and accuracy in crowded events remains an open issue, especially with the growing capabilities and resources of attackers. This challenge becomes more difficult in dynamic environments with moving users/devices. Unfortunately, the current privacy-preserving methods suffer from several drawbacks that include reliability and accuracy of results, the need to fully trust a third party, or the incurrence of heavy overheads. This research presents a novel approach that is enhanced by peer cooperation, which is one of the most suitable techniques for crowded environments. The proposed approach is called “Smart Monikers with Multi-Peer Cooperation (SM2Peer)”. The SM2Peer addresses all the drawbacks of the traditional peer cooperation approach through two scenarios. In addition, the SM2Peer exploits the fog computing layer to control the cooperation among peers effectively, where each fog node manages several peers with smart moniker management. Moreover, SM2Peer provides multiple caches to relax the total overhead. The simulation and comparison with other common privacy approaches show the superiority of the SM2Peer in many aspects and metrics of privacy without a significant effect on performance.

Information doi: 10.3390/info17050470

Authors: Yi Chen Yang Wang Hao Xu Anning Wang

This study addresses the need to transform enterprise scientific and technological intelligence (STI) services from a discrete resource-supply model toward a more systematic value-creation approach, an important challenge in the digital transformation of knowledge-intensive industries. As an exploratory qualitative inquiry, this work combines large language model-assisted analysis with grounded theory to examine the construction logic and operational mechanisms of an embedded intelligent STI service system. Drawing on in-depth interviews with STI professionals, a qualitative corpus was analyzed using human–machine collaborative coding to systematically derive and organize key constructs. The findings yield a preliminary three-layer conceptual framework: “supply-demand interactive matching, organizational embedded services, and digital-intelligent platform support.” Specifically, the supply–demand matching layer facilitates targeted alignment through demand insight, dynamic response, and quality closed-loop management; the organizational embedded service layer delivers intelligence through scenario integration, process integration, and responsibility–authority integration; and the digital-intelligent platform support layer enables core capabilities via data element induction, intelligent diffusion, and tacit knowledge conversion. The proposed framework offers an initial, structured perspective on how embedded intelligent STI services may operate, providing a foundational reference for both research and practice in this emerging domain.

Information doi: 10.3390/info17050469

Authors: Jose-Amelio Medina-Merodio Mikel Ferrer-Oliva José Fernández López Alejandro Ruiz-Zambrano Adrián Domínguez-Díaz

Credit card fraud constitutes a core component of the contemporary cybercrime economy, in which dark web carding forums play a pivotal role in coordinating, commoditising, and disseminating illicit activities. While prior research has primarily focused on transaction-level fraud detection, comparatively limited attention has been devoted to the systematic analysis of the social and organisational ecosystems within which these practices are enacted. This study addresses this gap by proposing and validating a domain-specific taxonomy for the automated classification of content in P2P carding forums. To this end, we adopt an iterative, data-driven methodology that integrates large language models (LLMs), lexical co-occurrence analysis, and semantic network analysis. Using a corpus of 3260 posts, we define and operationalise a taxonomy structured around four predicates: activity context, actor role, products and services, and technical tools, supported by a locally deployed LLM (Llama 4 Scout). A human-annotated subset was additionally used to evaluate inter-annotator agreement and standard classification metrics, complementing the coverage-based assessment and enabling comparison against a keyword-based baseline. Evaluation was further strengthened through manual benchmarking, confidence intervals, sensitivity analysis of key pipeline components, and comparison with alternative open-weight models. The results indicate that the proposed taxonomy achieves broad corpus-level representational coverage, with at least one semantic dimension identified in 98.71% of posts. However, coverage is uneven across predicates: activity-context is highly explicit, whereas actor-role and product-service show only moderate coverage and technique-tool remains substantially underrepresented and ambiguous. Overall, the findings show that combining domain-specific taxonomies with LLM-assisted classification and network analysis offers a robust framework for understanding and monitoring carding ecosystems in the dark web.

Information doi: 10.3390/info17050468

Authors: Sumukhi Nandam Carlos D. Paternina-Arboleda

Small and medium-sized enterprises (SMEs) frequently depend on spreadsheet-based financial reporting due to limited budgets and constrained access to enterprise analytics systems. As transaction volumes increase, manual profit and loss computation becomes time-intensive and prone to inconsistencies. This study proposes and evaluates a modular robotic process automation (RPA) framework designed to enhance spreadsheet-centric financial analytics without requiring enterprise system replacement. The framework is implemented as a unified pipeline using UiPath. Statistical anomaly detection mechanisms are integrated to identify abnormal revenue deviations and expense spikes in operational data. Experimental benchmarking compares manual spreadsheet processing with automated workflow execution using execution time, error exposure, reporting latency, and scalability as evaluation criteria. Empirical evaluation across five datasets spanning 300 to 3000 transactions demonstrates time reductions of 88.6% to 95.5% and error reductions of 93.3% to 95.5% relative to manual spreadsheet processing. Scalability analysis confirms linear growth of automated runtime with transaction volume, in contrast to the superlinear growth observed in manual processing. A cost feasibility analysis further indicates that lightweight RPA can significantly reduce operational costs in SME environments up to 88.6%. The study contributes a structured automation architecture that integrates spreadsheet automation with statistical monitoring to support financial oversight and decision support. The findings suggest that interface-level automation provides a viable transitional pathway for SMEs seeking incremental digital transformation while preserving existing spreadsheet infrastructures.

Information doi: 10.3390/info17050467

Authors: Basma Elsharkawy Amira Abdelatey O. G. El Barbary Hatem Abdelkader Nesma Mahmoud

Drug–drug interaction (DDI) prediction plays a critical role in optimizing therapeutic outcomes and enhancing patient safety. DDIs pose challenges in drug discovery, often leading to adverse effects, reduced efficacy, or unexpected outcomes. AI in DDIs acts as an effective tool for analyzing and predicting DDIs which introduced efficient computational approaches to DDI prediction. This paper aims to provide a comprehensive understanding of how ML and DL models perform in DDI prediction. This paper presents a comparative analysis based on key performance metrics such as accuracy, precision, recall and F-score for different ML and DL Models. We used Synthetic Minority Oversampling Technique (SMOTE) and the Gray Wolf Optimizer (GWO) which achieved the best accuracy of 95.42%. Combining the GWO with SMOTE addresses both optimization and data imbalance challenges in DDI prediction. Effectively, SMOTE addresses the class imbalance issue that leads to poor performance. SMOTE improves model performance by generating synthetic examples of the minority class rather than merely duplicating existing ones. This helps create a balanced dataset, enabling the model to learn the decision boundaries more accurately. SMOTE reduces the risk of overfitting. The GWO serves as a metaheuristic optimization framework that enhances model performance by guiding optimal feature selection subsets. This optimization process improves the model’s ability to capture complex, non-linear interaction patterns, leading to enhanced results. In our result, we achieve an accuracy of over 94% which helps in drug safety and therapeutic decision-making in health informatics.

Information doi: 10.3390/info17050466

Authors: Jingjing Chen Xu Cheng Yuxin He Qi Zhang Xiaoling Liu Qin Luo Kwok-Leung Tsui

Transit delays can adversely affect passengers, operational efficiency, and daily lives. It is important to develop effective methods to identify and analyze train stations vulnerable to delays. This paper proposes a three-stage analytical framework for analyzing train station delays. In the first stage, the 3-sigma rule defines normal passenger volume ranges and establishes a time window affected by delays. Next, a multivariate time series clustering method identifies stations with stable demand and high volume, considering passenger volume differences both among and within stations. In the final stage, the effects of delays on these key stations are assessed by examining starting, duration, and ending times, and passenger volume variation, providing a comprehensive analysis of delay impact. The proposed framework is illustrated using two real-world incidents: the 2021 delay incident at Longyang Road Station of Shanghai Metro and the 2019 delay incident on the Taoyuan–Luohu section of Shenzhen Metro. Case studies revealed that affected stations are not limited to the specific line or direction of the delay, but also include opposite-direction and transfer stations. Station impacts exhibit phased onset and recovery patterns. Additionally, both increases and decreases in passenger volumes due to the delay present considerable implications. While both incidents exhibit common propagation and recovery patterns, the Shanghai incident displays wider passenger impacts and longer recovery periods, whereas the Shenzhen incident exhibits narrower impacts and faster recovery. Our results will aid transit managers in better managing delays, thereby improving passenger satisfaction and operational efficiency. This paper also offers an integrated station-level analytical framework and initial cross-case empirical evidence, while broader validation remains needed.

Information doi: 10.3390/info17050465

Authors: Mona Almutairi Samia Dardouri

Panoramic radiographs are widely used in dental practice due to their ability to provide a comprehensive view of the teeth, jaws, and surrounding anatomical structures in a single examination. However, automated interpretation remains challenging because multiple conditions may co-exist within a single image, class distributions are highly imbalanced, and several findings exhibit subtle radiographic characteristics. This study presents a deep learning framework for multi-label dental findings classification using panoramic radiographs from the publicly available VZRAD2 dataset. Following a label curation process, eleven clinically relevant classes were retained, including diseases, treatments, and anatomical structures. The proposed EfficientNet-B4-CBAM model integrates an EfficientNet-B4 backbone with a Convolutional Block Attention Module (CBAM) to enhance feature representation through channel and spatial attention. EfficientNet-B4 and ResNet50 were used as baseline models for comparison under a unified training protocol. The training pipeline incorporates data augmentation, weighted sampling to address class imbalance, AdamW optimization, and Binary Cross-Entropy with Logits loss for multi-label learning. On the validation set, the proposed model achieved the highest micro-F1 score of 0.8567, compared to 0.8424 for EfficientNet-B4 and 0.8469 for ResNet50. ROC analysis showed comparable separability across models, with micro-AUC values of 0.946 (EfficientNet-B4-CBAM), 0.947 (EfficientNet-B4), and 0.960 (ResNet50). Class-wise evaluation indicated strong performance for visually distinct findings such as impacted tooth, implant, filling, and root canal treatment, while anatomically diffuse or underrepresented classes remained more challenging. Grad-CAM visualizations suggest that the model focuses on clinically relevant regions, supporting interpretability. Overall, the results indicate that attention-enhanced convolutional models can provide effective and interpretable support for multi-label dental findings classification. However, the observed performance improvements are modest, and further validation on independent datasets, along with clinical evaluation, is required to confirm generalizability and real-world applicability.

Information doi: 10.3390/info17050464

Authors: Jorge Fernando Sosa Gallardo Vivian Felix López Batista María N. Moreno-García María Dolores Muñoz Vicente Aldo Fernand Sosa Gallardo

Reducing cement consumption in mortar systems is essential for lowering the environmental impact of cement-based materials. Conventional mix design approaches rely mainly on particle size distribution and fineness modulus, which do not fully capture the effects of aggregate packing, morphology, and petrographic composition on paste demand and mechanical performance. Fourteen fine aggregates of distinct geological origins were experimentally characterized in terms of physical and petrographic properties. A dataset of 211 mortar mixtures, yielding 633 transverse-strength observations, was used to train a Random Forest Regressor (RFR) model for strength prediction. The model achieved R2=0.762 (RMSE = 0.223 kN; MAE = 0.165 kN), demonstrating its reliability as a surrogate screening tool. This study presents a hybrid framework that integrates particle packing theory with machine learning to optimize fine aggregate blends. By introducing a Paste Demand Index (PDI)—combining normalized uncompacted void content, surface texture, and shape—the framework enables the identification of mixtures that minimize paste demand while maintaining mechanical performance under strength constraints. Results confirm that the proposed PDI and strength-based filtering are robust, offering a physically grounded decision-support methodology for narrowing the design space. Ultimately, this approach provides an efficient strategy for resource optimization, effectively bridging the gap between computational screening and laboratory validation in cement-reduction initiatives driven by the cement-based tile manufacturing industry.

Information doi: 10.3390/info17050463

Authors: Cyril S. Ku Daniel Weiner Meera Wells Andrew Huang Morgan R. Peltier

Objective: This study examines the potential of natural language processing and text mining to automate the systematic review process in clinical psychiatry, a field that traditionally relies on domain experts and can be time-consuming, prone to human bias and errors. The study compares the classification of review articles by domain experts with that facilitated by machine algorithms. Methods: Using data from PubMed, 160 abstracts related to “transcranial magnetic stimulation” and “autism” were classified into “treatment” and “non-treatment” categories by both human reviewers and a computer algorithm. The computer algorithm, employing topic modeling in text mining, was compared to human reviewers, including two psychiatrists, a biostatistician, and a medical student. Results: The accuracy of human classifications ranged from 68% to 85%, with inter-rater reliability (Kappa statistic) between 0.40 (fair to moderate) and 0.64 (substantial). Intra-rater reliability, tested by reclassification after three months, varied from 0.38 to 0.82. Conclusions: The findings highlight the consistency and reproducibility of computational approaches compared to human classification, which exhibited both inter-rater and intra-rater variability. Differences in reviewer performance were observed; however, these patterns should be interpreted cautiously, as the study was not designed to directly assess cognitive or decision-making processes.

Information doi: 10.3390/info17050462

Authors: Igor Milojevic Dragana Rejman Petrovic Marina Milanovic Bojan Lekovic Marko Slavkovic

This study proposes an integrative model that combines technological factors and top management support to examine their impact on successful ERP implementation and its direct effect on the efficiency, effectiveness, and flexibility of business processes in the context of firms operating in Serbia. Employing the Technology–Organization–Environment (TOE) framework, this research analyzes data collected from 123 managers using a structured questionnaire and Partial Least Squares Structural Equation Modeling (PLS-SEM). The findings reveal that technological complexity positively affects ERP implementation success, contrary to common assumptions that complexity is a barrier. Technological compatibility and readiness show no direct significant effects but demonstrate conditional influences when moderated by top management support. Top management involvement significantly moderates the relationship between technological factors and ERP success, with balanced managerial engagement being critical to avoid potential negative impacts of over-control in complex projects. Moreover, ERP implementation significantly enhances operational efficiency, effectiveness, and flexibility. This study concludes that ERP success depends on the interplay between technological attributes and well-balanced leadership support, emphasizing the need for holistic management of technology, people, and processes. These insights contribute to theoretical understanding and practical guidance for organizations aiming to optimize ERP outcomes and operational performance.

Information doi: 10.3390/info17050459

Authors: En Lee Thian Song Ong Yvonne Lee

Household vulnerability assessment in Malaysia has traditionally relied on income-based indicators, which do not adequately capture multidimensional deprivation. To address this limitation, this study employs Random Tree–Dirichlet Process Mixture Model (RT-DPMM) to identify latent heterogeneity in spending patterns and their associated socioeconomic characteristics. Using microdata from Household Expenditure Survey (HES), this study performs clustering on 5130 stable household head samples with nine spending proportional features to model their joint distribution as mixtures of Dirichlet distributions, while five socioeconomic covariates inform cluster allocation through Random Tree embeddings. The proposed RT-DPMM identifies four distinct spending clusters: Balanced Budget Households (Cluster 1, N = 2883), Mobility and Home-Support Households (Cluster 2, N = 642), Basic Essentials-Focused Households (Cluster 3, N = 977), and Luxury Households (Cluster 4, N = 628). Cluster 1 and 3 are characterized as relatively vulnerable groups. These clusters have lower income levels and allocate a larger budget share in Food and Beverages, consistent with the Engel Law’s interpretation of higher food percentage in lower income households. Cluster 1 households primarily allocate their budget evenly across essential and non-essential spending. Cluster 3 are mostly elderly household heads with the highest budget shares in essential spending. In contrast, Cluster 2 and 4 appear relatively better off financially, given their higher income and larger spending share to non-essential categories. These findings suggest that social assistance policies should target expenditure patterns, rather than relying solely on income-based targeting.

Information doi: 10.3390/info17050461

Authors: Zhiang Chen Miao Song Ningge Wu

Early mild cognitive impairment (EMCI) is the earliest intervenable stage of Alzheimer’s disease (AD). Although graph neural networks (GNNs) have begun to exploit brain network topology, traditional fMRI-based diagnostic methods often neglect these structural patterns by relying on vectorized features. Furthermore, existing GNNs frequently disregard inter-regional functional heterogeneity and group-level discriminative patterns, leading to limited accuracy and biomarker interpretability. To address these challenges, we propose HF-BrainGNN, an end-to-end hierarchical graph learning framework for EMCI identification. Our method introduces a functional affinity region convolution (FAR-Conv) layer to learn region-adaptive kernels, a Differential Focus Pooling (DF-Pool) module to identify disease-salient brain regions by maximizing inter-group distinctiveness, and a hierarchical integration classifier (HIC) to fuse multi-level graph representations. The framework is optimized using classification, focus separation, and consistency regularization losses. Experiments on the ADNI dataset (104 EMCI, 114 Cognitively Normal) show that HF-BrainGNN achieves 86.78% accuracy, outperforming the best baseline (Hi-GCN) by 4.64%. Furthermore, the automatically identified regions, such as the bilateral hippocampus and default mode network hubs, align with established EMCI biomarkers. Ultimately, HF-BrainGNN provides an efficient, interpretable artificial intelligence tool for precise brain network characterization and early AD intervention.

Information doi: 10.3390/info17050460

Authors: Bo Zhao Weihua Ou

Parameter-efficient fine-tuning (PEFT) enables large language models to adapt to downstream tasks with low computational cost. As a representative high-rank PEFT method, MoRA (High-Rank Updating for Parameter-Efficient Fine-Tuning) improves update expressiveness through a compression–transformation–decompression reparameterization mechanism. However, its bottleneck subspace is still modeled using a freely learned linear transformation. In addition, grouped compression may project information from different original directions into shared bottleneck coordinates. This may reduce subspace separability and lead to inefficient utilization of the effective update space. To address this limitation, we propose GL-log-MoRA, which introduces a learnable general linear transformation into the MoRA bottleneck subspace and applies log-determinant regularization to encourage a more balanced spectral structure. In this way, the proposed method improves directional coordination and subspace expressiveness without imposing hard structural constraints or causing noticeable memory overhead. We evaluate GL-log-MoRA on five benchmarks: LogiQA, Financial PhraseBank, GSM8K, FinQA, and HotpotQA. The results show that GL-log-MoRA achieves the best performance on these downstream tasks and yields small but consistent improvements over MoRA under the same parameter budget. Compared with MoRA, GL-log-MoRA improves LogiQA from 42.50% to 45.45% and Financial PhraseBank from 81.60% to 83.02%. It also improves GSM8K from 63.1% to 64.6%, FinQA from 10.02% to 10.23%, and HotpotQA from 70.6% to 70.8%. Meanwhile, the average empirical effective-rank indicator increases from 1.05 to 2.80. Peak GPU memory changes only slightly, from 18.21 GB to 18.28 GB.

Information doi: 10.3390/info17050458

Authors: Muhammad Azhar Adeen Amjad Muhammad Arman Deshinta Arrova Dewi

This paper addresses emotion recognition in low-resource language settings for healthcare and human-computer interaction (HCI). Most existing multimodal systems rely on resource-intensive transformers or high-resource languages, limiting their applicability to low-resource languages like Urdu. We propose an efficiency-driven, lightweight multimodal framework for Urdu emotion detection integrating facial expressions, speech, and text. We utilize DistilBERT for text, CNN-BiGRU for audio, and MobileViT-XXS for visual processing with a dual-level fusion strategy. We evaluate on the publicly available UMED corpus, the only multimodal Urdu emotion dataset. Our system recognizes expressed emotional signals rather than internal affective states. Experimental results demonstrate competitive performance (83.72% accuracy) while requiring 76.5% fewer parameters and 4.4× faster inference than heavyweight baselines, enabling accessible, real-time emotion recognition in low-resource contexts.

Information doi: 10.3390/info17050457

Authors: Mobolaji Orisatoki Weihua Sheng Ebubekir Pinar Ali Rasoulzadeh Mahdi Amouzadi Arash M. Dizqah

Rapid exploration of unknown environments is critical in engineering applications such as disaster response and autonomous inspection. This paper presents an informative path planning approach for autonomous mapping of fully unknown, non-convex environments using a mobile robot with an uncertain narrow-beam range sensor. The artificial intelligence contribution lies in approximating the global optimal exploration solution under uncertainty using a sequential decision-making algorithm. The engineering contribution is the formulation and introduction of a benchmark solution, and the validation of the proposed algorithm against this benchmark through simulation and real-world experiments. Results show that the method achieves approximately 70% of the benchmark efficiency, measured as map expansion per unit distance travelled, with near-linear map growth. Sensitivity analysis demonstrates robust performance under varying initial conditions, confirming its applicability for real-world autonomous robotic systems.

Information doi: 10.3390/info17050456

Authors: Cesar Felipe Henao Villa David Alberto Garcia Arango Luis Fernando Garcés Giraldo Rosana Alejandra Meleán Romero Alejandro Valencia-Arias José Alexander Velásquez Ochoa

This systematic review (PRISMA 2020) examines 89 studies—64 peer-reviewed articles and 25 arXiv preprints (2007–2026)—addressing the gap between AI research and operational predictive maintenance (PdM) deployment in complex manufacturing systems. Analyzing five thematic clusters in non-stationary and stochastic environments, we evaluated predictive performance and deployment readiness. Deep learning dominates remaining useful life (RUL) forecasting; however, 65.6% of studies employ weak or unclear validation protocols (Tier 0–1), lacking real-world robustness testing. Fault diagnosis increasingly integrates Edge-AI, yet Explainable AI (XAI) adoption remains scarce (15.6%), undermining industrial trustworthiness. No study reached operational field validation beyond temporal or cross-domain split, reflecting a systematic disconnection from deployed manufacturing systems. We introduce a novel Deployment Readiness Score (DRS) framework and identify critical barriers: data scarcity, environmental non-stationarity, computational constraints, and black-box model distrust. Recommendations include standardized temporal validation protocols, multi-site field studies, and architecture-integrated explainability. The 25 arXiv preprints (2024–2026) exhibit a mean DRS nearly three times that of the peer-reviewed corpus, signaling nascent convergence toward deployment-mature research. This review was not pre-registered.

Information doi: 10.3390/info17050455

Authors: Tianci Sun Wanqiu He Changbin Shao Shang Zheng Hualong Yu

In real-world scenarios, large-scale datasets often exhibit a long-tailed data distribution. Training deep neural networks on such data typically leads to a bias towards head classes. Existing studies have demonstrated that the reweighting strategy is an effective means to alleviate the long-tailed issue. Recent studies suggest that incorporating class difficulty into reweighting can yield superior results. However, the method of quantifying class difficulty by an independent validation set has shown limitations in practical applications, i.e., wasting training samples and inaccurate estimations. To address this issue, this study proposes a novel model based on K-fold cross-validation, called the adaptive combination validation model, which contains two main innovations: first, both class and sample difficulty are quantified by using a more comprehensive and authentic estimation strategy, i.e., K-fold cross-validation, to obtain accurate and robust estimations; second, we extract the prediction probability distributions of samples, which reflect sample difficulty, from different model branches and design a distribution-harmonized loss to simultaneously focus on the effects of reweighted and original distributions. Extensive experiments on several popular long-tailed image recognition datasets (CIFAR10-LT and CIFAR100-LT, with several varying imbalance rates, and ImageNet-LT) demonstrate that the proposed method can effectively alleviate the long-tailed issue and achieve state-of-the-art performance on most datasets.

Information doi: 10.3390/info17050454

Authors: Shaozhen Hong Yingqi Liu

The rapid digitalization of the media and publishing industry has deepened systemic asymmetries in resources, power, and institutional rights. These asymmetries create fundamental barriers to the economic–institutional sustainability of digital content dissemination. Existing governance frameworks have not yet comprehensively addressed the resulting competitive and informational imbalances. Adopting China’s publishing and media industry as a focal case, this study draws on symmetry theory to develop an integrated analytical framework. It reconceptualizes government regulation as a multi-dimensional governance mechanism operating across three dimensions: resource allocation, technological innovation, and rights protection. We test this framework empirically using Xinbang Index data covering the top 10 publishing and media enterprises from 24 January 2025 to 7 December 2025. Multiple regression analysis and Spearman rank correlation are applied to assess each dimension’s differential impact on content dissemination efficiency. The results yield four key findings. First, all three regulatory dimensions contribute positively to content dissemination efficiency. Second, technological innovation is the most potent symmetry-restoring lever, exerting a statistically robust direct effect on dissemination outcomes. Third, resource allocation provides a necessary foundational contribution, while rights protection operates conditionally—its effect is fully realized only alongside adequate technological and resource inputs. Fourth, an integrated multivariate regression confirms the cross-dimensional hierarchy: the standardized Beta coefficient for technological innovation (β = 0.394) exceeds those for rights protection (β = 0.294) and resource allocation (β = 0.125). No single regulatory instrument is sufficient to achieve dynamic equilibrium. A synergistic, technology-centered combination of all three dimensions is required. This study proposes a tripartite symmetry-based governance strategy for media platform ecosystems. The symmetry framework developed here offers an analytical template for diagnosing analogous asymmetries in other platform-dependent sectors. Empirical validation beyond the Chinese publishing and media context is recommended as a priority for future research.

Information doi: 10.3390/info17050453

Authors: Babu George Divya Choudhary

Conventional web analytics treats the human user as its fundamental unit of analysis, assuming stable preferences, identifiable intentions, and behavioral patterns that unfold over time. That assumption is under strain. Crawlers and traditional bots already account for a substantial fraction of online interactions, and autonomous AI agents are emerging as a further class of actors layered on top of this automated traffic. Unlike either, these agents do not possess persistent identities or psychologically grounded motivations. They are task-specific, dynamically instantiated processes whose behaviors are contingent and often orchestrated by external systems. Their presence weakens the interpretive value of core metrics, including sessions, engagement, conversion, and retention. A click may reflect an optimization routine, a proxy objective, or a recursive agent-to-agent exchange rather than meaningful human intent, and traditional inference frameworks cannot reliably distinguish among these possibilities. This is a position paper. It synthesizes literature across bot and agent detection, agent architecture, web measurement validity, governance of automated systems in adjacent sectors, and the epistemology of digital trace data, and it argues that web analytics should supplement, and in places replace, its human-centered model with an agent-aware model focused on interaction dynamics within hybrid ecosystems of human and non-human actors. The paper develops a working taxonomy of crawlers, traditional bots, AI agents, LLM-powered agents, and autonomous agents; identifies three properties of LLM agents (identity discontinuity by design, task-based instantiation, agent-to-agent loops) that distinguish the present challenge from prior bot-detection problems; examines opaque agent objectives, synthetic traffic loops, and the indistinguishability between human-originated and agent-mediated signals; and proposes five candidate measurement primitives (task chain, actor class, interaction provenance, objective alignment, signal authenticity) with explicit operational definitions. Governance machinery from energy systems and critical infrastructure offers a partial template, and we delimit which dimensions transfer and which do not. The contribution is conceptual and programmatic, presenting a vocabulary, set of candidate primitives, and research agenda for a field whose foundational unit of analysis is becoming unreliable.

Information doi: 10.3390/info17050452

Authors: Prasit Puttamapadungsak Sumaman Pankham Somchai Lekcharoen

Against the backdrop of high digital uncertainty in the 2023 Thai General Election, this study examines how social media reshapes voting intentions through a novel hybrid framework integrating Fuzzy Rough Set Theory (FRST), Hierarchical Confirmatory Factor Analysis (CFA), and Random Forest Regression (RFR). A three-stage design—combining 23 expert opinions with survey data from 812 voters—overcomes expert ambiguity and non-linear dynamics. The findings reveal a hierarchy in digital campaigning: while Party Image (Importance = 0.3056) is the primary predictor for initial voter attention, substantive Campaign Policy (β = 0.98) remains the definitive driver of final commitment. Other perceptual constructs, including Trust, Loyalty, and Perceived Quality, function as reinforcing dimensions that validate policy claims within the digital ecosystem. This suggests a shift where traditional broadcasting is superseded by interactive digital streaming, allowing voters to scrutinize policies through replays and public comments. The model’s robustness, validated through 10-fold Random Forest Cross-Validation, demonstrates high predictive stability (Mean CV R2 = 0.840) and minimal error (MAE = 0.064). This study offers a sensitive instrument for emerging democracies and provides actionable insights, showing that substantive policy remains the ultimate driver of voter choice, even when mediated through Party Image in interactive digital environments.

Information doi: 10.3390/info17050451

Authors: Pruthviraj Pawar Gregory Epiphaniou

Deploying deep neural networks-based intrusion detection systems on resource-constrained edge devices demands inference strategies that balance latency, energy, and accuracy under shifting threat landscapes. This paper presents Edge-Prioritize IDS, a class-prioritized early-exit framework that accelerates inference for high-risk attack classes without post-deployment retraining. A lightweight K-dimensional control vector encodes per-class runtime priorities and steers samples toward earlier exits via adaptive normalization and cost-sensitive training. Evaluation across five benchmarks NSL-KDD, CIC-IDS2017, UNSW-NB15, WISDM, and CIFAR-10 on an NVIDIA Jetson TX2 shows that Edge-Prioritize IDS preserves baseline accuracy (up to 99.6%) while reducing latency by up to 55% and energy by up to 50% for prioritized classes. Ablation studies isolate each component’s contribution, and a controlled distribution-shift experiment demonstrates the sliding-window heuristic’s ability to recover near-baseline latency within 500 samples under synthetic class-frequency drift. Once trained under the proposed framework, the model requires no additional retraining, firmware updates, or additional memory beyond the priority vector itself when runtime priorities change.

Information doi: 10.3390/info17050450

Authors: Alexander Bradley Samuels

Municipalities in South Africa face increasing pressure to improve service delivery, operational efficiency, and sustainability amid growing urbanisation and governance challenges. The integration of smart city dimensions such as smart governance, mobility, and infrastructure offers a transformative approach to improve public sector supply chain management. However, limited empirical research exists on how these dimensions are being applied in South African municipal contexts. This study aimed to evaluate the extent to which smart city dimensions are integrated into supply chain management practices within a South African municipality and to assess the impact of these initiatives on supply chain efficiency, transparency, and sustainability. A qualitative, exploratory case study design was employed. Twenty senior managers and key stakeholders from the supply chain department of the selected municipality were purposively sampled. Data were collected through semi-structured face-to-face interviews and analysed thematically using NVivo software. Lincoln and Guba’s trustworthiness framework guided the study’s rigour. The findings revealed partial and uneven integration of smart city dimensions, with notable developments in smart governance and mobility, but limited progress in areas such as infrastructure digitalisation and citizen-centric data platforms. Participants highlighted both innovation drivers and institutional barriers affecting the transition to smart-enabled supply chain practices. Smart city dimensions present significant potential to improve municipal supply chain management; however, effective integration requires structural alignment, digital investment, and organisational readiness. This study provides context-specific insights into the uneven and fragmented integration of smart city dimensions within municipal supply chain systems in a developing country context, emphasising the impact of institutional constraints, digital capability gaps, and governance misalignments on implementation outcomes.

Information doi: 10.3390/info17050449

Authors: Amir M. Behrouzian Marco Meleti Maria Teresa Colangelo Elena Calciolari Carlo Galli

Large language models (LLMs) are increasingly investigated for abstract screening in systematic reviews, yet it remains unclear whether screening errors attributed to linguistic complexity arise from intrinsic semantic sensitivity or from its interaction with decision architecture. We examined how five polarity variants of logically equivalent eligibility criteria—affirmative inclusion, antonymic exclusion, predicate negation, verb-level negation, and double negation—affect screening outcomes in a controlled biomedical task. Using 1000 abstracts from a reconstructed Cochrane review corpus (50 TARGET; 950 non-targets), we implemented four abstract-visible criteria within a sequential hard-gated pipeline, where failure at any step triggered irreversible exclusion. Under hard gating, linguistic polarity alone produced substantial and statistically significant variation in recall. For GPT-5.1, recall ranged from 0.72 to 0.32 despite identical logical predicates and input data. Replication with GPT-3.5 Turbo yielded a similar divergence (0.92–0.18), confirming generalization across model generations. TARGET losses were concentrated at criteria typically satisfied but inconsistently reported in abstracts, indicating conservative exclusion under evidential under-specification. To assess whether this effect is semantic or architectural, we reimplemented screening using a scoring-based evidence-accumulation framework in which each criterion contributed graded support and inclusion was determined by a tunable threshold. Under scoring, recall increased across all variants and converged within a high-sensitivity regime, while residual polarity effects were attenuated but remained detectable. Linguistic differences shifted from structural recall collapse to controlled precision–recall trade-offs. These findings show that negation sensitivity is strongly mediated by decision architecture. Irreversible gating amplifies local uncertainty into false-negative exclusion, whereas cumulative scoring preserves uncertainty and enables controllable operating thresholds.

Information doi: 10.3390/info17050448

Authors: Muhammad Azhar Muhammad Arman Adeen Amjad Deshinta Arrova Dewi Muhammad Usman Ahmad Shafiq Hussain

The public health concern of suicide continues to rise and is increasingly prevalent on social media. The severity of this growing issue highlights the need for improved methods for detecting suicide risk. Many current deep learning approaches do not possess the required level of explainability for application in clinical settings. This study proposes the development of a transformer-based framework called “CrisisFormer,” which was trained on an imbalanced dataset containing 40,000 Reddit posts from the Suicide Watch subreddit and enhanced using DistilBERT. Additionally, the CrisisFormer framework uses three forms of explainable artificial intelligence for interpreting results: SHapley Additive exPlanations (SHAP), Local Interpretable Model-Agnostic Explanations (LIME), and transformer attention visualizations. The CrisisFormer framework achieved superior results for detecting the risk of suicide, with 96.25% accuracy, 96.30% precision, 96.25% recall, 96.25% F1 score, and 0.9944 AUC, compared to traditional models such as CNN, LSTM, and BiLSTM. Furthermore, by including clinically relevant suicide terms in its results, CrisisFormer demonstrates a high potential for incorporation into real-world mental health systems for intervention during ongoing mental health crises.

Information doi: 10.3390/info17050447

Authors: Jesús F. Rodríguez-Aragón Carolina Zato Fernando De la Prieta

Encrypted storage systems increasingly rely on governance mechanisms such as delegation, revocation, key updates, and policy evolution. While existing approaches provide strong guarantees for access enforcement, integrity, and transparency, they do not address a fundamental question: under which conditions can an observed sequence of governance events be accepted as a semantically valid evolution of authorization state? This work introduces a formal semantic framework for governance validity based on observable evidence. Governance is modeled as an admissibility-constrained state transition system in which events are accepted only if they satisfy explicit authorization, reference, temporal, revocation, and evidence conditions. The framework defines valid governance histories as sequences of admissible events; characterizes the conditions for deterministic state reconstruction; and establishes invariants capturing correctness properties such as revocation soundness, policy-constrained evolution, evidence completeness, non-equivocation, and temporal coherence. It also defines event-specific evidence obligations that support independent verification. The proposed approach is architecture-independent and does not prescribe specific enforcement or logging mechanisms, focusing instead on the semantic conditions required for accepting governance histories as valid from observable evidence. In addition, the framework can be instantiated as an independent verification layer that operates over observable governance traces without requiring access to internal system states.

Information doi: 10.3390/info17050446

Authors: Mohamed Ilyas El Harrak Omar Bahou Karim El Moutaouakil Ahmed Nuino Eddakir Abdellatif Alina-Mihaela Patriciu

While Graph Autoencoders (GAEs) have become a standard for unsupervised representation learning, their reliance on integer-order convolutions inherently restricts information propagation to immediate local neighborhoods. This paper introduces the Fractional Graph Autoencoder (FGAE) and its variational extension (FVGAE) to move beyond these local constraints. By integrating fractional Laplace operators, our framework generalizes conventional GAEs and enables tunable non-local propagation. We show that the fractional order α acts as a structural regularizer, utilizing the Green’s function of anomalous diffusion to induce a form of structural memory within the latent space. This allows the model to recover long-range dependencies that are typically lost in standard architectures. Systematic benchmarking across eight datasets—ranging from homophilic citation networks to heterophilic and dense product graphs—shows that these fractional variants consistently outperform both foundational and state-of-the-art baselines (ARGA, SIG-VAE, and GraphMAE). Notably, on the Amazon Computers and Citeseer datasets, our methods achieve relative increases in Normalized Mutual Information (NMI) of 77.55% and 67.28%, respectively. Statistical analysis confirms these gains are robust, with large effect sizes (Cohen’s d>0.80) and significance at p<0.05. These findings suggest that fractional graph autoencoding offers a mathematically grounded inductive bias for capturing the complex, multi-scale dynamics of real-world networked systems.

Information doi: 10.3390/info17050445

Authors: Yahia Jazyah

In modern computing environments characterized by high variability and complex workloads, traditional load-balancing algorithms such as Round Robin and Least Connections are often found to be less effective in distributing tasks and maintaining optimal performance. In this paper, a hybrid load-balancing algorithm is proposed, where the strengths of Fastest Response Load Balancing (FRLB) and Priority-Based Load Balancing (PBLB) are combined. Through this adaptive approach, response times are minimized and load distribution across heterogeneous server environments is balanced more effectively. In the recent literature, the need for enhanced load-balancing solutions that can adapt to dynamic conditions in cloud computing, IoT, and large-scale web services has been increasingly emphasized. By integrating a hybrid mechanism, a robust solution is provided by the hybrid algorithm, which is designed to merge between FRLB and PBLB. As demonstrated through simulations, a noticeable improvement in performance is achieved, with a significant reduction in average response time when compared to FRLB and PBLB. The proposed algorithm outperforms the classical FRLB and PBLB.

Information doi: 10.3390/info17050444

Authors: Peter Kokol Helena Blažun Vošner Jernej Završnik Bojan Žlahtič

Background: The convergence of Evolutionary Game Theory (EGT) and Artificial Intelligence (AI) has established the field of Intelligent Evolutionary Games (IEGs). While IEG applications have flourished in general systems and social sciences, their operationalization within healthcare (IEG Health) remains significantly underdeveloped. This study identifies a “knowledge void” in the literature, where the bottleneck is not a lack of clinical data but a scarcity of frameworks that integrate intelligent strategic modelling into clinical practice. Methods: We employ the Synthetic Near-Empty Review (SNER) framework, utilizing Synthetic Knowledge Synthesis (SKS) and bibliometric triangulation via VOSviewer. Three distinct corpora—IEG Health, EG Health, and IEG All (IEG)—were harvested from Scopus and mapped to identify thematic clusters and translation pathways. Results: The analysis reveals that IEG Health is a nascent domain currently focused on service regulation in elderly care and chronic disease management. We demonstrate a “Translation Framework” to bridge the research void, mapping concepts like Social Trust and Reputation Management from the broader IEG literature into clinical-specific models, such as Doctor-AI Adoption and Adaptive Coordination Games. Conclusions: By shifting from static Replicator Dynamics to Adaptive Learning Strategies (e.g., MARL and Bayesian updating), IEG Health can address critical challenges like algorithm aversion and clinical deskilling. Furthermore, transitioning these models into clinical environments requires the incorporation of structured ethical guidelines, such as ALTAI, to ensure algorithmic accountability. This study provides a structured foundation for future research to transition from theoretical modelling to AI-augmented clinical decision-making.

Information doi: 10.3390/info17050443

Authors: Mark Colgate Orla Colgate

AI is now embedded in frontline service at scale, yet the design frameworks managers reach for were built around human agents and do not translate cleanly to systems that generate rather than retrieve, that automate rather than augment. This paper argues that three design challenges sit at the heart of the problem, though they are rarely treated as a connected set. Generative AI can produce fluent, confident outputs that are simply wrong, which is a qualitatively different kind of reliability failure from anything SERVQUAL was designed to address. AI can reply instantly while leaving the customer no closer to resolution, exposing a gap between speed and what we might call felt responsiveness. And it faces an awkward relational tension. Overclaiming warmth triggers distrust, yet there are genuine service contexts in which the non-human nature of the system is a feature rather than a liability. The RRR Design Framework developed here extends established service quality dimensions to the AI context, organising fifteen prescriptive design principles around reliability, responsiveness, and relational quality, each reconceptualised for AI-mediated service. The principles follow a prevent-and-recover logic within each dimension and are tied together by a single strategic proposition, which is to automate to protect relationships. Four empirically testable propositions are derived from the framework, each operationalised with measurable constructs, moderating conditions, and falsifiable null cases. The framework is most applicable to hybrid human-AI frontline systems where customers are actively working toward a resolution.

Information doi: 10.3390/info17050442

Authors: Aziz Boujeddaine Hamid Khalifi Youssef Ghanou Sara Riahi Walid Cherif

Early detection of suicidal ideation and depressive risk remains a critical challenge, particularly when individuals express distress implicitly through metaphorical or obfuscated language. Existing approaches primarily rely on explicit linguistic signals, limiting their effectiveness in real-world settings. This paper proposes a unified multi-aspect transformer-based framework that integrates multi-source learning, multi-task optimization, affective feature fusion, and adversarial training to detect implicit psychological risk indicators in textual data. The model jointly learns suicidal ideation detection, depression severity classification, and perceived threat detection, while incorporating emotional representations derived from valence, arousal, and polarity signals. To improve robustness, an adversarial training strategy is employed to simulate obfuscated expressions, enhancing robustness and generalization under linguistic perturbations. Interpretability is ensured through a hybrid explainable AI approach combining attention mechanisms and SHAP-based feature attribution. Extensive experiments conducted on four benchmark datasets demonstrate that the proposed approach achieves state-of-the-art performance (F1-score = 0.91), with statistically significant improvements over strong baselines. Additional analyses, including ablation studies, adversarial evaluation, and calibration assessment, confirm the effectiveness, robustness, and reliability of the proposed framework. These results highlight the potential of the model for deployment in high-stakes applications such as clinical triage and online risk monitoring, where early and interpretable detection of concealed psychological distress is essential.

Information doi: 10.3390/info17050441

Authors: Hengfei Wang Ting Yu Yuanzheng Xin Zonghui Lu Shuangjian Li Yingting Luo Guoan Ye Helin Gong Tao Zhu

Nuclear Fuel Reprocessing literature contains critical experimental parameters, safety information, theoretical relations, and process data that are highly heterogeneous and subject to strict logical constraints. Manually interpreting complex charts and handling tedious database schema mappings imposes a high cognitive load on experts. Although existing Large Multimodal Models (LMMs) have demonstrated strong potential in information extraction, they often face engineering bottlenecks—such as poor structural compliance and a tendency to confuse entity logic—when dealing with domain databases containing complex foreign key constraints. To address this, we propose GenForge, a schema-aware extraction framework. By taking the target database schema as an explicit constraint, GenForge achieves automatic task decomposition and formatting self-correction via a “Generation–Execution–Reflection–Reforging” iterative loop. Additionally, a Local ID mechanism is introduced to ensure data lineage consistency. We evaluated GenForge on four internal evaluation corpora from nuclear fuel reprocessing literature, each aligned with a distinct database schema: Safety Event and Causal Context Extraction Schema, Property-Condition Data Extraction Schema, Model-Parameter Association Schema, and Process Topology and Stream Mapping Schema. On the independent test set, GenForge achieved 88.0% precision, 83.0% recall, and a 98.6% Schema Compliance Rate (SCR). These results indicate that GenForge, as an expert-assisted framework, reduces the need for manual JSON debugging and supports practical schema-constrained knowledge extraction under four schema-specific evaluation settings within the Nuclear Fuel Reprocessing domain.

Information doi: 10.3390/info17050440

Authors: Ndukwe Ukeje Jairo A. Gutierrez Krassie Petrova

The government’s adoption of cloud computing is critical for digital transformation, but it faces persistent concerns over information security, privacy, governance, and risk. This study examines the factors influencing a government’s intention to adopt cloud services, adapting the Unified Theory of Acceptance and Use of Technology (UTAUT) with constructs tailored to the public sector. A cross-sectional survey was conducted across 90 Nigerian government organisations, producing 230 valid responses from IT professionals, administrators, and policy personnel. The statistical analysis of the data was conducted using SPSS and structural equation modelling in AMOS. Validity and reliability were confirmed through composite reliability, Cronbach’s alpha, and discriminant validity measures. Findings show that privacy (β = 0.11, p < 0.05), governance framework (β = 0.34, p < 0.001), performance expectancy (β = 0.38, p < 0.001), and information security (β = 0.10, p < 0.05) significantly influence government intention to adopt cloud services. Performance expectancy emerged as the strongest predictor. Contrary to expectations, perceived risk did not significantly moderate the relationships, and interaction terms were non-significant. The final model explained 45% of the variance in adoption intention (R2 = 0.45). The study highlights the importance of strengthening governance frameworks, emphasising tangible performance outcomes, and positioning information security and privacy as an enabler of adoption rather than a barrier. By adapting UTAUT to the government context and disentangling the role of perceived risk, the study offers both theoretical refinement and practical guidance for policymakers aiming to accelerate digital transformation and secure cloud adoption.

Information doi: 10.3390/info17050439

Authors: Le Song Jie Liu Maizura Yasin Marzni Mohamed Mokhtar

Health chatbots have been growing into a necessary tool for dealing with risky and important contexts, such as medical and health information seeking. Meanwhile, trust towards chatbots influences people’s willingness to embrace technology and use it consistently. Thus, it is important to explore the mechanism of forming trust towards the health chatbots. The TAM has been introduced to explain the mechanism. This study extends the TAM framework by incorporating perceived risk and self-efficacy to develop an expanded model that explains the mechanisms underlying trust formation in health chatbots, applying a survey and investigating 480 Chinese chatbot users on the Credamo. The findings show that perceived risk reduces trust both directly and indirectly through perceived usefulness, perceived ease of use, and self-efficacy. Both parallel and serial mediation pathways were supported. These results offer a more complete insight into trust formation in high-risk AI contexts and provide practical guidance for chatbot design and governance in health communication.

Information doi: 10.3390/info17050438

Authors: Parul Naib Jaeyoung Park Paniz Abedin Christian King Varadraj Gurupur

Leveraging Artificial Intelligence (AI) ethically in connected healthcare systems requires a quantifiable framework that measures not only outcome correctness, but also the clarity, auditability, and ethical acceptability of model explanations in high-stakes clinical and cybersecurity workflows. This manuscript first presents a narrative review of ethical risks and countermeasures in Healthcare Internet of Things (HIoT) and explains why existing performance metrics are insufficient for trustworthy deployment. We then formalize a quantitative metric called Ethical Explainability (Ee) as a composite index integrating (1) a Human Agreement Ratio (HAR), capturing concordance between AI recommendations (and their rationale) and a calibrated expert consensus, and (2) an Entropy Reduction Index (ERI), capturing the proportional reduction in expert uncertainty after receiving an explanation, operationalized via probability-elicitation questionnaires mapped to Shannon entropy. Designed for HIoT security monitoring, Ee links transparency with governance-ready evidence of trustworthiness for human–AI collaboration.

Information doi: 10.3390/info17050434

Authors: Enrico Barbierato

Modern cyber–physical infrastructures rely heavily on alarm and notification systems to direct human attention when abnormal conditions occur. These mechanisms support timely and safe responses by informing operators and occupants about potential hazards. At the same time, research in human factors has shown that repeated or excessive alerts can weaken vigilance, slow reactions, and reduce confidence in warning systems. This behavioral pattern is commonly described as alarm fatigue. This paper examines how that vulnerability can be exploited intentionally. We refer to this adversarial strategy as alarm poisoning: the deliberate injection of false or misleading alerts in order to increase alarm pressure, erode trust in the monitoring infrastructure, and degrade organizational responsiveness over time. To study this process, we develop a stochastic Cybersecurity Dynamics model representing the interaction among attackers, defenders, alarm infrastructure, and a population of employees. Employee behavior is modeled through evolving trust and fatigue levels, while the overall system is formulated as a continuous–time Markov chain and simulated using the Gillespie Stochastic Simulation Algorithm. A Monte–Carlo campaign is used to analyze the resulting socio–technical dynamics under alternative attacker strategies. The study evaluates time-dependent trust, fatigue, and alarm-pressure trajectories, the distribution of times to behavioral collapse, and defender timing through Trust–Resilience–Agility–Mitigation (TRAM) metrics. The revised analysis also includes replication-sufficiency diagnostics, one-at-a-time sensitivity analysis, and threshold-robustness checks for the collapse criterion. The results show that false alarms with high perceived severity drive alarm pressure upward and degrade trust faster than nuisance-dominated campaigns, even when the total fake-alarm intensity is held constant across strategies. Collapse timing remains highly variable across stochastic realizations, and a non-negligible fraction of runs do not reach the collapse threshold within the simulation horizon. Sensitivity analysis indicates that the main qualitative ranking of attacker strategies is robust across most tested perturbations, with fatigue recovery and defender escalation emerging as particularly influential mechanisms. Overall, the findings support the view that alarm poisoning is a credible socio–technical attack vector and highlight the importance of rapid mitigation, robust alarm management, and human-centered defensive design in cyber–physical security systems.

Information doi: 10.3390/info17050437

Authors: Iraklis Katsaris Sakellaris Sfakiotakis Ilias Logothetis Nikolas Vidakis

Adaptive recommendation mechanisms are widely used to personalise digital learning environments; however, many existing approaches prioritise algorithmic optimisation while providing limited insight into how recommendation behaviour aligns with pedagogically structured instructional artefacts, such as worksheets. To address this gap, this paper proposes a hybrid recommendation approach for adaptive worksheet generation that integrates content-based and collaborative filtering with explicit pedagogical constraints derived from Bloom’s Revised Taxonomy. The system ranks and selects learning and evaluation objects across cognitive levels by combining learner profiles, behavioural signals, and similarity-based information within a unified scoring framework. A simulation-based evaluation was conducted to examine the internal behaviour, stability, and instructional alignment of the recommendation engine under controlled conditions, using Bloom-aligned worksheets and synthetic learner profiles. The analysis focuses on expected–actual alignment and adaptive variation across cognitive levels rather than learning outcomes. Results indicate strong alignment with the intended instructional structure at lower cognitive levels, while bounded and interpretable adaptive variation emerges at higher levels. Evaluation object recommendations showed high agreement with the instructional design, exceeding 95% across simulated conditions. Overall, the study demonstrates how hybrid recommendation mechanisms can support adaptive content selection in pedagogically structured learning scenarios, offering a transparent and robust foundation for information-driven educational systems.

Information doi: 10.3390/info17050435

Authors: Kingsley Attai Daniel Asuquo Kingsley Akputu Okure Obot Cornelia Thomas Faith-Valentine Uzoka Ekerette Attai Christie Akwaowo Faith-Michael Uzoka

Urinary Tract Infections (UTIs) represent one of the most prevalent bacterial infections globally, posing significant health burdens, especially in low- and middle-income countries (LMICs), due to delayed diagnoses, limited access to laboratory services, and rising antimicrobial resistance. This study presents a machine learning (ML)-based diagnostic support framework for early UTI detection, leveraging structured clinical data and explainable artificial intelligence (XAI) techniques to enhance interpretability and trust among healthcare providers. A patient dataset containing 4865 records was used in the study to train and test Extreme Gradient Boosting (XGBoost), Decision Tree (DT) and Random Forest (RF) classifiers, while class imbalance was addressed using Synthetic Minority Over-sampling Technique (SMOTE). The performance of the models was evaluated through accuracy, precision, recall, F1-score, Log Loss, and AUC-ROC, and random forest showed the best results (accuracy: 86.43%, F1-score: 86.71%, AUC-ROC: 0.8695). To ensure that such models can be adopted by stakeholders in the health sector, Local Interpret-able Model-agnostic Explanations (LIME) were integrated, which identified painful urination, urinary frequency, and suprapubic pain as primary predictors in the model. This study shows that interpretable ML models can be helpful in resource-limited regions in predicting UTIs, thereby rendering a solution to improve the management of infections in these regions.

Information doi: 10.3390/info17050436

Authors: Vijayasri Sundarapuram Soundayan Sina Vafi

The paper presents a modification of the interleaver used to construct Parallel Concatenated Block (PCB) codes with flexible lengths. This is accomplished for PCB codes whose interleaved message blocks have at most two bits of a message block. For this purpose, a two-step permutation is implemented, ensuring that the minimum weight of PCB codes has low multiplicity and is obtained from messages with weight one. Conducted analysis and simulations confirm that the new interleaver improves the performance of the PCB code (by 0.25 dB), which is evident at medium to high signal-to-noise ratios. Such improvement is evident for long-length PCB codes.