Information

80 pages, 5436 KB

Open AccessArticle

Global Virtual Prosumer Framework for Secure Cross-Border Energy Transactions Using IoT, Multi-Agent Intelligence, and Blockchain Smart Contracts

by Nikolaos Sifakis

Information 2026, 17(4), 396; https://doi.org/10.3390/info17040396 - 21 Apr 2026

Viewed by 529

Abstract

Global decarbonization and the rapid growth of distributed energy resources increase the need for information-centric mechanisms that can support secure, scalable, cross-border coordination under heterogeneous technical and regulatory conditions. This paper proposes a Global Virtual Prosumer (GVP) framework that integrates IoT sensing, multi-agent [...] Read more.

Global decarbonization and the rapid growth of distributed energy resources increase the need for information-centric mechanisms that can support secure, scalable, cross-border coordination under heterogeneous technical and regulatory conditions. This paper proposes a Global Virtual Prosumer (GVP) framework that integrates IoT sensing, multi-agent coordination, and permissioned blockchain smart contracts to operationalize cross-border energy services as auditable service commitments rather than physical power exchange. Building on prior work that validated MAS-based power management and blockchain-secured operation within individual Virtual Prosumers, the present contribution lies in the cross-border coordination layer and its associated contractual and evaluation mechanisms, not in the constituent technologies themselves. A layered IoT–AI–blockchain architecture is introduced, where off-chain optimization produces allocations and admissibility indicators and on-chain contracts enforce identity, feasibility guards, delegation and partner-assignment rules, oracle verification, and settlement time compliance outcomes. The contractual lifecycle is formalized through four smart-contract algorithms covering trade registration, conditional delegation, cooperative fulfillment, and cross-border settlement with explicit failure semantics and event-based audit trails. The framework is evaluated on a global case study with seven Virtual Prosumers and quantified using contract-centric KPIs that capture registration time rejections, settlement success versus non-compliance, oracle-driven failure attribution, and full lifecycle traceability. The results demonstrate internal consistency of the proposed lifecycle and the practical value of KPI-driven accountability for cross-border energy service coordination. At the same time, the evaluation is based on synthetic parameterization and an emulated contract environment; realistic deployment constraints—including consensus latency, cross-region communication reliability, and regulatory overlap—are discussed as explicit limitations and directions for future empirical validation. Full article

(This article belongs to the Special Issue IoT, AI, and Blockchain: Applications, Security, and Perspectives)

► Show Figures

Figure 1

28 pages, 1008 KB

Open AccessReview

Deep Learning for Credit Risk Prediction: A Survey of Methods, Applications, and Challenges

by Ibomoiye Domor Mienye, Ebenezer Esenogho and Cameron Modisane

Information 2026, 17(4), 395; https://doi.org/10.3390/info17040395 - 21 Apr 2026

Cited by 1 | Viewed by 1452

Abstract

Credit risk prediction is central to financial stability and regulatory compliance, guiding lending decisions and portfolio risk management. While traditional approaches such as logistic regression and tree-based models have long been the industry standard, recent advances in deep learning (DL) have introduced architectures [...] Read more.

Credit risk prediction is central to financial stability and regulatory compliance, guiding lending decisions and portfolio risk management. While traditional approaches such as logistic regression and tree-based models have long been the industry standard, recent advances in deep learning (DL) have introduced architectures capable of capturing complex nonlinearities, temporal dynamics, and relational dependencies in borrower data. This study provides a comprehensive review of DL methods applied to credit risk prediction, covering multi-layer perceptron, recurrent and convolutional neural networks, transformer, and graph neural networks. We examine benchmark and large-scale datasets, highlight peer-reviewed applications across corporate, consumer, and peer-to-peer lending, and evaluate the benefits of DL relative to classical machine learning. In addition, we critically assess key challenges and identify emerging opportunities. By synthesising methods, applications, and open challenges, this paper offers a roadmap for advancing trustworthy deep learning in credit risk modelling and bridging the gap between academic research and industry deployment. Full article

(This article belongs to the Special Issue Predictive Analytics and Data Science, 3rd Edition)

► Show Figures

Figure 1

20 pages, 1137 KB

Open AccessArticle

Diagonal Adaptive Graph: Revisiting Channel Dependency in Multivariate Time Series Forecasting

by Xiang Li, Yanping Zheng and Zhewei Wei

Information 2026, 17(4), 394; https://doi.org/10.3390/info17040394 - 21 Apr 2026

Viewed by 524

Abstract

Adaptive graph learning has become a widely adopted paradigm for multivariate time series forecasting when explicit physical topology is unavailable. In these approaches, node embeddings are typically used to construct dense adjacency matrices based on pairwise similarity, implicitly coupling representation learning with relational [...] Read more.

Adaptive graph learning has become a widely adopted paradigm for multivariate time series forecasting when explicit physical topology is unavailable. In these approaches, node embeddings are typically used to construct dense adjacency matrices based on pairwise similarity, implicitly coupling representation learning with relational modeling. However, we observe that under identical training settings but different random initializations, the learned adjacency matrices can vary substantially while predictive performance remains nearly unchanged, indicating that the relational structure is often underdetermined by the forecasting objective. This observation suggests a mismatch between similarity-based structural learning and the forecasting objective. In this work, we revisit node embeddings from a sequence approximation perspective and propose a Diagonal Adaptive Graph (DiAG) module that restricts adaptive learning to diagonal elements. The diagonal coefficients are derived from channel-independent predictions, while off-diagonal interactions are constructed from the similarity of input sequences. This design decouples representation learning from relational modeling, allowing variables to adaptively switch between channel-independent and channel-dependent regimes. Experiments on multiple datasets show that DiAG improves forecasting performance without modifying the channel-independent backbones. These results indicate that channel-dependent forecasting can be achieved as a prediction-driven refinement over channel-independent backbones, without requiring fully learned dense relational structures. Full article

(This article belongs to the Special Issue Deep Learning Approach for Time Series Forecasting)

► Show Figures

Figure 1

35 pages, 1381 KB

Open AccessArticle

Formality Requirements in the Era of Smart Contracts: A Mixed-Methods Analysis of Emerging Challenges

by Nabeel Mahdi Althabhawi, Ra’ed Fawzi Aburoub, Rizal Rahman, Faris Kamil Hasan Mihna and Hazim Akram Sallal

Information 2026, 17(4), 393; https://doi.org/10.3390/info17040393 - 21 Apr 2026

Viewed by 918

Abstract

Smart contracts raise persistent challenges regarding compliance with traditional contract formalities, including writing, signature, notarization, and in certain transactions, registration. These issues are particularly significant in high-value and public-facing transactions such as real estate, where formalities determine legal validity, evidentiary sufficiency and publicity [...] Read more.

Smart contracts raise persistent challenges regarding compliance with traditional contract formalities, including writing, signature, notarization, and in certain transactions, registration. These issues are particularly significant in high-value and public-facing transactions such as real estate, where formalities determine legal validity, evidentiary sufficiency and publicity effects. While existing scholarly work has examined these challenges from either doctrinal or technological perspectives, limited attention has been given to how the functional roles of formalities interact with blockchain architecture, practitioner perceptions and institutional legal frameworks. This study addresses this gap through a mixed-methods approach combining doctrinal legal analysis with qualitative socio-legal research based on 27 semi-structured interviews with legal professionals including attorneys, judges, and academic scholars. The analysis is grounded in a civil law framework, with particular reference to the Jordanian legal system, while references to the European Union’s eIDAS Regulation are used illustratively to demonstrate regulatory approaches to digital authentication. The findings demonstrate that blockchain-based systems can effectively support the evidentiary and attribution functions of contractual formalities through cryptographic verification, consensus mechanisms, and automated execution. However, they do not independently satisfy formalities that perform cautionary, constitutive, protective or public order function, namely notarization and registration, which remain dependent on institutional validation and legal recognition. The analysis further shows that practitioner concerns reflect not only doctrinal constraints but also institutional roles and varying levels of technical familiarity. To address these limitations, the study proposes a function-based analytical framework for evaluating smart contract formalities and identifies two complementary pathways for legal adaptation: (i) institutional integration, including registry-linkage systems and hybrid contracts; and (ii) technological adaptation, including digital authentication frameworks and legal oracles that connect on-chain execution to off-chain legal conditions. The study concludes that smart contract formalities’ challenges arise not solely from technological limitations, but from the interaction between legal doctrine, institutional structures, and system design. It advances a functional framework for aligning automation with the evidentiary, protective, and publicity functions of contractual formalities. Full article

(This article belongs to the Special Issue Recent Advances in Smart Contract and Blockchain Analysis)

► Show Figures

Figure 1

22 pages, 1994 KB

Open AccessArticle

Haipai New Year Paintings Segmentation Design Based on PSE-Net

by Yueyang Zhao, Jingru Zhang, Jin Liu and Damin Ding

Information 2026, 17(4), 392; https://doi.org/10.3390/info17040392 - 21 Apr 2026

Viewed by 487

Abstract

Chinese Haipai New Year paintings are an important part of the country’s intangible cultural heritage, and their digital preservation holds great significance. This paper proposes PSE-Net (Pyramid Scale Expansion Network), a deep learning-based segmentation method specifically designed to handle the complex textures and [...] Read more.

Chinese Haipai New Year paintings are an important part of the country’s intangible cultural heritage, and their digital preservation holds great significance. This paper proposes PSE-Net (Pyramid Scale Expansion Network), a deep learning-based segmentation method specifically designed to handle the complex textures and intricate compositions of these artworks. By constructing a dedicated large-scale dataset, we trained PSE-Net to achieve high-precision segmentation by incorporating attention mechanisms and multi-scale feature fusion to better capture detailed features. Experimental results demonstrate that the proposed method outperforms existing approaches (such as ResNet) in terms of segmentation performance, yielding superior results in edge preservation. This work establishes the first automated tool for the pixel-level analysis of Haipai New Year paintings, thereby facilitating museum digitization, art history research, and education. Furthermore, it offers new insights for the image processing and digital preservation of other traditional artworks. Full article

(This article belongs to the Section Artificial Intelligence)

► Show Figures

Figure 1

37 pages, 808 KB

Open AccessArticle

Re-Examining Organisational Performance: An Empirical Study on the Relationships Between Revenue, Net Profit, Cash Flow per Share, and Earnings per Share in Australian Energy Firms

by Kabossa A. B. Msimangira, Shirley Wong and Sitalakshmi Venkatraman

Information 2026, 17(4), 391; https://doi.org/10.3390/info17040391 - 20 Apr 2026

Viewed by 842

Abstract

New approaches to improve organisational performance in firms are evolving in this data-driven age. However, there is lack of studies in examining the relationship between revenue, net profit, cash flow per share, and earnings per share. The energy sector remains under-researched regarding the [...] Read more.

New approaches to improve organisational performance in firms are evolving in this data-driven age. However, there is lack of studies in examining the relationship between revenue, net profit, cash flow per share, and earnings per share. The energy sector remains under-researched regarding the multi-dimensional drivers of profitability. Existing research shows inconclusive evidence with studies predominantly examining revenue—performance relationship limiting to a single factor and not guiding potential investors regarding future earnings per share in the energy industry. This paper aims to bridge the gap in literature by proposing a data-driven approach to analyse the relationships between revenue, net profit, cash flow per share, and earnings per share. We examine these relationships by conducting an empirical analysis using secondary data derived from published annual reports of the energy firms listed on the Australian Securities Exchange (ASX). Our empirical study uses Pearson correlations and regression techniques to test the hypotheses on the relationships between revenue, net profit, cash flow per share, and earnings per share. Also, we use market capitalisation as a control variable and predictor of earnings per share in the energy industry. The data analysis results in four findings: (i) revenue positively influences earnings per share because higher revenue expands the firm’s earnings capacity within the financial performance, (ii) net profit has a strong positive effect on earnings per share, consistent with profitability theory and the direct derivation of EPS from net income, (iii) cash flow per share influences earnings per share because liquidity supports operational stability, investment decisions, and earnings sustainability (e.g., heavy capital expenditure contexts), and (iv) the combined effects of revenue, net profit, and cash flow per share provide a stronger and more holistic prediction of earnings per share than any single variable, consistent with multidimensional organisational performance theory (a more holistic valuation model than looking at single factors). In addition, the results indicate that market capitalisation (control variable) has both strong prediction of earnings per share and strong association with earnings per share. The results of this study can offer practitioners and investors in Australia and other countries for a better understanding of the relationships between revenue, net profit, cash flow per share, and earnings per share from energy companies. The data will help investors to make good investment data-driven decisions in the energy industry or other industries. It also motivates researchers to conduct similar studies in different contexts. We further provide recommendations, including a closed-loop Artificial Intelligence (AI) data-driven approach integrated into energy accounting and operational processes to enhance profitability. This approach operationalises the revenue and earnings-per-share (EPS) strategies identified in our empirical analysis, offering practical value for industry practitioners and guiding future research in this direction. Full article

(This article belongs to the Section Information Applications)

► Show Figures

Graphical abstract

21 pages, 8832 KB

Open AccessSystematic Review

Transforming Financial Reporting: A Systematic Literature Review on the Synergistic Role of Artificial Intelligence and Blockchain

by Jinfeng Wang, Jiaqi Chen, William Yeoh and Jingzhu Chen

Information 2026, 17(4), 390; https://doi.org/10.3390/info17040390 - 20 Apr 2026

Viewed by 878

Abstract

As global digital transformation accelerates, artificial intelligence (AI) and blockchain technologies have evolved from theoretical concepts into practical tools within the field of accounting, particularly in financial reporting. This study conducts a systematic review of 62 sources drawn from major academic databases to [...] Read more.

As global digital transformation accelerates, artificial intelligence (AI) and blockchain technologies have evolved from theoretical concepts into practical tools within the field of accounting, particularly in financial reporting. This study conducts a systematic review of 62 sources drawn from major academic databases to develop a comprehensive framework for classifying application scenarios. The findings indicate that the application of artificial intelligence and blockchain technology can help improve the efficiency of financial report generation, enhance the reliability of data, and promote innovation in the auditing process. Nevertheless, persistent challenges remain, including concerns related to data security, technological limitations, and regulatory gaps. The study proposes a structured roadmap for the implementation of these technologies, underscoring their transformative potential in advancing the digital evolution of accounting, while also identifying key directions for future research. Full article

(This article belongs to the Section Information Systems)

► Show Figures

Figure 1

17 pages, 913 KB

Open AccessArticle

An Empirical Study of Knowledge Graph-Enhanced RAG for Information Security Compliance

by Dimitar Jovanovski, Marija Stojcheva, Mila Dodevska, Petre Lameski, Igor Mishkovski and Dejan Gjorgjevikj

Information 2026, 17(4), 389; https://doi.org/10.3390/info17040389 - 20 Apr 2026

Viewed by 1897

Abstract

Information security compliance has become critical for organizations worldwide, with the ISO/IEC 27000 family serving as the most widely adopted framework for establishing information security management systems. Despite their global acceptance, these standards present significant interpretation challenges due to their formal language, abstract [...] Read more.

Information security compliance has become critical for organizations worldwide, with the ISO/IEC 27000 family serving as the most widely adopted framework for establishing information security management systems. Despite their global acceptance, these standards present significant interpretation challenges due to their formal language, abstract structure, and extensive cross-referencing across 97 documents. Traditional retrieval-augmented generation (RAG) systems, which rely on independent text chunking and dense vector retrieval, prove inadequate for such highly interconnected regulatory materials, often fragmenting contextual relationships and reducing accuracy. This study introduces a privacy-preserving RAG framework that integrates LightRAG, a knowledge graph-based retrieval system, with locally hosted open-source language models. Unlike chunk-based RAG systems that treat document segments independently, the system in this study constructs a semantic knowledge graph that explicitly models relationships between clauses through typed edges representing cross-references, semantic similarity, and hierarchical dependencies. To enable rigorous evaluation, we developed a curated benchmark dataset of 222 multiple-choice questions with authoritative ground-truth answers, systematically constructed from official ISO standards, certification preparation materials, and academic sources. Through systematic evaluation on this benchmark, we show that knowledge graph-based retrieval achieves higher accuracy than chunk-based RAG and non-retrieval LLM baselines within the evaluated setup. The analysis indicates that embedding model quality is strongly associated with system performance, that hybrid retrieval modes combining local and global graph traversal tend to yield better accuracy, and that mid-sized open-source models paired with strong retrievers can approach the performance of larger proprietary systems. The best configuration achieves 90.54% accuracy, demonstrating the promising effectiveness of graph-structured retrieval for multiple-choice regulatory questions. Full article

(This article belongs to the Special Issue Advanced Retrieval-Augmented Generation Systems Based on Large Language Models)

► Show Figures

Figure 1

34 pages, 2746 KB

Open AccessArticle

AdaptiveNet: A Novel Architecture for Reducing Computation Complexity to Fake Review Classification

by Deepalakshmi Perumalsamy, Sharon Roji Priya Cornelius and Rajermani Thinakaran

Information 2026, 17(4), 388; https://doi.org/10.3390/info17040388 - 20 Apr 2026

Viewed by 627

Abstract

The exponential rise of e-commerce platforms has resulted in a dramatic increase in online reviews, which creates a challenge in distinguishing fake reviews that erode consumer confidence and harm commerce ecosystems. Traditional approaches for fake review detection employ computationally expensive deep learning networks [...] Read more.

The exponential rise of e-commerce platforms has resulted in a dramatic increase in online reviews, which creates a challenge in distinguishing fake reviews that erode consumer confidence and harm commerce ecosystems. Traditional approaches for fake review detection employ computationally expensive deep learning networks which are resource-intensive and difficult to use in practice. In this paper, we describe AdaptiveNet, a new lightweight neural architecture that achieves fake review detection with much lower computational resources while maintaining a higher detection and classification precision. The model proposed in this paper is based on three original innovations: a Multi-Scale Semantic Fusion (MSSF) layer for hierarchical feature extraction, Dynamic Attention Scaling (DAS) with complexity measure attention, and Adaptive Parameter Sharing (APS) context-gated networks. With thorough evaluation on Amazon, Yelp, and TripAdvisor datasets of reviews totalling 1.2 million reviews, AdaptiveNet attains 94.8% accuracy while achieving 65% computational overhead in comparison to traditional models. The architecture outperformed all other state-of-the-art models, BERT-base (92.1%), RoBERTa (91.8%), and other more recent efficient models, requiring 70% lower parameters and 60% lower energy consumption. This work markedly advances the other efficient deep learning architectures for text classification and allows for the practical implementation of fake review detection systems in resource-limited settings as process innovation. Full article

(This article belongs to the Section Information Applications)

► Show Figures

Figure 1

23 pages, 1085 KB

Open AccessReview

A Scoping Analysis of Literature on the Enhancement in Security in Financial Messaging Systems

by Unarine Madzivhandila and Colin Chibaya

Information 2026, 17(4), 387; https://doi.org/10.3390/info17040387 - 20 Apr 2026

Viewed by 718

Abstract

The security of financial messaging systems is critical to maintaining trust in digital financial platforms. Despite advances in cryptography, many contemporary systems remain vulnerable to channel-based and cryptographic threats, including eavesdropping, interception, tampering, and unauthorized access. Hybrid cryptographic models that combine asymmetric encryption [...] Read more.

The security of financial messaging systems is critical to maintaining trust in digital financial platforms. Despite advances in cryptography, many contemporary systems remain vulnerable to channel-based and cryptographic threats, including eavesdropping, interception, tampering, and unauthorized access. Hybrid cryptographic models that combine asymmetric encryption for secure key exchange with symmetric encryption for efficient data protection have emerged as effective approaches for strengthening confidentiality, integrity, and authenticity in financial message communications. This study presents a scoping review of literature published between 2015 and 2025, mapping research on user vulnerabilities in financial messaging systems and examining the role of hybrid cryptographic models in mitigating these risks. Guided by the PRISMA-ScR reporting standards, 615 articles were identified across nine scholarly databases. Forty-four studies met the inclusion criteria after systematic screening. The findings reveal a growing emphasis on hybrid encryption strategies, particularly RSA–AES and ECC–AES combinations, due to their balance of security strength and computational efficiency. However, significant gaps persist in empirical validation, real-world deployment, and user-centred security design, especially in mobile-first and resource-constrained environments. Existing research largely prioritizes theoretical performance and algorithmic efficiency, with limited attention to practical integration, usability, and operational constraints. This review highlights the need for holistic security frameworks that integrate cryptographic robustness with usability, regulatory compliance, and contextual deployment considerations. It provides a structured foundation for future research focused on developing scalable, user-centric, and resilient security solutions for financial messaging systems. Full article

(This article belongs to the Section Information Systems)

► Show Figures

Figure 1

21 pages, 24433 KB

Open AccessArticle

A Novel Deep Learning Model for Predicting University English Proficiency Achievement of Students

by Yan Yang, Xiaowei Wang, Mohan Liu, Huiwen Xue and Laixiang Xu

Information 2026, 17(4), 386; https://doi.org/10.3390/info17040386 - 19 Apr 2026

Viewed by 480

Abstract

The rapid expansion of English major enrollment has exposed critical limitations in traditional academic assessment methods regarding efficiency and accuracy, constraining educational quality enhancement. This paper introduces an English proficiency assessment approach utilizing an improved RegNet architecture integrated with a dual attention mechanism. [...] Read more.

The rapid expansion of English major enrollment has exposed critical limitations in traditional academic assessment methods regarding efficiency and accuracy, constraining educational quality enhancement. This paper introduces an English proficiency assessment approach utilizing an improved RegNet architecture integrated with a dual attention mechanism. The multidimensional academic data processed by our model include attendance, online participation, language practice, and assessment scores for listening, speaking, reading, and writing from undergraduate English majors. The initial downsampling module of RegNet is optimized through a dual convolutional structure to augment shallow feature extraction. Subsequently, a deformable attention mechanism (DAT) is incorporated to enhance focus on salient features, while a graph attention network (GAT) facilitates interaction and fusion among academic node features. Experimental results demonstrate that the proposed method achieves an average accuracy of 99.46% in proficiency assessment, substantially outperforming mainstream models including EfficientNet and AlexNet. Additionally, it demonstrates robust edge deployment capabilities, providing an effective technical solution for intelligent academic management of English programs within smart campus frameworks. Full article

(This article belongs to the Section Artificial Intelligence)

► Show Figures

Figure 1

26 pages, 1033 KB

Open AccessArticle

Measuring the Awareness of Cyber Guardians, Situation Awareness Predicts Operator Performance

by Håvard Jakobsen Ofte and Sokratis Katsikas

Information 2026, 17(4), 385; https://doi.org/10.3390/info17040385 - 19 Apr 2026

Viewed by 470

Abstract

Situation Awareness (SA) of operators in Cyber Security (CS) has been assumed important for effective incident response in critical infrastructure. Many previous studies have proposed tools and methods to improve SA, but there is a general lack of empirical evidence on the impact [...] Read more.

Situation Awareness (SA) of operators in Cyber Security (CS) has been assumed important for effective incident response in critical infrastructure. Many previous studies have proposed tools and methods to improve SA, but there is a general lack of empirical evidence on the impact of SA on performance in this domain. In this study, we present such empirical evidence from experiments done within the domain of critical infrastructure. Eleven professional CS operators from the power sector participated in a realistic simulated network-related incident response task. SA, experience, and performance were scored for each participant. SA was measured using the Situation Awareness Global Assessment Technique (SAGAT). Statistically significant results confirmed three hypotheses: Higher SA was correlated with higher performance, and longer experience as an operator was correlated with higher performance. Additionally, higher SA is predictive of higher performance independently of longer experience. Accordingly, initial empirical evidence supporting the conjecture that cyber-SA is positively associated with performance now exists. Implications include that further CS research aimed at tool development and operator training should use recognized SA measurements as demonstrated in this study. This study thus contributes to bridging the existing knowledge gap of cyber-SA. Full article

(This article belongs to the Section Information Security and Privacy)

► Show Figures

Figure 1

21 pages, 1703 KB

Open AccessArticle

Three-Way Multimodal Learning with Severely Missing Modalities

by Hanrui Wang, Yu Fang, Xin Wang and Fan Min

Information 2026, 17(4), 384; https://doi.org/10.3390/info17040384 - 19 Apr 2026

Viewed by 565

Abstract

Missing modalities remain a major obstacle to the real-world deployment of multimodal learning systems, as incomplete inputs can substantially degrade model performance. Existing methods often suffer from biased imputation under high missing rates and lack uncertainty-aware, differentiated processing. Inspired by three-way decision, a [...] Read more.

Missing modalities remain a major obstacle to the real-world deployment of multimodal learning systems, as incomplete inputs can substantially degrade model performance. Existing methods often suffer from biased imputation under high missing rates and lack uncertainty-aware, differentiated processing. Inspired by three-way decision, a framework for handling uncertainty by adding a deferment option to acceptance and rejection, we propose three-way multimodal learning with severely missing modalities (3WML-SMMs), a novel framework that introduces a three-way decision mechanism into both missing-modality imputation and feature regularization for the first time. Specifically, 3WML-SMM treats variance not merely as a descriptive measure of uncertainty, but as a decision signal for adaptive processing. Based on this idea, the framework incorporates (1) a variance-guided three-way imputation strategy with accept–delay–reject decisions to reduce unreliable reconstruction when only a limited number of complete samples are available and (2) a dimension-wise adaptive feature enhancement module that performs fine-grained regularization according to perturbation uncertainty. Experiments on the CMU Multimodal Opinion Sentiment Intensity (CMU-MOSI) and Multimodal Internet Movie Database (MM-IMDb) datasets show that 3WML-SMM consistently outperforms representative baselines, including reconstruction-based methods, complete-input multimodal methods, and missing-modality-specific methods under severe missing-modality settings, with statistically significant improvements over the multimodal learning with severely missing modality (SMIL) baseline (

p < 0.05

). These results demonstrate the effectiveness of the proposed framework, even in extreme settings where only 10% of the text modality is available. Full article

(This article belongs to the Section Artificial Intelligence)

► Show Figures

Graphical abstract

28 pages, 8399 KB

Open AccessArticle

Machine Learning-Enabled Secure Unified Framework for Remote Electrocardiogram Monitoring via a Multi-Level Blockchain System

by Chathumi Samaraweera, Dongming Peng, Michael Hempel and Hamid Sharif

Information 2026, 17(4), 383; https://doi.org/10.3390/info17040383 - 18 Apr 2026

Viewed by 467

Abstract

Timely classification of cardiovascular diseases is crucial to improve medical outcomes. Emerging remote patient monitoring systems help achieve this by enabling continuous monitoring of electrocardiogram signals in home environments. However, these systems struggle with unique challenges like missing genuine medical emergencies, rising energy [...] Read more.

Timely classification of cardiovascular diseases is crucial to improve medical outcomes. Emerging remote patient monitoring systems help achieve this by enabling continuous monitoring of electrocardiogram signals in home environments. However, these systems struggle with unique challenges like missing genuine medical emergencies, rising energy demands, scalability challenges, handling vast medical databases, data processing delays, and safeguarding patient records. To overcome these challenges, we propose a single framework with three main phases: (a) an embedded hardware-driven K-Nearest Neighbor (KNN)-assisted real-time ECG monitoring and classification method; (b) a differentiated communication strategy (DCS) formed with a priority-based ECG data packaging framework and multi-layered security protocols; and (c) a multi-level blockchain network (MLBN) architecture armed with adaptive security mechanisms and real-time cross-chain medical data communication bridges. Simulations are conducted using the ECG signals (1000 fragments) dataset and the Ganache Ethereum development framework. The classification accuracies obtained for patient urgent categories U1 to U5 are 91.43%, 95.71%, 94.23%, 90.00%, and 91.43%, respectively. The performance evaluation results of the KNN-guided classification method, along with DCS and MLBN simulation results obtained from average gas consumption analysis, confirms reliability and viability of our framework, while also revolutionizing remote patient monitoring technology and addressing critical challenges in existing systems. Full article

(This article belongs to the Special Issue Machine Learning and Simulation for Public Health)

► Show Figures

Graphical abstract

27 pages, 6204 KB

Open AccessArticle

A Crossover Study on VR and Traditional Instruction in Engineering Education

by Petru-Iulian Grigore, Corneliu Octavian Turcu, Andrei Zaharia and Valentin Nedeff

Information 2026, 17(4), 382; https://doi.org/10.3390/info17040382 - 18 Apr 2026

Viewed by 457

Abstract

Virtual reality (VR) is increasingly used as an interactive instructional medium in engineering education, yet evidence on practical implementation and student-reported experience remains limited. This study examined students’ perceived experience and usability across VR and traditional instruction within a crossover design in a [...] Read more.

Virtual reality (VR) is increasingly used as an interactive instructional medium in engineering education, yet evidence on practical implementation and student-reported experience remains limited. This study examined students’ perceived experience and usability across VR and traditional instruction within a crossover design in a UV-C water disinfection lesson. Using a mixed 2 × 2 crossover design, 52 undergraduate engineering students completed both a VR lesson (Meta Quest 3; Unreal Engine 5.4) and a content-aligned traditional session delivered with slides and a physical UV disinfection stand. After each session, participants reported perceived flow (short Flow Index) and engagement (adapted User Engagement Scale); the System Usability Scale (SUS) was completed after the VR session only. A brief knowledge quiz and open-ended feedback were also collected and used descriptively. Students reported higher perceived flow and engagement in the VR condition than in the traditional condition, and VR usability was generally rated acceptable-to-excellent, with higher SUS scores observed in the VR-first sequence than in the traditional-first sequence. Qualitative feedback emphasized clarity and interactivity, and most participants expressed a preference for a blended approach. Overall, the results support the practical feasibility and positive user acceptance of the VR lesson in this instructional context. The findings also suggest that perceived usability may be associated with instructional sequence, although this pattern should be interpreted cautiously within the perception-based scope of the study. Full article

(This article belongs to the Section Information Applications)

► Show Figures

Figure 1

22 pages, 2661 KB

Open AccessArticle

Generative Design and Evaluation of Industrial Heritage for Tourism Development Based on Kansei Engineering-KANO Model-TOPSIS Method: The Case of Shanghai Libo Brewery

by Qichao Song and Huiling Zhang

Information 2026, 17(4), 381; https://doi.org/10.3390/info17040381 - 18 Apr 2026

Viewed by 690

Abstract

Adaptive reuse of industrial heritage from a tourism perspective presents a complex design challenge requiring a balance between heritage preservation, functional innovation, and diverse stakeholder expectations. However, current practices often face issues such as ambiguous demand interpretation and a disconnect between design generation [...] Read more.

Adaptive reuse of industrial heritage from a tourism perspective presents a complex design challenge requiring a balance between heritage preservation, functional innovation, and diverse stakeholder expectations. However, current practices often face issues such as ambiguous demand interpretation and a disconnect between design generation and systematic evaluation. Addressing these limitations, this paper proposes and illustrates a human–machine collaborative design paradigm that integrates generative AI into a closed-loop process of “demand analysis–intelligent generation–comprehensive evaluation.” The method first employs Kansei Engineering and the KANO model to qualitatively extract and quantitatively prioritise heterogeneous user needs, translating subjective perceptions into structured design constraints and optimisation objectives. Next, these needs are encoded as text prompts to drive targeted spatial exploration by the generative AI tool Nano Banana AI. Finally, the TOPSIS method is applied for multi-criteria performance evaluation and solution selection. A case study of Shanghai Libo Brewery suggests that this paradigm can enhance design efficiency and show potential to outperform traditional methods across dimensions such as historical preservation, public accessibility, ecological integration, social inclusivity, and formal innovation. The research offers a quantifiable and systematically documented intelligent design methodology for industrial heritage renewal, while acknowledging the exploratory nature of the generative phase. Furthermore, it provides a visitor-demand-driven innovation pathway for developing industrial heritage tourism destinations, thereby potentially enhancing cultural experiences and tourism appeal at heritage sites. This research illustrates a move from an experience-driven paradigm toward a data- and value-driven approach, contributing theoretical methodologies to the intersection of cultural tourism and artificial intelligence. Full article

(This article belongs to the Topic The Applications of Artificial Intelligence in Tourism)

► Show Figures

Figure 1

20 pages, 2397 KB

Open AccessArticle

Towards Sustainable AI: Benchmarking Energy Efficiency of Deep Neural Networks for Resource-Constrained Edge Devices

by Rohail Qamar, Raheela Asif and Syed Muslim Jameel

Information 2026, 17(4), 380; https://doi.org/10.3390/info17040380 - 17 Apr 2026

Viewed by 897

Abstract

Deep learning models represent one of the most advanced and effective approaches in predictive modeling. Their hierarchical architectures enable the extraction of complex, non-linear feature relationships and the identification of latent patterns within data, making them highly suitable for tasks involving high-dimensional or [...] Read more.

Deep learning models represent one of the most advanced and effective approaches in predictive modeling. Their hierarchical architectures enable the extraction of complex, non-linear feature relationships and the identification of latent patterns within data, making them highly suitable for tasks involving high-dimensional or unstructured inputs. However, these models are computationally demanding, requiring significant processing resources and time. Furthermore, their predictive performance is largely contingent upon the availability of large-scale datasets. In this study, a Deep Green Framework is employed for the prediction of two computer vision tasks. CIFAR-10 and CIFAR-00 have been taken for image classification. Fifteen convolutional neural network (CNN) variants categorized into light-weight and heavy-weight are trained for the prediction of these two datasets. Based on energy footprint, time, memory usage, Top-1 accuracy, Top-3 accuracy, model size, and model parameters. The study highlights that MobileNetV3-Small produces the best outcomes when compared to other trained models having low task latency and higher efficiency, making it highly suitable for edge environments where resources are scarce. Full article

(This article belongs to the Special Issue Machine Learning for Predictive Analytics: Models, Applications, and Challenges)

► Show Figures

Graphical abstract

20 pages, 5500 KB

Open AccessArticle

DTWICA: A Novel Method for Constructing Character Templates in Imaginary Handwriting

by Jiaofen Nan, Panpan Xu, Gaodeng Fan, Xueqi Jin, Shuyao Zhai, Yanting Li, Yongquan Xia, Yinghui Meng, Liqin Yue and Duan Li

Information 2026, 17(4), 379; https://doi.org/10.3390/info17040379 - 17 Apr 2026

Viewed by 402

Abstract

Imaginary handwriting is an important research paradigm in the field of brain-controlled typing. Neural signals exhibit high complexity, low signal-to-noise ratio, and strong temporal and environmental variability, leading to significant inter-trial differences in the temporal dynamics of character-related signals. These factors pose significant [...] Read more.

Imaginary handwriting is an important research paradigm in the field of brain-controlled typing. Neural signals exhibit high complexity, low signal-to-noise ratio, and strong temporal and environmental variability, leading to significant inter-trial differences in the temporal dynamics of character-related signals. These factors pose significant challenges for segmenting character-related signals and accurately decoding imaginary handwriting. To address these issues, this study proposes a Dynamic Time Warping Independent Component Analysis (DTWICA) framework. This framework employs FastDTW to construct individualized warping functions for each trial, followed by FastICA-based decomposition to separate the signal into distinct temporal and neuronal factors. The decomposed temporal factors are then mapped and transformed using the warping function and subsequently merged with the neuronal factors to reconstruct the signal. A sliding time window is then applied for adaptive processing, yielding the transformed signal. Finally, the transformed signals from multiple trials are averaged to generate a template for each character. Results based on a publicly available neural signals dataset for imaginary handwriting indicate that, compared with mainstream time warping models such as Shift, Linear, Piecewise, and TWPCA, the proposed model improves the character decoding accuracy for 31 characters by 14%, 13%, 7%, and 2%, respectively. This study not only constructs effective character signal templates but also facilitates accurate character segmentation during unlabeled imagined typing in an offline setting, providing a promising methodological basis for future real-time imagined typing decoding systems. Full article

► Show Figures

Figure 1

25 pages, 767 KB

Open AccessArticle

A Qualitative Synthesis of Cyberattack Trends in Managed Service Providers: Analyzing Multi-Tenant Vulnerabilities and Mitigation Strategies

by Shiva Ram Neupane, Neeraj Shrestha and Weiqing Sun

Information 2026, 17(4), 378; https://doi.org/10.3390/info17040378 - 17 Apr 2026

Viewed by 1246

Abstract

Managed Service Providers (MSPs) have increasingly become prime targets for cyberattacks due to their privileged access across multiple client environments. Utilizing a qualitative thematic synthesis and an Open-Source Intelligence (OSINT) methodology, this study examines a purposive sample of major MSP-targeted cyber incidents from [...] Read more.

Managed Service Providers (MSPs) have increasingly become prime targets for cyberattacks due to their privileged access across multiple client environments. Utilizing a qualitative thematic synthesis and an Open-Source Intelligence (OSINT) methodology, this study examines a purposive sample of major MSP-targeted cyber incidents from 2020 to 2025 to identify common attack patterns, exploited vulnerabilities, and operational impacts on downstream clients, particularly small and medium-sized businesses. Analysis of publicly reported incidents reveals a clear trend toward attacks leveraging centralized management platforms, remote access tools, and multi-tenant architectures, resulting in cascading disruptions from limited initial compromise. The synthesis highlights extortion-driven ransomware, supply chain compromises, and the exploitation of unpatched edge devices as dominant threats. To counter these systemic risks, this study outlines contextualized mitigation strategies such as zero trust principles, strict identity controls, tenant isolation, and continuous monitoring tailored to balance security requirements with MSP operational constraints. While these strategies are evidence-informed and grounded in observed trends, they remain proposed solutions that require further empirical validation. The findings emphasize the critical need for proactive, collaborative security practices among MSPs, clients, and regulators to manage evolving cyber threats effectively. Full article

(This article belongs to the Special Issue Security and Privacy Approaches Against Cyber Threats: Innovations, Challenges, and Practical Solutions)

► Show Figures

Graphical abstract

28 pages, 904 KB

Open AccessArticle

Supervised Machine Learning-Based Multiclass Classification and Interpretable Feature Importance Analysis of Teacher Job Satisfaction

by Bouabid Qabliyane, Zakaria Khoudi, Abdelamine Elouafi, Abderrahim Salhi and Mohamed Baslam

Information 2026, 17(4), 377; https://doi.org/10.3390/info17040377 - 17 Apr 2026

Viewed by 702

Abstract

This study examines the increasing concern regarding teacher job satisfaction, which has a direct impact on retention, instructional quality, and student outcomes. Traditionally, teacher satisfaction has been evaluated through questionnaires, which present limitations in terms of data efficiency and analyses. In this study, [...] Read more.

This study examines the increasing concern regarding teacher job satisfaction, which has a direct impact on retention, instructional quality, and student outcomes. Traditionally, teacher satisfaction has been evaluated through questionnaires, which present limitations in terms of data efficiency and analyses. In this study, machine learning techniques were applied to data from the PISA 2022 teacher questionnaire in Morocco (N = 2998 lower-secondary teachers). Two multiclass classification targets were defined: overall job satisfaction (SATJOB_class) and satisfaction with the teaching profession (SATTEACH_class), each categorised into three balanced classes: low (<−0.5), medium (−0.5 to 0.5), and high (>0.5) classes. The methodology comprised four key stages. Initially, comprehensive pre-processing was conducted to address missing values, retaining features with fewer than 300 missing entries and applying mode imputation. Subsequently, nine classifiers, including logistic regression, K-nearest neighbours, multinomial naïve Bayes, support vector machine, decision tree, random forest, XGBoost, AdaBoost, and a feed-forward Artificial Neural Network, were evaluated using identical train/test splits and hyperparameter tuning. Third, the model performance was assessed using accuracy, precision, recall, and F1-score. Finally, the feature importance was derived from tree-based and permutation methods. The results indicated that XGBoost outperformed the other models for SATJOB_class with an accuracy (0.61), precision (0.62), recall (0.61), and F1-score (0.61), followed by Random Forest (accuracy = 0.59), Logistic Regression (accuracy = 0.59), and AdaBoost (accuracy = 0.59). For SATTEACH_class, Random Forest led with accuracy (0.59), followed closely by XGBoost (0.58), ANN (0.57), and AdaBoost (0.56). Key predictors of teacher job satisfaction included workload-related variables and school-environment factors, which consistently emerged as the most important features across the best-performing models. The methodology and open-source pipeline provide a reproducible framework for evidence-based interventions to improve teacher retention and instructional quality, offering valuable insights for policymakers and educational administrators. Full article

(This article belongs to the Special Issue AI Technology-Enhanced Learning and Teaching)

► Show Figures

Figure 1

25 pages, 845 KB

Open AccessArticle

AI Museum Guides Acceptance for History Learning: Design Attributes, Dual Affective Pathways, and Largely Invariant Gender Effects

by Li Wang, Xuezhen Wu, Yifan Zhuo, Chaohui Wang and Gang Ren

Information 2026, 17(4), 376; https://doi.org/10.3390/info17040376 - 17 Apr 2026

Viewed by 634

Abstract

As AI-powered learning tools become more common in educational settings, understanding their acceptance mechanisms is increasingly important. This study examines how the design attributes of AI museum guides—anthropomorphism, interactivity, and personalization—are associated with the acceptance intention and perceived learning outcomes among Chinese high [...] Read more.

As AI-powered learning tools become more common in educational settings, understanding their acceptance mechanisms is increasingly important. This study examines how the design attributes of AI museum guides—anthropomorphism, interactivity, and personalization—are associated with the acceptance intention and perceived learning outcomes among Chinese high school students with prior museum experience. Using structural equation modeling with 324 participants, we test whether these features relate to acceptance through two affective pathways: perceived warmth and anxiety reduction. The results reveal distinct patterns: anthropomorphism shows an indirect-only association with anxiety reduction through perceived warmth; interactivity is associated with anxiety reduction through responsive feedback; and personalization serves dual functions, enhancing both pathways. Anxiety reduction shows strong positive associations with both acceptance intention and perceived learning outcomes. The multi-group analysis shows that most pathways function equivalently across genders, with one exception where anxiety reduction more strongly predicts learning outcomes for females than males. These findings reveal distinct psychological functions within the Chinese educational context: anthropomorphism influences anxiety reduction exclusively through perceived warmth, while personalization and interactivity provide both affective and cognitive support. The implications for AI museum guide design in similar contexts are discussed. The generalizability to other cultural contexts and populations, such as Western students or adult learners, requires further investigation. Full article

► Show Figures

Figure 1

23 pages, 516 KB

Open AccessArticle

Edge-Centric Federated Subgraph Isomorphism Counting via Residual Graph Neural Networks

by Jianjun Shi, Qinglong Wu and Xinming Zhang

Information 2026, 17(4), 375; https://doi.org/10.3390/info17040375 - 16 Apr 2026

Viewed by 457

Abstract

Subgraph isomorphism counting is a fundamental yet computationally challenging task in graph analysis, with broad applications in bioinformatics and social network mining. With the tightening of data privacy regulations and the emergence of data silos, traditional centralized Graph Neural Network (GNN) approaches face [...] Read more.

Subgraph isomorphism counting is a fundamental yet computationally challenging task in graph analysis, with broad applications in bioinformatics and social network mining. With the tightening of data privacy regulations and the emergence of data silos, traditional centralized Graph Neural Network (GNN) approaches face significant deployment hurdles. Existing federated subgraph counting methods are primarily designed for database federation scenarios, focusing on exact queries and the privacy and security concerns of databases. However, this rigid focus on exactness and heavy cryptographic security severely limits their scalability and generalizability to complex, arbitrary query patterns. To bridge this gap, we propose a general Federated Edge-Centric Framework for Subgraph Isomorphism Counting (FedCount), shifting the paradigm from exact querying on federated databases to neural approximate counting under federated architectures. Rather than relying on heavy cryptographic techniques, we exclusively leverage the inherent structural isolation of federated learning as a lightweight empirical privacy measure. While this framework does not theoretically defend against advanced gradient-based inference attacks, it successfully prevents the direct leakage of raw graph topology and node features, achieving high-precision approximate counting without the prohibitive cryptographic overheads. Specifically, we introduce two key technical innovations to enhance local counting capability: (1) we integrate a provable edge encoding scheme into the interaction-based GNN architecture, explicitly modeling edge-to-edge interactions to break the expressiveness bottleneck of standard message passing; (2) we design a Residual Edge-Centric Readout mechanism that mitigates the gradient vanishing problem, enabling the effective training of deeper networks to capture high-order topological dependencies. Extensive experiments on multiple benchmark datasets demonstrate that our framework significantly outperforms existing distributed enumeration baselines in terms of generalization and efficiency, approaching the performance of centralized state-of-the-art models. Full article

(This article belongs to the Special Issue Graph Learning and Graph Neural Networks: Techniques and Applications)

► Show Figures

Figure 1

23 pages, 3485 KB

Open AccessArticle

Physical Key Extraction in Galvanic Coupling Communications: Reliability and Security Analysis

by Giacomo Borghini, Stefano Caputo, Anna Vizziello, Pietro Savazzi, Antonio Coviello, Maurizio Magarini, Sara Jayousi and Lorenzo Mucchi

Information 2026, 17(4), 374; https://doi.org/10.3390/info17040374 - 16 Apr 2026

Viewed by 359

Abstract

The evolution toward sixth-generation (6G) networks envisions humans as active nodes within a fully interconnected digital ecosystem, supported by data collected from in-body and on-body sensors. Since many of these devices are not equipped to connect directly to 6G networks, Wireless Body Area [...] Read more.

The evolution toward sixth-generation (6G) networks envisions humans as active nodes within a fully interconnected digital ecosystem, supported by data collected from in-body and on-body sensors. Since many of these devices are not equipped to connect directly to 6G networks, Wireless Body Area Networks (WBANs) serve as an essential intermediate layer. However, conventional radio-frequency technologies face limitations in terms of energy efficiency, security, and data integrity, motivating the adoption of lightweight security mechanisms. Physical Layer Security (PLS), and in particular Physical Key Extraction (PKE), offers a promising solution by enabling legitimate devices to derive shared cryptographic keys from the reciprocal properties of the communication channel. Galvanic coupling (GC) communication has recently emerged as an on-body transmission technology alternative to radio-frequency (RF), which exploits low-power electrical signals propagating through biological tissue. Building on prior feasibility studies, this work proposes a PKE framework tailored to GC channels, integrating a lightweight key reconciliation method, based on Hamming (7,4) error-correction codes, and evaluating system performance through dedicated reliability and security Key Performance Indicators (KPIs). Results reveal a trade-off shaped by electrode placement and channel quantization parameters. Among the ones tested, the optimal configuration is achieved with a 3 cm transverse inter-electrode spacing at both transmitter and receiver, and a 3 cm longitudinal separation between transmitter and receiver, by quantizing the channel impulse response with two quantization bits. While this work focuses on validating the method in controlled conditions in order to establish a reliable study framework, future developments will focus on enhanced reconciliation, privacy amplification, and analysis of the GC channel considering physiological and environmental variations. Full article

(This article belongs to the Special Issue Advances in Wireless Communications Systems, 3rd Edition)

► Show Figures

Figure 1

28 pages, 6084 KB

Open AccessArticle

Symmetric Cross-Entropy: A Novel Multi-Level Thresholding Method and Comprehensive Study of Entropy for High-Precision Arctic Ecosystem Segmentation

by Thaweesak Trongtirakul, Sos S. Agaian, Sheli Sinha Chauhuri, Khalifa Djemal and Amir A. Feiz

Information 2026, 17(4), 373; https://doi.org/10.3390/info17040373 - 16 Apr 2026

Viewed by 527

Abstract

Arctic sea ice is a critical indicator of global climate dynamics, directly influencing maritime navigation, polar biodiversity, and offshore engineering safety. The precise mapping of diverse ice types, such as frazil ice, slush, melt ponds, and open water, is essential for environmental monitoring; [...] Read more.

Arctic sea ice is a critical indicator of global climate dynamics, directly influencing maritime navigation, polar biodiversity, and offshore engineering safety. The precise mapping of diverse ice types, such as frazil ice, slush, melt ponds, and open water, is essential for environmental monitoring; however, it remains a formidable challenge in satellite remote sensing. These difficulties arise from low-contrast imagery, overlapping spectral signatures, and the subtle textural nuances characteristic of polar regions. Traditional entropy-based thresholding techniques often falter when segmenting these complex scenes, as they typically rely on Gaussian distribution assumptions that do not align with the stochastic nature of Arctic data. To address these limitations, this paper presents a novel unsupervised segmentation framework based on symmetric cross-entropy (SCE). Unlike standard directional measures, SCE provides a more robust objective function for multi-level thresholding by simultaneously maximizing intra-class cohesion and minimizing inter-class ambiguity. The proposed method uses an optimized search strategy to identify intensity levels that best delineate complex Arctic features. We conducted an extensive entropy-based comparative study that benchmarked SCE against 25 state-of-the-art entropy measures, including Shannon, Kapur, Rényi, Tsallis, and Masi entropies. Our experimental results demonstrate that the SCE method: (i) achieves superior accuracy by consistently outperforming established models in segmentation precision and boundary definition; (ii) provides visual clarity by producing segments with significantly reduced noise, making them ideal for identifying small-scale melt ponds and slush zones; and (iii) demonstrates computational robustness by providing stable threshold values even in datasets with non-Gaussian class distributions and poor illumination. Ultimately, these improvements deliver high-quality ice feature data that enhance risk assessment, operational planning, and predictive modeling. This research marks a major step forward in Arctic sea studies and introduces a valuable new tool for wider image processing and computer vision communities. Full article

(This article belongs to the Section Information Systems)

► Show Figures

Figure 1

29 pages, 1651 KB

Open AccessArticle

TSQA: Integrating Text Summarization and Question Answering to Improve Information Retrieval from Documents Using Retrieval-Augmented Generation

by Ahmed Sami Jaddoa, Jaber Karimpour and Pedram Salehpour

Information 2026, 17(4), 372; https://doi.org/10.3390/info17040372 - 15 Apr 2026

Viewed by 617

Abstract

Here, we propose a composite system that uses text summarization (TS) and question answering (QA) to supplement the IR process of long documents. Most previous studies have used separate approaches, i.e., either TS or QA. The aim of this paper is to develop [...] Read more.

Here, we propose a composite system that uses text summarization (TS) and question answering (QA) to supplement the IR process of long documents. Most previous studies have used separate approaches, i.e., either TS or QA. The aim of this paper is to develop an interaction between TS and QA in three stages to enhance IR performance. First, SBERT is used for summarization. Second, an RAG method is employed to retrieve information and generate answers. In the architecture of RAG, retrieval of the document is fulfilled via all-MiniLM-L6-v2, while answer generation is performed via the T5 and BART-large-cnn models. Third, the retrieved answers are assessed and compared with a baseline system in which the documents are treated without summarization. The proposed system aims to improve the quality of retrieved information and accuracy of answers generated by TSQA in a unified pipeline. Experimental evaluation conducted on the NIPS dataset demonstrates that the proposed approach significantly enhances summary informativeness and answer accuracy compared with traditional single-task approaches. The simulation results show improvements of 20.83% in text similarity and 2.38% in BERT scores for answer generation compared with the standard RAG baseline without summarization. Full article

(This article belongs to the Section Artificial Intelligence)

► Show Figures

Figure 1

37 pages, 570 KB

Open AccessReview

Autonomous Supply Chains: Integrating Artificial Intelligence, Digital Twins, and Predictive Analytics for Intelligent Decision Systems

by Mohammad Shamsuddoha, Honey Zimmerman, Tasnuba Nasir and Md Najmus Sakib

Information 2026, 17(4), 371; https://doi.org/10.3390/info17040371 - 15 Apr 2026

Viewed by 1924

Abstract

Autonomous supply chains (ASC) are the next generation of digitally empowered logistics and operations systems that can make adaptive, data-driven, and intelligent decisions. Innovations in artificial intelligence (AI), digital twins (DT), and predictive analytics (PA) are transforming traditional supply chains into integrated and [...] Read more.

Autonomous supply chains (ASC) are the next generation of digitally empowered logistics and operations systems that can make adaptive, data-driven, and intelligent decisions. Innovations in artificial intelligence (AI), digital twins (DT), and predictive analytics (PA) are transforming traditional supply chains into integrated and interactive networks to detect disruptions, simulate the future, and automatically modify operational decisions. This paper reviews the ASC mechanism and summarizes the increasing literature on the technologies and analytical capabilities available to support intelligent supply chain decision systems. A structured literature review was conducted using Scopus, Web of Science, and Google Scholar, resulting in 52 relevant studies after screening and eligibility assessment. The paper discusses the recent advances in AI-based forecasting, simulation environments using digital twins, data integration using the Internet of Things (IoT), and predictive analytics. These technologies can help an organization gain real-time visibility of the supply chain networks. They improve the precision of demand forecasting, optimize inventory and production planning, and dynamically coordinate logistics operations. Digital twins allow the development of virtual models of supply chain ecosystems, which could be used to test scenarios, analyze risks, and plan strategies. These capabilities combined can be used to create predictive and self-adaptive supply networks capable of being responsive to uncertainty and market volatility. Besides examining the technological foundations, the paper also tracks key challenges related to the move towards autonomous supply chains, such as data governance, system interoperability, cybersecurity risks, algorithm transparency, and the necessity of successful human-AI collaboration in decision-making. The synthesis leads to a multi-layered framework that integrates data acquisition, analytics, simulation, and execution for autonomous decision-making in supply chains. Future research directions in relation to resilient supply networks, intelligent automation, and adaptive supply chain ecosystems are also provided in the study. Through integrating existing information on the new forms of intelligent technology and how it can be incorporated into the supply chain systems, this review contributes to the literature on next-generation supply chains. It will also offer information to both researchers and practitioners aiming at designing autonomous as well as data-driven supply networks. Full article

(This article belongs to the Special Issue Artificial Intelligence in Sustainable Supply Chains: Innovations, Applications, and Future Directions)

► Show Figures

Figure 1

17 pages, 434 KB

Open AccessArticle

MACD: Multi-Agent Collaborative Approach for Cybersecurity Defense Strategy Generation

by Nanfang Li, Xiang Li, Zongrong Li, Denghui Ma, Lijun Yan, Haishan Cao, Wenqian Zhang, Xu Wang and Yu Liu

Information 2026, 17(4), 370; https://doi.org/10.3390/info17040370 - 15 Apr 2026

Viewed by 703

Abstract

Cybersecurity defense strategy generation transforms threat intelligence into actionable defense measures against sophisticated multi-stage cyberattacks. Existing approaches lack multi-dimensional coordination of technical, tactical, and threat actor expertise, with limited benchmarks for evaluating defense strategy quality. To address these gaps, we introduce MACD (Multi-Agent [...] Read more.

Cybersecurity defense strategy generation transforms threat intelligence into actionable defense measures against sophisticated multi-stage cyberattacks. Existing approaches lack multi-dimensional coordination of technical, tactical, and threat actor expertise, with limited benchmarks for evaluating defense strategy quality. To address these gaps, we introduce MACD (Multi-Agent Collaborative Defense), a novel framework that orchestrates specialized AI agents to generate ATT&CK-aligned defense strategies. MACD deploys three expert agents for technical defense, kill chain phase analysis, and APT profiling, coordinated through a synthesizing agent, while leveraging retrieval-augmented generation to mitigate hallucination risks in threat mapping. Additionally, we construct CyberDefBench, a comprehensive benchmark combining real-world APT cases and synthetic scenarios with dual-layer annotations for reactive and proactive defenses. Experimental results demonstrate that MACD achieves 84.6% technique mapping accuracy and 72.3% defense coverage, significantly outperforming baseline methods and validating the effectiveness of multi-agent collaboration for cybersecurity defense. Full article

(This article belongs to the Section Artificial Intelligence)

► Show Figures

Figure 1

28 pages, 1855 KB

Open AccessSystematic Review

AI-Powered Knowledge Management Systems Across Industries: A Systematic Review of Applications, Implementation Barriers, and Ethical Challenges

by Edmund Evangelista and Ghazala Rizvi

Information 2026, 17(4), 369; https://doi.org/10.3390/info17040369 - 14 Apr 2026

Viewed by 2054

Abstract

This systematic literature review (SLR) evaluates the existing literature on the benefits, implementation challenges, and ethical concerns associated with Artificial Intelligence (AI)-driven Knowledge Management Systems (KMS) across industries. The SLR followed PRISMA guidelines to identify studies from Scopus, Web of Science, JSTOR, and [...] Read more.

This systematic literature review (SLR) evaluates the existing literature on the benefits, implementation challenges, and ethical concerns associated with Artificial Intelligence (AI)-driven Knowledge Management Systems (KMS) across industries. The SLR followed PRISMA guidelines to identify studies from Scopus, Web of Science, JSTOR, and Google Scholar, using inclusion and exclusion criteria. Critical Appraisal Skills Programme (CASP) checklists were used to assess methodological quality and risk of bias in the included studies, and a structured narrative synthesis was employed to synthesize the findings. The review of 21 articles reveals benefits like improved knowledge capture and creation, storage, retrieval, personalization, and efficient dissemination, which lead to effective decision-making and performance improvements. The implementation barriers are categorized as organizational, technological, ethical, and financial, which generate a lack of trust, inability to manage, lack of interoperability, and monetary constraints. These barriers can be overcome by adopting Kotters’ Eight Stage Change Model, developing interoperability frameworks, evolving ethics benchmarks and standard guidelines for governance, and using viability analyses that incorporate both financial and non-financial considerations. In addition to bridging the gap between AI and KMS theories, the paper also provides practical and actionable insights about managing implementation and governance challenges. Full article

(This article belongs to the Section Artificial Intelligence)

► Show Figures

Figure 1

14 pages, 579 KB

Open AccessArticle

Wearable Sensor-Free Adult Physical Activity Monitoring Using Smartphone IMU Signals: Cross-Subject Deep Learning with Window-Length and Sensor Modality Studies

by Mussa Turdalyuly, Ay Zholdassova, Tolganay Turdalykyzy and Aydin Doshybekov

Information 2026, 17(4), 368; https://doi.org/10.3390/info17040368 - 14 Apr 2026

Viewed by 799

Abstract

Human activity recognition (HAR) using inertial sensors is essential for health monitoring and wellness applications, yet robust classification in real-world adult scenarios remains challenging due to subject variability and activity transitions in smartphone sensing environments. This study investigated smartphone-based physical activity recognition using [...] Read more.

Human activity recognition (HAR) using inertial sensors is essential for health monitoring and wellness applications, yet robust classification in real-world adult scenarios remains challenging due to subject variability and activity transitions in smartphone sensing environments. This study investigated smartphone-based physical activity recognition using accelerometer and gyroscope signals under a cross-subject evaluation protocol. To reduce label ambiguity and improve generalization, the original activity set was grouped into a reduced 6-class taxonomy. We evaluated lightweight deep learning models, including a smartphone-only convolutional neural network (CNN) and a multimodal fusion model combining smartphone and smartwatch signals. Using GroupKFold cross-subject validation, the smartphone-only CNN achieved competitive performance with Macro-F1 ≈ 0.46, while multimodal fusion did not provide consistent improvements. We also examined temporal segmentation and showed that shorter windows (2.0 s) yield better results than longer windows. Sensor ablation confirmed the importance of gyroscope information, and per-class analysis indicated that dynamic activities could be recognized reliably, whereas stairs and static categories remained difficult. Overall, the results demonstrate the practicality of smartphone-based activity recognition using built-in smartphone sensors without external wearable devices for adult activity monitoring and provide recommendations for window length and sensor selection in cross-subject HAR. Full article

► Show Figures

Figure 1

19 pages, 1982 KB

Open AccessArticle

Mapping Research Trends with the CoLiRa Framework: A Computational Review of Semantic Enrichment of Tabular Data

by Luis Omar Colombo-Mendoza, Julieta del Carmen Villalobos-Espinosa, María Elisa Espinosa-Valdés and Elías Beltrán-Naturi

Information 2026, 17(4), 367; https://doi.org/10.3390/info17040367 - 14 Apr 2026

Viewed by 768

Abstract

This article introduces the CoLiRa (Computational Literature Review & Analysis) framework, a novel integration of established computational algorithms designed to quantitatively analyze and map the evolution of scientific fields. Employing a human-in-the-loop epistemological approach, CoLiRa combines the scalability of automated algorithms with the [...] Read more.

This article introduces the CoLiRa (Computational Literature Review & Analysis) framework, a novel integration of established computational algorithms designed to quantitatively analyze and map the evolution of scientific fields. Employing a human-in-the-loop epistemological approach, CoLiRa combines the scalability of automated algorithms with the semantic coherence of expert-driven qualitative research. The multi-stage pipeline incorporates Latent Dirichlet Allocation (LDA) for thematic discovery, cluster analysis (K-Means and Multidimensional Scaling) for conceptual mapping, and Ordinary Least Squares (OLS) regression to monitor temporal trends. Algorithmic outputs are structurally validated by domain experts using quantitative metrics. The framework’s end-to-end capabilities are demonstrated through a proof-of-concept case study on the semantic enrichment of tabular data, encompassing studies up to 2024 that utilize Semantic Web ontologies, Linked Data, and knowledge graphs. The analysis identifies three core research topics and finds no statistically significant linear trends, suggesting thematic coexistence. This work provides a validated, hybrid computational approach for conducting robust literature reviews and mapping research trajectories. Full article

(This article belongs to the Special Issue Advances in Information Studies)

► Show Figures

Figure 1

Journal Menu

Journal Browser

Information, Volume 17, Issue 4 (April 2026) – 86 articles

Further Information

Guidelines

MDPI Initiatives

Follow MDPI