Search Results (5,646)

Advanced oxidation processes (AOPs) are widely applied to degrade recalcitrant organic contaminants in municipal effluents, industrial wastewaters, and water-reuse streams. Their deployment, however, remains constrained by matrix scavenging, high energy or reagent demand, catalyst/electrode ageing, and the possible formation of toxic transformation products. Artificial intelligence and machine learning (AI/ML) have been proposed as tools for prediction, optimization, catalyst discovery, mechanism inference, and process control, but high accuracy on curated laboratory datasets is often confused with actionable knowledge for real treatment systems. This narrative review evaluates AI/ML-enabled AOPs through an evidence-to-deployment framework built on three principles: real wastewater is treated as the primary inference domain; mechanistic claims are graded according to convergent evidence; and AI/ML contributions are linked to explicit decisions rather than to model accuracy alone. We argue that progress depends less on black-box complexity than on standardized reporting, benchmark matrices, curated datasets, uncertainty-aware validation, and pilot-scale demonstrations that satisfy contaminant removal, energy efficiency, byproduct safety, and operational constraints simultaneously. A six-gate decision framework and a targeted research agenda are proposed to guide future studies toward deployment-grade evidence. Full article

(1) Background: Artificial intelligence tools have become integrated into undergraduate students from academic assignments to seek help with psychological concerns, particularly during the crises period. Scales measuring Artificial Intelligence-help-seeking behavior (AI-HSB) are still limited. This study aims to develop a new bilingual scale (Arabic and English) to assess AI-HSB by providing a reliable and useful tool for researchers worldwide. (2) Methods: We conducted a methodological cross-sectional design among 416 undergraduate students in United Arab Emirates (AUE) between the period of 1 October 2025 and 10 December 2025, using an online Google Form. The development, translation, validation, and reliability processes were conducted for the AI-HSB scale. (3) Results: It has been found that 13 items (two factors) are strong indications of factorial validity, reliability, and construct validity of AI-HSB scale. The two factors explained about 58% of the total variance. The confirmatory factor analysis confirmed the two-factor structure with all items loading above recommended thresholds and the goodness-of-fit indices of AI-HSB all exceeded 0.90. (4) Conclusions: The AI-HSB is a valid and reliable tool for assessing AI-based psychological help-seeking behavior among university students in the UAE. This scale will allow universities, counselors, and policymakers to use a well-validated scale to measure the extent to which students are using AI for psychological coping. Full article

Advanced cyber threats increasingly exploit volatile memory to execute malicious payloads without touching persistent storage, rendering traditional disk-centric forensic tools insufficient for comprehensive digital investigations. This paper presents Forensic-BERT, an AI-driven forensic framework that automatically extracts and classifies potentially relevant artifacts from unstructured memory dumps across heterogeneous operating environments. The framework combines byte-boundary-preserving Hex-to-ASCII conversion, sliding-window Shannon entropy filtering ($H>7.2$ bits per byte, 256-byte windows) to isolate high-probability artifact regions, and a binary-aware WordPiece tokenizer extended with 2048 domain-specific tokens covering hexadecimal byte patterns, Windows API names, and Linux system-call sequences. These components feed a transformer-based classifier fine-tuned from bert-base-uncased (110 M parameters) on memory-derived text, with sliding-window inference and majority-vote aggregation for large images. A SHAP DeepExplainer module and averaged 12-head attention heatmaps provide transparent, analyst-accessible explanations for classification decisions. We evaluate the framework on a multi-source corpus of 735 labeled memory segments drawn from 197 distinct images across four independent collections, MemLabs, the DARPA Transparent Computing program, Digital Corpora, and live sandbox execution traces from Any.run and Joe Sandbox, spanning Windows XP through Windows 11, Ubuntu Linux 16.04/18.04, and FreeBSD. Source-stratified five-fold cross-validation yields an overall F1-score of $0.92\pm 0.02$ and AUC-ROC of $0.95\pm 0.01$ (95% CI). Forensic-BERT outperforms all six baselines, Volatility with YARA rules (F1 $=0.71$), Random Forest (F1 $=0.82$), BiLSTM with GloVe embeddings (F1 $=0.85$), MRm-DLDet (F1 $=0.87$), SPECTRE (F1 $=0.89$), and SecBERT (F1 $=0.90$), with every pairwise difference statistically significant under the McNemar test with Bonferroni correction. Explainability quality is independently confirmed by a Spearman rank correlation of $\rho =0.81$ between model SHAP token rankings and expert forensic-indicator rankings and by a System Usability Scale score of 73.2 among certified examiners. The complete pipeline processes 512 MB memory images in 7.5–10.2 s (GPU) or 38–52 s (CPU-only), scaling to 4 GB images with near-linear throughput. These results indicate that, on the corpus evaluated here, combining domain-adapted NLP preprocessing, transformer-based sequence modeling, and quantified explainability can improve the effectiveness and usability of analyst decision support and investigative triage for volatile memory analysis. Full article

The proliferation of generative artificial intelligence (GenAI) has precipitated an urgent, global reassessment of how higher education evaluates critical thinking, creative agency, and academic integrity. However, scholarly and institutional responses remain fragmented across cultural contexts, impeding the development of robust, flexible, and discipline-adaptable assessment frameworks. Responding to the imperative to move beyond the traditional standardized assessment paradigm, this study conducts a comparative discourse analysis of 5368 academic articles in Anglophone/Western scholarly discourse (Web of Science, WoS) and Chinese (CNKI). Using LDA topic modeling and Word2Vec-enhanced semantic analysis, the study identifies two divergent orientations: an Anglophone/Western discourse that frames AI as an instrument for cognitive augmentation, efficiency optimization, and functional human–AI collaboration; and a Chinese discourse that emphasizes epistemic sovereignty, the reconstruction of creative subjectivity, and systemic institutional rebuilding against technological alienation. These pathways are mapped onto a tripartite framework of Tools, Creative Subjectivity, and Organizational Ecosystems. The findings demonstrate that AI integration is culturally embedded rather than technically determined, carrying profound implications for assessment validity, academic integrity policy, and equitable access to AI-enhanced learning. The study synthesizes these insights into a culturally responsive assessment framework that redirects evaluation from standardized, product-centric outputs toward process-oriented, transparent, and ethically governed human–AI co-authorship. By centering critical autonomy, AI literacy, and epistemological diversity, this framework offers actionable strategies for inclusive assessment redesign, institutional policy development, and sustainable competency cultivation in the GenAI era. Full article

Background: Adverse drug reactions (ADRs) represent a major and often underestimated source of morbidity, hospitalization, and functional decline in older adults. The convergence of age-related pharmacokinetic and pharmacodynamic changes, multimorbidity, polypharmacy, and frailty creates a clinical environment in which ADR risk is not static but evolves along progressive trajectories—from mild, early manifestations toward severe, potentially irreversible outcomes. Understanding these trajectories is essential for rational geriatric prescribing. Methods: This narrative review synthesizes evidence from epidemiological studies, systematic reviews, Cochrane analyses, and clinical trials published between 2000 and 2025, focusing on adults aged 65 years and older with two or more chronic conditions. Sources were identified through a structured, non-systematic literature search of PubMed, EMBASE, Cochrane Library, Web of Science, and Scopus using the terms ‘adverse drug reactions’, ‘polypharmacy’, ‘multimorbidity’, ‘frailty’, ‘deprescribing’, and ‘pharmacokinetics’ in older adults, alone and in combination. Evidence quality was assessed narratively, distinguishing trial evidence from observational and expert consensus data. Results: ADRs in older adults are best classified using complementary frameworks—the augmented Type A to withdrawal Type E and failure-of-therapy Type F taxonomy (Types A–F), the Dose-Time-Susceptibility (DoTS) classification, and the EIDOS mechanistic scheme—which together capture the heterogeneity of drug-related harm in this population. Age-related pharmacokinetic changes (altered absorption, increased volume of distribution of lipophilic drugs, reduced hepatic and renal clearance) and pharmacodynamic shifts (heightened receptor sensitivity, baroreflex impairment, increased blood–brain barrier permeability) interact with polypharmacy and frailty to amplify ADR trajectories from mild to severe. Anticholinergic burden, prescribing cascades, and inappropriate polypharmacy function as structural accelerators of these trajectories. Medication review and deprescribing improve prescribing quality but evidence for hard outcome benefits remains of low to very low certainty. Emerging AI-enabled digital tools show promising accuracy for identifying frailty and pharmacological vulnerability, but this performance relates to frailty classification and has not yet been shown to prevent ADR trajectories; they require validation for routine clinical use. Conclusions: Recognizing ADRs in older adults as dynamic trajectories rather than isolated events repositions prescribing review and deprescribing from optional to essential clinical acts. An integrated approach combining pharmacological vigilance, comprehensive geriatric assessment, structured deprescribing, and emerging digital decision-support tools offers the most realistic pathway to reduce the trajectory-related burden of drug-related harm in complex older patients. Full article

Stroke remains a leading cause of death and long-term disability worldwide, making prevention strategies a global health priority. Emerging technologies—including artificial intelligence (AI), wearable devices, digital health applications, and drone-assisted emergency systems—are increasingly being explored to improve stroke prevention and early management. In primary prevention, machine learning models can identify individuals at high risk of stroke using clinical and behavioral data with high reported predictive accuracy, although most models are derived from retrospective, single-center datasets and still require prospective external validation. Digital devices and wearable technologies enable continuous monitoring of cardiovascular risk factors and support behavioral interventions aimed at reducing vascular risk. In secondary prevention, AI-based tools are being developed to predict stroke recurrence, identify modifiable risk factors, and detect patients at risk of poor medication adherence. In the acute setting, AI-assisted neuroimaging platforms are already integrated into clinical and telestroke workflows, supporting rapid triage and treatment decisions. In parallel, drone-based emergency systems may contribute to improved outcomes by reducing prehospital delays and facilitating telemedicine-based triage in remote or resource-limited settings, although current evidence is derived largely from out-of-hospital cardiac arrest pathways rather than stroke-specific trials. Although advanced neurotechnological systems capable of real-time neurophysiological monitoring and closed-loop neuromodulation exist in other neurological disorders, their role in stroke prevention remains largely theoretical. Overall, these technologies offer promising opportunities to reshape the continuum of stroke prevention and care, but further validation, integration into clinical workflows, and evidence of real-world effectiveness are required before widespread implementation. Full article

The evolution of digital technologies is reshaping parliaments worldwide, driving fundamental changes in their operations. Parliaments, being traditionally conservative institutions, typically lean toward “mature” emerging or disruptive technologies through cautious, incremental digital transformation attempts, resulting in complex digital parliamentary environments for their users, processes, systems, and tools. The paper employs an integrative literature review as its methodological tool, examining the concept of the “digital parliament” and the technologies that enable it. Using a PRISMA-informed methodology as a guide, we conducted an integrative review covering the period 2006–2025, and in this context, we retrieved 535 publications, screened 260, thoroughly examined 57, and analyzed and synthesized 34 studies addressing digital parliamentary technologies, digital platforms, and cooperative workspaces. We found that while specific parliamentary technology (ParlTech) applications—including big data analytics, artificial intelligence (AI), and hybrid parliamentary tools—are reaching institutional maturity, the concept of a digital parliament remains fragmented, lacking a unified definitional and operational framework. Key research gaps have been identified concerning user classification, the digitization of parliamentary functions, operations, and processes, as well as the institutionalization of cooperation platforms. Based on these findings, we propose strategic directions toward establishing a responsible, inclusive, and evidence-based digital parliament. This research contributes as a guideline for parliamentary organizations seeking to create, retain, and disseminate public value through the responsible adoption of emerging digital technologies. Full article

Background/Objectives: Accurate staging of mediastinal lymph nodes is essential for therapeutic decisions and prognostic assessment in non-small cell lung cancer (NSCLC). This systematic review evaluates diagnostic performance, validation strategies, and clinical significance of artificial intelligence (AI)-based 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET/CT) radiomics models for mediastinal nodal staging in NSCLC. Methods: Systematic literature searching was conducted in PubMed, ScienceDirect, and Scopus according to the PRISMA 2020 guidelines. Eligible studies used radiomic or AI-based approaches for mediastinal lymph node (LN) evaluation in NSCLC, with histopathology as a reference standard. Extracted data included study design, cohort characteristics, imaging method, validation strategy, and diagnostic performance metrics. Methodological quality was assessed by the QUADAS-2 tool. Results: Thirteen studies were included which are mainly retrospective in their designs with cohort sizes varying between 87 and 3265 patients. Models evaluated on external or prospective validation cohorts generally showed lower performance compared with training or internal datasets. However, clinically significant discriminative ability has been preserved across heterogeneous populations. In studies that directly compared methods, composite models integrating radiomic features with clinical factors and conventional PET metrics, sometimes including deep learning-derived features, consistently outperformed radiomics-only models. Additionally, selected approaches addressing FDG-related false-positive uptake improved distinction between benign and metastatic mediastinal lymph nodes; this is reflected by reduced false-positive classifications plus higher specificity compared with conventional PET/CT interpretation. Conclusions: AI-based 18F-FDG PET/CT radiomics show a promising discriminative capacity for mediastinal nodal staging in NSCLC, especially when it is integrated with clinical and conventional imaging variables. Although the model performance remains clinically significant within independent validation cohorts, attenuation compared with training datasets is commonly observed. Methodological heterogeneity, predominantly retrospective study designs, and the scarcity of prospective multicenter validation currently limit routine clinical implementation. Full article

Artificial intelligence (AI) is increasingly embedded in analytics-driven decision systems operating in high-stakes environments, where predictive performance alone may not ensure robustness, traceability, or readiness for governance. While explainable artificial intelligence (XAI) is commonly treated as an interpretability tool, its role as a measurable diagnostic component remains underexplored. This study evaluates explainability beyond predictive accuracy using a controlled dual-pipeline design with identical data, model, and validation settings. Using the UCI Default of Credit Card Clients dataset, an XGBoost model achieved strong predictive performance (AUC ≈ 0.78; AP ≈ 0.56) while maintaining high decision stability under retraining (agreement ≈ 99.49%). Global explanations were highly reproducible across runs (Spearman ρ ≈ 0.99), and entropy-based local explanation analysis revealed substantially higher attribution dispersion in false-negative cases (odds ratio ≈ 6.58), linking explanatory uncertainty to misclassification-prone regions. The findings demonstrate that explainability diagnostics can reveal measurable patterns of stability, reproducibility, and uncertainty that are not captured by predictive metrics alone. The study advances a governance-oriented perspective in which explainability serves as a reproducible diagnostic layer that supports monitoring, validation, and audit-oriented assessment in analytics-driven decision systems. Full article

Influenza viruses continue to pose a significant threat to both animal and public health due to their rapid evolution and the frequent emergence of antiviral resistance. Host-directed antiviral (HDA) strategies, which target host factors essential for viral replication, may represent an alternative to conventional virus-targeting approaches. However, the identification of reliable and therapeutically actionable host targets remains a major challenge, primarily due to the complexity and context dependency of host–virus interactions. Recent advancements in multi-omics technologies, including functional genomics, transcriptomics, and proteomics, have facilitated the systematic characterization of host factors involved in influenza virus infection. These methodologies have unveiled intricate regulatory networks that govern viral replication and host immune responses. Nonetheless, translating large-scale datasets into biologically meaningful targets necessitates robust integrative frameworks. In this context, artificial intelligence (AI) and machine learning methods offer powerful tools for data integration, target prioritization, and predictive modeling. In this Review, we summarize current insights into host factors that regulate influenza virus infection and discuss how multi-omics and AI-driven approaches are expediting host target discovery. Furthermore, we highlight the potential of these strategies to enhance antiviral development while addressing key challenges related to specificity, safety, and translational application. Collectively, these advancements lay a foundation that may support the rational design of next-generation host-directed antivirals. Full article

Background/Objectives: Novel vision-language models (VLMs) can integrate patient textual data with image data to support medical diagnosis. Recent studies reported conflicting results regarding the performance of multimodal VLMs compared to other models and physician performance. This systematic review aims to assess the diagnostic performance of multimodal VLMs integrating both patient textual and image data across diverse real-world hospital settings. Methods: We performed comprehensive searches of eight resources, including Embase, MEDLINE, and SCOPUS, on 17 December 2025. Eligible studies reporting diagnostic performance of VLMs integrating both image and patient history textual data from real-world adult patients compared to that of other models and physicians were included. The review adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The Prediction model study Risk Of Bias Assessment Tool + AI (PROBAST + AI) was used to assess the quality and risk of bias. The study protocol was registered in the PROSPERO database (CRD420251244054). This review received no external funding. Results: We screened 11,026 records, of which 18 studies met the inclusion criteria. Six studies comparing multimodal and unimodal models demonstrated the consistent superiority of the multimodal models. Four studies evaluating VLM accuracy as standalone agents compared with physician performance reported conflicting evidence. One study assessing VLMs as a clinical copilot demonstrated higher accuracy from the group of physicians using VLM assistance. A meta-analysis could not be performed due to the heterogeneity across study populations and outcomes. The majority of the studies were assessed as having a high risk of bias due to dataset quality. Primary limitations identified across studies include small sample size, a lack of external validation, and the need for prospective clinical deployment studies. No study provided documented considerations regarding model safety or data security. Conclusions: This systematic review suggests that multimodal VLMs consistently outperform unimodal models with access to only image or text. While model performance as standalone agents compared to humans remains inconclusive, a copilot model has demonstrated high diagnostic accuracy. Given substantial methodological concerns across studies, cautious interpretation is required, No firm clinical recommendation can be made regarding the use of standalone VLMs. Further research employing high-quality datasets is needed to ensure the reliability and clinical applicability of future VLMs. Full article

The global sugar industry is facing increasing challenges due to climate variability, sustainability requirements, and the need for improved operational efficiency. These pressures are driving the search for advanced technological solutions to enhance productivity and resource management. Artificial intelligence (AI) has already demonstrated significant potential across various agricultural sectors; however, a comprehensive evaluation of AI applications across the entire sugar industry value chain from crop cultivation to industrial processing and supply chain management remains limited. This review provides a detailed assessment of the current state of AI and internet of things (IoT) implementation in the sugar beet industry. It examines key applications, including precision agriculture for sugarcane and sugar beet cultivation, intelligent monitoring systems for early disease detection, and AI-driven decision support tools for resource optimization. In addition, the study explores the role of AI in sugar manufacturing processes, where machine learning and data-driven models are used to optimize milling operations, improve product quality control, and enable predictive maintenance of industrial equipment. AI technologies are also shown to enhance supply chain efficiency through improved demand forecasting, logistics optimization, and real-time data analytics. Monitoring volatile organic compounds (VOCs) is becoming increasingly important in sugar beet and sugarcane storage. Microbial activity during storage and fermentation can release VOCs such as ethanol, which act as early indicators of crop degradation and spoilage. Detecting these gases using modern gas sensors enables continuous monitoring of storage conditions and crop health. When sensor data is integrated with AI and IoT systems, it can be analyzed in real time to identify early signs of microbial activity, improve storage management, and optimize processing decisions. Such intelligent monitoring systems have the potential to reduce losses and enhance overall efficiency in the sugar production chain. Full article

Viral vaccine development requires both mechanistic understanding of protective immunity and translational study designs that connect preclinical data with human outcomes. Animal models remain important for early assessment of safety, immunogenicity and protective efficacy, but their predictive value depends on the question being asked, the pathophysiology of infection, the immune mechanisms expected to mediate protection, and the biomarkers chosen to bridge animal and human data. This review focuses on viral vaccines and examines innate and adaptive mechanisms of vaccine-induced protection, including B cell and antibody responses, Fc-mediated functions, Fc glycosylation, T cell memory and CD8+ cytotoxic responses. We discuss common reasons for clinical failure and show how preclinical endpoints can be classified as human-counterpart, surrogate or comparative/mechanistic readouts. Influenza and COVID-19 examples illustrate how different models can be combined across discovery, challenge, transmission and late-stage bridging studies. Emerging tools such as systems serology, omics, AI/ML and new approach methods can improve candidate prioritization, but their value depends on assay standardization, biological validation and cautious interpretation. A mechanism-driven model cascade, paired with human-relevant immunological readouts, can improve preclinical interpretation and reduce the risk of advancing candidates that are unlikely to succeed in clinical trials. Full article

This article examines how professionals who use generative AI in their daily work reshape their temporal experience. Drawing on 21 semi-structured interviews with experienced AI users and developers in Romania, and building on Flaherty’s concept of “time work”, it introduces the notion of delegated time work: a form of temporal agency in which individuals transfer part of the time-shaping effort to an AI tool while retaining judgment over the temporal structure of activity. The results show clear support for delegated time work in three dimensions of temporal experience: duration, sequence, and allocation. Evidence for frequency, timing, and taking time is limited: delegation succeeds in the dimensions professionals control individually and fails in those governed by shared institutional rhythms. Delegation also generates its own temporal costs through learning and verification overheads, unevenly distributed between developers and users. Drawing on the “time capital” framework of Matei and Preda, the analysis traces three outcomes of the freed time: accumulation as a personal resource, conversion into professional or economic capital, and absorption by rising expectations, leaving workers faster but not freer. In Romania’s dependent market economy, market exposure shapes who keeps the time that AI frees. Full article

The primary objective of this article is to examine the organizational, economic, and sustainability-related implications of implementing artificial intelligence (AI) systems in small and medium-sized enterprises (SMEs) in Poland. The study combines a survey of 112 SMEs in the Kuyavian–Pomeranian region, including 70 AI-using firms, with 13 in-depth interviews with managers. The quantitative analysis applies logit models to identify determinants of perceived AI effects on internal processes: working time and workload reduction, automation, cost effects, and creativity. The qualitative component explains how AI is adopted and embedded in business practice. The results show that AI adoption in SMEs is increasingly common but remains uneven and mostly operational. The strongest effects concern workload reduction and time efficiency, particularly in service firms and where AI is used intensively. Advanced AI adoption increases the probability of perceiving workload and cost-related effects. However, these effects should not be interpreted simply as direct cost reduction. Rather, AI improves productivity and work capacity while creating new costs related to paid tools, data preparation, integration, output verification, and governance. The interviews show that AI implementation follows a staged path: from curiosity-driven experimentation, through cognitive work augmentation, to workflow integration and, in selected cases, AI-enabled business model innovation. The transition from ad hoc use to strategic implementation depends less on firm size alone and more on process maturity, capabilities, and data readiness. Barriers also change with maturity: early-stage firms face a lack of knowledge, time, and clear use cases, whereas advanced users encounter data quality, hallucinations, security, integration, and governance problems. The study finds that sustainability considerations, particularly environmental impacts and ESG-related implications of AI, remain largely unperceived in SME decision-making. Entrepreneurs primarily interpret sustainability through the lenses of organizational resilience, long-term competitiveness, adaptability, and responsible digital transformation rather than through formal environmental metrics. The findings suggest that SME managers should implement AI gradually, link adoption to measurable process-level outcomes, and invest in AI literacy and governance. They should also integrate responsible AI principles into organizational strategy to support sustainable digital transformation. The study contributes to the literature by showing that AI adoption in SMEs should be understood not only as a productivity-enhancing process but also as a broader organizational transition shaping long-term sustainability and resilience. Full article