Public higher education systems face increasing financial pressures from expanding student populations, rising operational costs, and persistent demands for equitable access. Artificial Intelligence (AI), including generative tools such as ChatGPT, learning analytics, intelligent tutoring systems, and predictive models, has been proposed as a means of enhancing efficiency and reducing costs. This study conducts a scoping review of the literature on AI applications in public higher education, based on systematic searches in Scopus and IEEE Xplore that identified 241 records, of which 21 empirical studies met predefined eligibility criteria and were thematically analyzed. The findings show that AI enables cost savings by automating administrative tasks, optimizing resource allocation, supporting personalized learning at scale, and applying predictive analytics to improve student retention and institutional planning. At the same time, concerns emerge regarding implementation costs, unequal access across institutions, and risks of widening digital divides. Overall, the thematic analysis highlights both the promises and limitations of AI-driven cost reduction in higher education, offering insights for policymakers, university administrators, and educators on the economic implications of AI adoption, while also pointing to gaps that warrant further empirical research.

Record linkage holds a crucial position in data management and analysis by identifying and merging records from disparate data sets that pertain to the same real-world entity. As data volumes grow, the intricacies of record linkage amplify, presenting challenges, such as potential redundancies and computational complexities. This paper introduces DynaHash, a novel randomized record linkage mechanism that utilizes (a) the MinHash technique to generate compact representations of blocking keys and (b) Hamming Locality-Sensitive Hashing (LSH) to construct the blocking structure from these vectors. By employing these methods, DynaHash offers theoretical guarantees of accuracy and achieves sublinear runtime complexities, with appropriate parameter tuning. It comprises two key components: a persistent storage system for permanently storing the blocking structure to ensure complete results, and an in-memory component for generating very fast partial results by summarizing the persisted blocking structure. Additionally, DynaHash leverages Multi-Probe matching to scan multiple neighboring blocks, in terms of their Hamming distances, in order to find matches. Our theoretical work derives a decrease factor in the space requirements, which depends on the Hamming threshold, compared with the baseline LSH. Our experimental evaluation against three state-of-the-art methods on six real-world data sets demonstrates DynaHash's exceptional recall rates and query times, which are at least 2x faster than its competitors and do not depend on the size of the underlying data sets.

Deep Learning has revolutionized Entity Resolution (ER) by enabling high-accuracy matching through dense vector embeddings. However, this paradigm shift transforms the traditional blocking step into a high-dimensional similarity search problem, introducing a massive computational bottleneck that threatens scalability. While modern ER frameworks increasingly rely on blocking, the so-called indexing, to generate candidate pairs, the architectural design of this retrieval stage is often treated as a black box, with little attention paid to the resource implications of high-dimensional models. This paper conducts a rigorous design space exploration of blocking architectures, evaluating nine candidate generation strategies across graph-based, partition-based, and hash-based families. A primary contribution is a systematic analysis of Vector Quantization – specifically Product Quantization and Scalar Quantization – demonstrating that quantization is not merely a memory optimization, but a critical requirement for deploying state-of-the-art embeddings at scale. We evaluate these architectures on seven real-world datasets, measuring the interplay between embedding granularity, recall, and computational cost. Our results reveal that while graph-based methods offer peak precision, quantized partitioning methods reduce memory footprints by up to 96% with negligible loss in recall, effectively democratizing high-performance ER on commodity hardware. The study culminates in a decision framework for designing scalable, resource-aware blocking pipelines.

Privacy-preserving data mining (PPDM) seeks to extract useful patterns from shared data without revealing sensitive information. Within PPDM, knowledge hiding—encompassing both association rule hiding (ARH) and frequent itemset hiding (FIH)—forms a coherent family of techniques that sanitize transactional databases before release. This focused survey synthesizes the main algorithmic paradigms for knowledge hiding (1999–2026), covering heuristic sanitization, border-based and exact optimization via integer linear programming, constraint-based and graph-based formulations, emerging learning-guided support mechanisms, and extensions to utility mining and non-relational structures. We use a PRISMA-style search and selection protocol to make the evidence base transparent and to mitigate selection bias. We trace the evolution from early disclosure-limitation heuristics to graph-guided and knowledge-graph approaches, and we treat deep-learning, GNN, and federated graph-learning work as adjacent tools that may support candidate selection, representation learning, or distributed deployment rather than as replacements for classical hiding validation. We identify persistent challenges around scalability, infeasibility in LP formulations, and evaluation standardization, and outline directions for future research. Unlike broader PPDM overviews, this review centers exclusively on transactional knowledge hiding. Beyond cataloging algorithms, it compares method families through their intervention mechanisms, side-effect profiles, scalability assumptions, and benchmark regimes, and it distills reporting recommendations for more reproducible empirical evaluation.

Background/Objectives: Assessment of Internet Gaming Disorder (IGD) relies largely on retrospective self-report and clinical interviews, which may be affected by recall bias, social desirability bias, and limited sensitivity to within-person behavioral change. This study evaluated an AI-enabled, privacy-preserving digital phenotyping framework for personalized IGD risk stratification under controlled simulation assumptions. Methods: A reproducible synthetic dataset of 1,000 virtual user profiles was generated; 20% were assigned to an elevated-risk class, and 5% balanced stochastic label noise was introduced to approximate imperfect ground truth. Four aggregated telemetry features were modeled: average session duration, sessions per week, Late-Night Index, and application-switching rate. Random Forest, Logistic Regression, and Gradient Boosting classifiers were evaluated against playtime-only baselines using a stratified 80:20 train–test split. Results: In the primary Random Forest model, accuracy was 0.880, balanced accuracy was 0.850, sensitivity was 0.800, specificity was 0.900, area under the receiver operating characteristic curve (ROC-AUC) was 0.909, area under the precision–recall curve (PR-AUC) was 0.779, and the Brier score was 0.089. All-feature models substantially outperformed playtime-only baselines. Feature-importance analyses recovered the known signal hierarchy encoded in the synthetic data-generating process, with application-switching rate and Late-Night Index showing the largest Gini-based and permutation-importance values. Performance degraded progressively as label noise increased from 0% to 20%. Conclusions: The framework demonstrates the methodological feasibility of transforming aggregated, privacy-preserving behavioral telemetry into interpretable simulated risk signals for IGD. The findings are hypothesis-generating and do not establish clinical validity or diagnostic performance. Longitudinal validation in clinically characterized cohorts using validated psychometric instruments and person-level calibration is required before practical deployment.

We used an Item Response Theory (IRT) model to assess the overall quality and the characteristics of each question included in the final test of the 2023 thematic unit of the postgraduate “Master in Information Systems” (PLS60) of the Hellenic Open University. The findings derived from this analysis were used to enhance the examination for the next academic year (2024). A comparison of the test quality between the two consecutive years, revealed a significant improvement, but also the potential for developing even higher quality assessments that are fair, well-structured, and capable of accurately evaluating student performance.

Artificial Intelligence (AI) is increasingly incorporated into medical devices, revolutionizing diagnostics, treatment planning, and patient monitoring. To ensure AI's safe and ethical use, the European Commission published the AI Act in 2024, which places stringent obligations on AI systems, especially those classified as high-risk, such as medical devices. This paper evaluates the impact of the EU AI Act on existing regulations such as the Medical Device Regulation (MDR) and the In Vitro Diagnostic Regulation (IVDR). It explores challenges related to compliance, certification processes, and potential conflicts between the AI Act and existing medical device frameworks while providing recommendations for harmonization.

Optical Coherence Tomography (OCT) has become an indispensable imaging modality in ophthalmology, providing high-resolution cross-sectional images of the retina. Accurate classification of OCT images is crucial for diagnosing retinal diseases such as Age-related Macular Degeneration (AMD) and Diabetic Macular Edema (DME). This study explores the efficacy of various deep learning models, including convolutional neural networks (CNNs) and Vision Transformers (ViTs), in classifying OCT images. We also investigate the impact of integrating metadata (patient age, sex, eye laterality, and year) into the classification process, even when a significant portion of metadata is missing. Our results demonstrate that multimodal models leveraging both image and metadata inputs, such as the Multimodal ResNet18, can achieve competitive performance compared to image-only models, such as DenseNet121. Notably, DenseNet121 and Multimodal ResNet18 achieved the highest accuracy of 95.16%, with DenseNet121 showing a slightly higher F1-score of 0.9313. The multimodal ViT-based model also demonstrated promising results, achieving an accuracy of 93.22%, indicating the potential of Vision Transformers (ViTs) in medical image analysis, especially for handling complex multimodal data.

The integration of Artificial Intelligence (AI) into healthcare has the potential to revolutionize patient care, diagnostics, and treatment planning. However, this integration also introduces significant challenges related to data governance, privacy, and compliance with emerging regulations. The European Union's (EU) AI Act proposes a comprehensive regulatory framework aimed at ensuring that AI systems are trustworthy and respect fundamental rights. This paper provides an in-depth analysis of the data governance requirements stipulated by the EU AI Act specifically within the context of healthcare AI. Furthermore, it explores strategies for compliance, examines the interplay with existing regulations such as the General Data Protection Regulation (GDPR), and addresses the ethical considerations inherent in deploying AI in healthcare settings.

Applying language models (LMs) and generative artificial intelligence (GenAI) to the study of Ancient Greek offers promising opportunities. However, it faces substantial challenges due to the language’s morphological complexity and lack of annotated resources. Despite growing interest, no systematic overview of existing research currently exists. To address this gap, a systematic literature review was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 methodology. Twenty-seven peer-reviewed studies were identified and analyzed, focusing on application areas such as machine translation, morphological analysis, named entity recognition (NER), and emotion detection. The review reveals six key findings, highlighting both the technical advances and persistent limitations, particularly the scarcity of large, domain-specific corpora and the need for better integration into educational contexts. Future developments should focus on building richer resources and tailoring models to the unique features of Ancient Greek, thereby fully realizing the potential of these technologies in both research and teaching.

Frequent itemset mining is a foundational technique in data mining, used extensively for discovering associations within large datasets. However, the presence of sensitive itemsets poses a significant challenge, necessitating data sanitization to protect privacy without compromising the utility of the mined itemsets. This paper introduces an AI-empowered approach to border sanitization that effectively hides sensitive itemsets while preserving the usability of non-sensitive itemsets. Our method involves generating frequent itemsets, identifying sensitive and non-sensitive borders, and applying optimization techniques such as safe sets and linear programming. The approach leverages advanced AI techniques, including Markov Random Fields, to optimize the sanitization process. Experimental results demonstrate the method’s effectiveness in maintaining data utility and privacy, highlighting the potential of AI in enhancing privacy-preserving data mining techniques.

In the age of Big Data, data is at the core of innovation, driving growth across industries such as healthcare, finance, retail, government, and the Internet of Things (IoT). However, this data deluge has caused critical concerns regarding human freedoms such as the privacy of individuals. As we collect, store, and analyze vast amounts of data, maintaining the balance between data utility and privacy becomes increasingly challenging and expensive. This paper provides a comprehensive exploration of privacy-preserving techniques and their applications, including traditional methods like k-anonymity, l-diversity, and t-closeness, as well as advanced approaches such as differential privacy, homomorphic encryption, and privacy-preserving machine learning. We investigate their historical development, real-world applications across various sectors, and the ethical challenges they pose, particularly in the context of Generative AI and Big Data analytics. The objective is to underscore how these techniques can enable the utilization of data without infringing on individual rights, ensuring in this way compliance with legal frameworks like GDPR and HIPAA. This paper also highlights research directions and future trends, emphasizing the importance of interdisciplinary collaboration in advancing privacy preservation while fostering innovation.

The effective management of Emergency Department (ED) overcrowding is essential for improving patient outcomes and optimizing healthcare resource allocation. This study validates hospital admission prediction models initially developed using a small local dataset from a Greek hospital by leveraging the comprehensive MIMIC-IV dataset. After preprocessing the MIMIC-IV data, five algorithms-Linear Discriminant Analysis (LDA), K-Nearest Neighbors (KNN), Random Forest (RF), Recursive Partitioning and Regression Trees (RPART), and Support Vector Machines (svmRadial)-were evaluated. Among these, RF demonstrated superior performance, achieving an Area Under the Receiver Operating Characteristic Curve (AUC-ROC) of 0.9999, sensitivity of 0.9997, and specificity of 0.9999 when applied to the MIMIC-IV data. These findings underscore the robustness of RF in handling complex datasets for admission prediction, establishing MIMIC-IV as a valuable benchmark for validating models based on smaller local datasets and providing actionable insights for steering ED management strategies in the right direction.

CONTEXT:Guideline-directed medical therapy of heart failure (HF) primarily targets neurohormonal activation. However, GH has emerged as a potential treatment for the multiple hormonal deficiency syndrome, which is associated with worse outcomes in HF. OBJECTIVE:This study evaluates the efficacy and safety of GH therapy in HF. DATA SOURCES:A systematic search was conducted in PubMed, Cochrane Library, and ClinicalTrials.gov, according to PRISMA guidelines. STUDY SELECTION:Randomized, placebo-controlled trials studying GH therapy in adult HF patients were included. Of the 1184 initially identified records, 17 studies (1.4%) met the inclusion criteria. DATA EXTRACTION:Two independent authors conducted the search, with any disagreements resolved by a third author. Study quality was assessed using predefined criteria, including randomization, blinding, and the presence of a placebo group. DATA SYNTHESIS:A random-effects model was applied due to heterogeneity across studies. GH therapy significantly improved left ventricular ejection fraction (+3.34%; 95% CI, 1.09-5.59; P = .0037), peak oxygen consumption (+2.84 mL/kg/min; 95% CI, 1.32-4.36; P = .0002), and New York Heart Association class (-0.44; 95% CI, -0.08 to -0.81; P = .023). GH therapy also reduced the composite of death, worsening HF or ventricular tachycardia by 41% (RR = .59; 95% CI, 0.39-0.90; P = .013). Subgroup analyses indicated that patients with ischemic cardiomyopathy, baseline ejection fraction ≥30%, and longer treatment duration experienced greater benefits. CONCLUSION:GH therapy demonstrated improvements in cardiac function, exercise capacity, and HF symptoms, along with a statistically significant trend toward improvements in hard endpoints. Event-driven trials are needed to validate these findings.

The present research focus on the use of Named Entity Recognition (NER) for the anonymization of qualitative educational data in Greek. More specifically, it investigates the effectiveness of NER models for the Greek language in identifying sensitive information, such as tutor names, within open-ended student responses collected from course evaluation surveys at the Hellenic Open University (HOU). Five different NER models were examined, both individually and in combinations, to determine their performance in identifying proper names. The findings show that while all models demonstrate high precision and recall for non-entity classes, however significant differences are found in their ability to identify named entities. The Toolkit model is the best performing model, achieving high recall and precision, with some combinations of NER models improving recall but resulting in lower precision. The findings underscore the criticality of choosing appropriate NER models, especially in cases that involve category imbalance, and presents several trade-offs between precision and recall that one should consider in NER and anonymization tasks.

This study examines the forecasting of all-cause hospitalizations in the Greek elderly population until 2032, using historical data from 2001 to 2019. We employed two forecasting models: Autoregressive Integrated Moving Average (ARIMA) and Prophet model. The ARIMA model demonstrated a conservative approach, generating stable forecasts with narrower confidence intervals, making it suitable for identifying gradual trends. In contrast, the Prophet model, with its flexibility in trend capture, produced forecasts with broader confidence intervals, capturing potential sharp increases but with greater uncertainty. Our findings underscore that forecasting accuracy varies across age groups, with the highest precision observed in the 80+ age cohort, reflecting the more predictable healthcare utilization patterns of older populations. These insights emphasize the value of a multi-model approach in healthcare planning, particularly for accurately predicting trends within aging populations and efficiently allocating healthcare resources.

Accurately assessing body fat percentage (BF%) is crucial for healthcare and fitness but is hindered by gold-standard methods that are costly and invasive. This study employs a dataset containing variables such as age, sex, Body Mass Index (BMI), and body circumferences, from individuals whose body fat percentage (BF%) was estimated via underwater weighing, to develop predictive machine learning models. Multiple regression techniques alongside a neural network were employed to compare model accuracies in estimating BF%. Ridge Regression emerged as the most effective model, demonstrating the highest R2 score. Notably, feature importance analysis using ElasticNet and SHAP revealed that abdominal circumference was the most significant predictor of BF%, challenging the adequacy of BMI as a measure of adiposity. These insights advocate for the broader adoption of circumference measurements in everyday practice to enhance the predictive accuracy of cost-effective and easily performed BF% estimation.

The ASCAPE project aims to improve the health-related quality of life of prostate cancer patients using artificial intelligence-driven solutions. This study tries to unravel the complex relationships between patient data variables and urinary incontinence (UI), and post-radical prostatectomy using the ASCAPE datasets. We employed a Generalized Additive Model to analyze patient-reported outcomes on UI (QLQ PR25 questionnaires over a 12-month period), and objective data derived from wearable devices. Our findings showcase age and comorbidities as the main predictors of incontinence severity, whereas physical activity failed to show any significance in our model. Our study highlights the importance of a personalized approach to incontinence care, where patient characteristics and recovery patterns are considered when developing treatment plans.

The discipline of Entity Resolution (ER), the process of identifying and linking records that refer to the same real-world entity, has been fundamentally reshaped by the adoption of high-dimensional vector embeddings. This transformation reframes ER as a large-scale Approximate Nearest Neighbor Search (ANNS) problem, making the choice of ANNS architecture a critical determinant of system performance. This paper provides a deep architectural comparison and a novel, large-scale empirical evaluation of the two dominant ANNS paradigms: graph-based methods (HNSW, DiskANN) and partition-based methods (Faiss-IVF+PQ, Scann). We introduce a new semi-synthetic benchmark tailored to the ER task, consisting of two one-million-vector datasets with a known ground truth. On this benchmark, we conduct a comprehensive evaluation, measuring not only total query time but also disaggregated blocking and matching times, alongside canonical ER quality metrics: precision, recall, and F1-score. Our findings reveal that partition-based methods, particularly Scann, offer superior performance in high-throughput, moderate-recall scenarios, while graph-based methods like HNSW and DiskANN are unequivocally superior for applications demanding the highest levels of matching quality. This work provides a nuanced, application-centric analysis that culminates in a set of actionable recommendations for practitioners designing modern data integration and retrieval systems.

Traditional methods for patient cohort identification from Electronic Health Records (EHRs) are often slow, labor-intensive, and fail to capture the rich semantic nuance embedded in unstructured clinical narratives. This paper introduces a scalable, end-to-end framework that creates a synergy between deep medical embeddings and Approximate Nearest Neighbor Search (ANNs) to overcome these limitations. We detail a complete pipeline that begins with preprocessing multi-modal EHR data and creating holistic patient representations using a domain-specific language model combined with an intelligent gated fusion mechanism. These high-dimensional embeddings are then indexed using an ANN method to enable near real-time retrieval. A comprehensive experimental evaluation was conducted on the MIMIC-III and MIMIC-IV datasets, comparing the retrieval performance of ClinicalBERT against BioBERT across several ANN algorithms. The results demonstrate that the combination of ClinicalBERT and HNSW consistently achieves the highest retrieval accuracy, with F1-Scores exceeding 0.78, and query latencies under 10 ms. This framework enables a paradigm shift towards high-speed, semantic patient similarity search, with significant implications for accelerating clinical trial recruitment, augmenting clinical decision support, and paving the way for a new era in data-driven precision medicine.