AI doi: 10.3390/ai7070249

Authors: George Dueñas Sergio Jimenez Geral Eduardo Mateus Ferro

Large Language Models (LLMs) achieve strong performance on standardized examinations, yet the language-mediated mechanisms through which they assess answer options in Multiple-Choice Questions (MCQs) and diverge from human difficulty judgments remain poorly understood. We argue that predicting the difficulty of MCQs provides a lens for studying how LLMs represent plausibility, uncertainty, and error across competing response options. reviewer1comments1mlAlthough recent deep machine learning approaches achieve competitive accuracy through large feature sets and complex architectures, their limited interpretability reduces their value for understanding model behavior. We propose an interpretable framework that decomposes item difficulty into LLM-based plausibility estimates over response options. These estimates are elicited through direct prompting and pairwise contrastive comparisons, and then integrated into a rational model that expresses item difficulty as a ratio between the plausibility of distractors and the plausibility of the correct option. We evaluated this approach on two high-stakes datasets. reviewer1comments1usmleUsing the United States Medical Licensing Examination (USMLE) dataset, reviewer1comments1rmsethe model achieved a Root Mean Squared Error (RMSE) of 0.277, comparable to previous approaches, while reducing the representation of the underlying elements from hundreds of features to only three parameters. Under Spearman rank correlation, the model reached ρ=0.427 on USMLE, representing a 70.8% relative improvement over previously reported results, and ρ=0.488 on ICFES, a new dataset. A complementary ranking analysis further reveals a systematic inversion between LLM-based difficulty judgments and those of experts, exposing divergences between model-internal assessments and human response patterns. These findings position option plausibility based on divide-and-conquer prompting as a principled framework for probing LLM decision processes, their rank-order misalignment with human response patterns, and their challenges in educational settings.

AI doi: 10.3390/ai7070248

Authors: Gilberto Huesca Yolanda Martinez-Trevino Claudia Gabriela Jiménez González David Alonso Cantú Delgado Christelle Navarrete Antonio Cedillo-Hernandez Ricardo Rafael Quintero Meza

The rapid integration of Generative Artificial Intelligence (GenAI) into programming education has raised important questions regarding its impact on learning processes, conceptual understanding, and technological dependency. This study analyzed the effects of four GenAI-supported instructional strategies in an introductory programming course for undergraduate engineering students. A multi-group quasi-experimental pre-test–post-test design was implemented involving 686 students distributed across 53 class groups, from 10 campuses, taught by 32 professors. The instructional conditions included Quizzes for Self-Regulation, Github-Copilot-assisted learning, Prompt Problems with Iterative Refinement, and Flipped Learning enhanced with GenAI, which were compared against a traditional teaching approach. Learning outcomes were measured using normalized learning gain, while statistical analyses were conducted using non-parametric methods due to deviations from normality and heteroscedasticity. Results indicate that GenAI integration did not produce statistically significant overall differences in learning gain when all GenAI-supported strategies were analyzed as a single cluster compared to traditional instruction. However, differences emerged between specific strategies, with Quizzes and Copilot-based approaches having higher median learning gains than Prompt Problems and Flipped Learning strategies. No statistically significant differences associated with gender were identified. These findings suggest that the effectiveness of GenAI in programming education depends less on the mere presence of the technology and more on the pedagogical conditions under which it is integrated into the teaching–learning process.

AI doi: 10.3390/ai7070247

Authors: Mishell Cadena-Yanez Angela Bernardini Marisol Gómez

This study examined how healthcare professionals (HCP) perceive and trust AI in clinical settings, and what kinds of explanations they need to use it effectively. Using interviews with six HCPsand a questionnaire (N=41), the findings show that trust was higher for administrative tasks and lower for direct clinical decisions, regardless of their prior AI or clinical experience or AI literacy. Clinical validation and algorithmic bias ranked above explainability as trust factors, indicating that explainability is not a primary trust-building mechanism. However, HCP consistently demanded explanations as a tool to support their own clinical reasoning, preferring to receive them across all clinical contexts rather than only under disagreement with the system, and valued grounding in medical evidence and consistency with clinical protocols over clarity or simplicity. These findings argue for an HCXAI design approach that treats explainability as a critical-reasoning tool rather than a primary trust mechanism.

AI doi: 10.3390/ai7070246

Authors: Yildiray Kabak Gokce B. Laleci Erturkmen Mert Gencturk Tuncay Namli A. Anil Sinaci Ruben Alcantud Corcoles Cristina Gómez Ballesteros Pedro Abizanda Volkan Atmis Asuman Dogac

Background: Evidence-based clinical guidelines are essential for high-quality care yet translating them into personalized clinical decision support remains resource-intensive and time-consuming. Large language models (LLMs) show promise for supporting clinical decision-making, but their limited access to patient-specific data and explicit guideline sources constrains trustworthiness, personalization, and clinical applicability. Retrieval-augmented generation (RAG) addresses part of this challenge by grounding model outputs in curated evidence sources; however, true personalization requires structured access to electronic health record data. Methods: This study presents FHIR-RAG-MEDS, a medical decision support system that integrates HL7 Fast Healthcare Interoperability Resources (FHIR) with an RAG-enhanced LLM to enable patient-specific, guideline-concordant clinical recommendations. Through SMART on FHIR, the system retrieves real-time patient data from FHIR servers, generates structured medical summaries, and incorporates this personalized context into the RAG pipeline, grounding responses in evidence-based clinical guidelines stored in a vector database. Results: FHIR-RAG-MEDS was evaluated using 139 physician-generated clinical questions covering dementia, chronic obstructive pulmonary disease, hypertension, and sarcopenia. Performance was assessed using automated metrics, RAG-specific evaluation frameworks, and independent expert physician review. The system consistently outperformed state-of-the-art medical LLMs, demonstrating higher semantic accuracy, improved faithfulness to guideline content, and stronger clinical relevance. Conclusions: Integrating HL7 FHIR with RAG-based LLMs enables trustworthy, personalized clinical decision support, bridging the gap between static language models and real-world, patient-centered care.

AI doi: 10.3390/ai7070245

Authors: Maryam Khanian Najafabadi Bushra Naeem Touraj Khodadadi Saman Shojae Chaeikar Zawar Shah

This study proposes a time series-aware Large Language Model (LLM)-driven feature engineering framework for tabular prediction tasks. Existing automated feature engineering methods, including LLM-based approaches such as CAAFE and OCT-Tree and established libraries such as tsfresh and Featuretools, can generate useful features for general tabular data but do not explicitly address temporal availability constraints in time series settings. This can lead to data leakage when variables that are only available after the prediction event are used directly during model training. To address this limitation, the proposed framework classifies variables into antecedent features, consequent features, and historical aggregated features. The key innovation is that consequent variables are not discarded to prevent leakage but are instead routed into a leakage-safe historical aggregation pipeline, recovering predictive signal from post-event variables through temporally valid past values. The framework guides an LLM to generate structured feature engineering configurations, applies temporally valid transformations, performs feature selection, and evaluates the selected features using predictive models. A formal leakage control mechanism ensures that all aggregations use strictly past observations, applied within entity groups and before the temporal train–validation–test split. The framework is evaluated on two time series tabular tasks: Tesla stock prediction and English Premier League match outcome prediction. The results show that the proposed approach improves predictive performance compared with raw-feature baselines and selected existing automated feature engineering methods. On the Tesla dataset, the framework reduced MAE compared with both the baseline and the reported OCT-Tree result. On the EPL dataset, it improved accuracy compared with the odds-only baseline and the reported Azure ML preprocessing result. These findings suggest that combining LLM reasoning with explicit temporal constraints is a practical direction for automated feature engineering in time series tabular machine learning.

AI doi: 10.3390/ai7070244

Authors: Meryem Chajia El Habib Nfaoui Soufiyan Ouali

Customer churn has become a major challenge in the retail industry, where customer loyalty directly affects business success and sustainability. Despite the significant progress in Artificial Intelligence, especially in prediction tasks, its use in the retail e-commerce domain remains limited and underexplored; this is due to the scarcity and limited quality of available datasets. To address these challenges, this paper proposes a churn prediction approach designed to handle data scarcity while ensuring accurate performance. We experimented with a combination of various feature selection techniques along with several Machine Learning and Deep Learning models to evaluate their performance on a limited tabular dataset. The impact of feature selection on predictive performance was also systematically analyzed. The results demonstrated that feature selection plays an important role in improving model performance by identifying the key features that have the most significance to the classification task. The analysis showed that the L1-based Logistic Regression feature selection method combined with the Extreme Gradient Boosting classifier achieved the best performance, with a Macro F1-score of 95.25%. Based on these results, companies can identify potential churners and implement retention strategies. These findings may provide a useful reference point for future researchers in the retail e-commerce industry.

AI doi: 10.3390/ai7070243

Authors: Ashley Allen Alexios Mylonas Stilianos Vidalis Nikolaos Pitropakis

The last five years have seen considerable growth in the topic of peer-to-peer (P2P) federated learning (FL). This framework removes the central coordinating server used in conventional federated learning and instead requires participating nodes to manage model training, peer selection, communication, aggregation, and trust directly. This provides a promising route for privacy-preserving and decentralised machine learning, but it also introduces unresolved challenges in topology selection, participant incentivisation, communication efficiency, security, and evaluation. Existing studies frequently evaluate proposed methods under narrow assumptions, such as static network membership, homogeneous devices, fixed bandwidth, limited topology choices, and public benchmark datasets. Existing surveys also tend to present taxonomies of decentralised federated learning rather than synthesising how topology, incentives, and communication algorithms jointly affect deployability. This paper reviews recent work on peer-to-peer federated learning across three connected dimensions: network topology, incentive mechanisms, and communication algorithms. We compare the topologies, datasets, experimental assumptions, incentive designs, communication strategies, and open issues reported in the literature. The review shows that highly connected topologies tend to improve convergence but increase communication overhead and vulnerability to bottlenecks; sparse and dynamic topologies improve efficiency but create challenges for convergence, reliability, and node drop-out. Incentive mechanisms increasingly combine reward, reputation, validation, and punishment but remain weakly validated under realistic churn, heterogeneous resources, and adversarial behaviour. Communication algorithms reduce bandwidth through gossip, sparsification, prediction, routing, and multi-step aggregation but often trade communication savings against accuracy, robustness, and generalisability. Across all three areas, the field lacks standardised benchmarks, reproducible experimental settings, and realistic evaluation under unstable peer-to-peer conditions. We conclude by identifying cross-cutting research gaps and recommending future work on dynamic topologies, heterogeneous devices, real-world datasets, incentive robustness, and comparable benchmarking.

AI doi: 10.3390/ai7070242

Authors: Mohammad Odeh Ahmad Hasasneh

Precise weather forecasting plays a critical role in sectors such as agriculture, transport, energy management, and climate change adaptation, and machine learning and deep learning algorithms have been widely used for data-driven time series forecasting problems. In this work, we explore the application of machine learning and deep learning models for multi-weather variable forecasting in a dataset recorded over a period of ten years (2015–2025) for five weather stations in Palestine. The dataset comprises measurements for temperature, relative humidity, wind speed, precipitation, atmospheric pressure, and sunshine hours. To avoid the issue of temporal leakage, a chronological training, validation, and test set splitting approach was used in the evaluation experiments. The models used in this study include ARIMA, SARIMA, Random Forest, XGBoost, CNN, LSTM, GRU, ConvLSTM, CNN-GRU, and CNN-LSTM with station embeddings. Our experimental results indicate that the XGBoost model achieved the highest performance in predicting temperature and relative humidity (R2 = 0.953 and R2 = 0.670, respectively), while deep learning methods exhibited high accuracy across several weather features. The CNN-LSTM model was successfully able to learn temporal–spatial patterns via station embeddings, while recurrent neural networks performed impressively in forecasting sunshine hours and atmospheric pressure.

AI doi: 10.3390/ai7070241

Authors: Yunfei Wang Weidong Jia Mingxiong Ou Xiang Dong Shiqun Dai Rong Zhang Yaning Wang Wenrui Zhu

Real-time kinematic (RTK) navigation, which enables centimeter-level positioning accuracy through carrier-phase differential correction, provides high-accuracy positioning for orchard robots, but short-term outages caused by canopy occlusion and signal interference may interrupt path guidance and increase lateral drift. To address this issue, this study proposes a vision-based takeover navigation method for orchard robots under short-term RTK failure conditions. First, an improved YOLOv11-based road segmentation and completion model, termed YOLOv11-VF, was developed. By introducing a Squeeze-and-Excitation (SE) channel attention mechanism, the model jointly perceives visible road regions and occluded road completion regions, thereby producing continuous and complete road semantic representations. Second, a structured geometric road representation was constructed from the segmentation results to extract the navigation reference line, and a joint direction-position constraint mechanism was established by integrating the reference line with the robot reference view axis. A hierarchical constraint strategy based on a travel corridor and a deadband region was further designed to jointly determine heading deviation and lateral drift. Finally, road segmentation, navigation-line extraction, parameter analysis, and vision-based takeover experiments were conducted in a standardized orchard environment. The results showed that YOLOv11-VF achieved Precision, Recall, AP50, mAP@0.5:0.95, and F1 values of 92.31%, 88.56%, 94.40%, 67.41%, and 90.40, respectively, showing the best overall segmentation performance among all compared models while maintaining good real-time performance. The proposed method also demonstrated high consistency in navigation-line extraction and maintained mean absolute deviations of 0.0176 ± 0.0041 m to 0.0718 ± 0.0138 m during RTK outage intervals over 10 repeated trials, indicating good path-following capability and operational stability.

AI doi: 10.3390/ai7070240

Authors: Lujain Abdulrahman Alawwad Mohamed El Bachir Menai

While aspect-based sentiment analysis (ABSA) has made significant progress in the identification of explicit opinion targets, the more challenging case of implicit aspects remains insufficiently studied. Implicit aspect extraction is particularly challenging, as it relies on contextual and semantic cues and requires systems to infer what reviewers mean rather than what they state explicitly. A four-component hybrid pipeline is proposed for explicit and implicit aspect extraction, formulating the task as controlled text generation. The pipeline combines (i) a fine-tuned decoder-only large language model as a generative baseline, (ii) an iterative residual generation strategy that recovers multiple aspects through successive masked generation passes, (iii) paraphrase-based input transformation to broaden the contextual signal, and (iv) domain-specific knowledge graphs activated by linguistic signals to infer implicit aspects. The novelty lies not in the individual components themselves but in their principled orchestration and the linguistically motivated gating logic governing the activation of each stage. Extensive experiments are conducted on eight benchmark ABSA datasets spanning both English and Arabic: SemEval-2014, SemEval-2015, SemEval-2016, ACOS, and M-ABSA for English; and SemEval-2016, HAAD, and M-ABSA for Arabic. The proposed solution outperforms strong baseline methods and recent state-of-the-art models on English datasets, with F1-scores of 0.8533, 0.713, 0.7859, 0.793, and 0.664, respectively. On Arabic datasets, the best-performing configurations achieve F1-scores of 0.7632, 0.4765, and 0.7656 on SemEval-2016, HAAD, and M-ABSA, respectively, with the knowledge-graph component providing consistent and statistically significant gains for implicit aspect identification in both languages. These results demonstrate the effectiveness of generative modeling, iterative generation, paraphrasing, and structured knowledge for aspect extraction and highlight the potential of the proposed approach for implicit aspect identification, in particular for morphologically rich languages such as Arabic, where annotated resources remain scarce.

AI doi: 10.3390/ai7070238

Authors: James Vakilian Fareed Ud Din Edmund J. Sadgrove Mohammadreza Haghighat Niusha Shafiabady

The integration of international students into host countries and their cultures is a multifaceted challenge that significantly impacts their academic success and well-being. This study leverages Explainable Artificial Intelligence (XAI) to model and interpret variables associated with the self-rated integration of 175 international students at Charles Darwin University (CDU) in Australia, using data from a 42-question survey. Employing machine learning models, including Decision Tree (DT) and Gradient Boosting Machine (GBM), we use XAI techniques to identify variables most strongly associated with students’ self-rated integration, including career confidence, perceived future happiness, and perceived career obstacles. SHAP analyses and partial dependence plots provide global and instance-level insights, revealing both the magnitude and directional effects of these features. The findings highlight the predictive relevance of psychological and social variables in students’ self-rated integration, offering exploratory insights that inform targeted support programs. By enhancing model transparency through XAI, this research fosters trust in AI-driven educational interventions, addressing ethical considerations and promoting equitable outcomes for diverse student populations.

AI doi: 10.3390/ai7070239

Authors: María Isabel A. Benítez Serafín Moral-García Joaquín Abellán

Artificial intelligence systems that learn or reason from finite empirical data often require uncertainty representations that go beyond a single precise probability distribution. This is especially relevant when observations are scarce, incomplete or not reliable enough to support precise probabilistic assessments. In current data-driven AI tools, empirical information extracted from data must often be converted into a structured uncertainty model before it can be used for reasoning, learning or decision support. The singleton intervals induced by NPI-M and A-NPI-M provide such a representation, since they express the predictive information obtained from the observed data without introducing externally chosen cautiousness parameters. Evidence theory is useful in this context because it allows partial support to be assigned to sets of alternatives, making it suitable for representing imperfect knowledge in AI systems. This paper studies how Non-Parametric Predictive Inference for multinomial data (NPI-M) can be connected with evidence theory through reachable probability intervals. Since the exact NPI-M model does not directly define a credal set, we focus on its approximated version, A-NPI-M, which preserves the NPI-M singleton bounds and represents them through reachable probability intervals. We analyze whether the resulting credal set can be represented exactly by a belief function, showing that this is not possible in general, although exact representations may exist in particular cases. Motivated by this limitation, we construct a basic probability assignment whose belief and plausibility values reproduce the A-NPI-M singleton bounds. The resulting belief function preserves the marginal interval information of A-NPI-M while adding an evidential structure on composite events, and its associated set of compatible probability distributions is included in the A-NPI-M credal set. The construction is presented by cases, illustrated with numerical examples and compared with the belief-function representation of the Imprecise Dirichlet Model. The proposed model provides a theoretical representation layer that may support uncertainty-aware AI procedures by transforming empirical predictive information into structured imperfect knowledge before reasoning, learning or decision-support criteria are applied.

AI doi: 10.3390/ai7070237

Authors: Igor Adamenko Orpaz Ben Aharon Yehudit Aperstein Alexander Apartsin

Urban environments are saturated with imaging sensors deployed for purposes unrelated to vehicle identification, from ATM and dashboard cameras to pole-mounted CCTV and smartphones. We term the use of such non-purpose-built sensors for secondary inference “opportunistic sensing”; its central question is where, under uncontrolled capture conditions, AI-enabled restoration remains reliable. This paper introduces recoverability maps, a task-agnostic methodology for quantifying that boundary, and applies it to oblique-view license plate recognition (LPR). It pairs a full-grid synthetic sweep of the degradation space with two summary measures: a boundary area-under-curve for coverage and a reliability score F for the frequency and depth of interior unrecovered pockets. For LPR, the space is the oblique-angle grid [0°,89°]2 sampled by Scrambled Sobol sequences, and the utility is plate-level optical character recognition (OCR) accuracy. Within this synthetic benchmark, approximately 90–92% of the angle grid is recoverable (best single model to union of restoration arms), recovery degrades sharply beyond roughly 80° in both axes, and lateral rotations are harder to reconstruct than elevational ones. Five restoration architectures cluster within a narrow AUC band of 0.89–0.93, and share the same α/β asymmetry, so the recoverable region is set primarily by sensing geometry, with architecture affecting efficiency and interior consistency; discriminative architectures outperform generative models. The methodology is validated on real plates: on CCPD and the Brazilian legacy and Mercosur layouts of RodoSol-ALPR, restoration raises held-out extreme-angle recognition by +15 to +38 exact-match points under plate-specialized recognizers, and the discriminative-over-generative ordering reproduces on real data.

AI doi: 10.3390/ai7070236

Authors: Leonardo Eleuterio Ariello Kelvin Wang David Newman-Toker Jee Bang David P. W. Rastall

Background: Artificial intelligence in medicine has largely relied on supervised training of disease-specific models, limiting scalability in conditions where labeled data are scarce. Large language models (LLMs), which encode broad medical knowledge through large-scale pretraining, offer an alternative paradigm in which structured physiological measurements can be interpreted directly without task-specific model training. Objective: To evaluate whether smartphone-derived ocular motor biomarkers can be translated into clinically meaningful probabilistic assessments of Huntington’s disease (HD) using general-purpose LLMs operating as inference engines. Methods: In this prospective proof-of-concept study, 26 participants (13 with genetically confirmed HD and 13 age-matched controls) completed a standardized ocular motor assessment using a custom smartphone application. Quantitative eye-movement metrics were validated against expert neurologist ratings. Structured physiological features were then provided to four general-purpose LLMs without task-specific training or diagnostic labels, and the models generated an AI-Assigned HD Probability Score (HAIPS). Discriminative performance and associations with clinical severity measures were evaluated. Results: Smartphone-derived ocular motor metrics showed strong agreement with clinician assessments (Spearman ρ = 0.76–0.95; all p < 0.001), confirming preservation of clinically meaningful physiological signals. LLM-derived HAIPS distinguished HD from controls with high accuracy (AUC 0.879–0.944), with no significant differences across models. Discrimination was statistically equivalent to a supervised logistic regression model trained on the same features. HAIPS correlated strongly with established measures of disease severity, including cognitive (MoCA, ρ = −0.86), functional (TFC, ρ = −0.74), and motor impairment (UHDRS, ρ = 0.85) (all p ≤ 0.003). Conclusions: Structured ocular motor biomarkers acquired using a consumer smartphone can be translated into clinically meaningful probabilistic assessments of HD by general-purpose LLMs without disease-specific model training. These findings support a framework in which physiologically grounded digital biomarkers are coupled with general-purpose inference models, potentially enabling scalable assessment in rare neurological diseases where labeled data are limited.

AI doi: 10.3390/ai7070235

Authors: Tim Geppert Andreas Block Maria Rothstein Mario Gellrich

Background: Artificial intelligence (AI) adoption is increasingly reported as a priority for organizations, yet they face a growing, fragmented body of evidence concerning the factors that influence successful AI integration. Method: To identify the relevant factors for organizational AI adoption, we conducted a systematic literature review (SLR) following PRISMA guidelines, which yielded 37 quantitative empirical studies. From these studies we extracted 1229 paper-item instances, of which 810 were retained after applying structured exclusion criteria to develop a domain model relevant to organizational AI adoption. The model’s content validity was assessed and supported through expert feedback using the Content Validity Index (CVI) methodology. Results: We organized 24 subclusters into nine main clusters across the three dimensions Technology (Enablers, Usability, Trust), Organization (Leadership, People, Process), and Environment (Market, Regulatory, Partner). Our analysis suggests that workforce skills, perceived intelligence, and resources are among the most frequently studied and positively associated antecedents of AI adoption, and that constructs related to AI explainability and control (human-in-the-loop oversight) have received little research attention and remain underrepresented despite growing regulatory requirements such as the EU AI Act. Conclusions: The resulting domain model provides an empirically grounded classification of organizational AI adoption factors and can serve as a foundation for future measurement instruments.

AI doi: 10.3390/ai7070234

Authors: Taehyeon Kim Kyung-Taek Lee

Group Relative Policy Optimization (GRPO) streamlines the alignment of Large Language Models (LLMs) and Vision–Language Models (VLMs) by eliminating the Critic model. However, its efficiency heavily depends on the group size, G. While a larger G improves reward estimation and stabilizes the Advantage, Ai, it drastically increases VRAM usage and reduces throughput. Standard heuristics like a fixed G of 64 create significant bottlenecks in resource-constrained settings. This paper introduces an Efficiency-Aware optimization framework utilizing Multi-fidelity Bayesian Optimization and Hyperband (BOHB) to dynamically identify the optimal group size, G∗. The method uses a multi-objective function that balances reward accuracy, Ai variance, and hardware utilization, applying z-score normalization. By employing Successive Halving to quickly evaluate candidates at low fidelity, the framework reduces search costs by up to 74% compared with random search. Tested across text-only LLMs (Qwen2.5-7B/1.5B) and multimodal VLMs (Qwen2.5-VL-3B), the framework demonstrates that the discovered G∗ saves up to 72.5% in VRAM compared with the baseline of 64, while maintaining reward accuracy within 5.8%. Sensitivity analyses on hyperparameters like λ, α, and β confirm the framework’s robustness. Rather than treating group size as a mere engineering heuristic, this study establishes a principled methodological advance by formalizing the trade-off between statistical estimation stability and hardware constraints into a unified optimization framework for resource-efficient RLHF.

AI doi: 10.3390/ai7070233

Authors: Rodrigo Garcia Horderlin Robles

Precision Livestock Farming (PLF) increasingly relies on edge-deployed sensors to monitor bovine behaviors, fostering improved welfare and management. However, behavioral data naturally expands over time and presents severe class imbalances due to animals’ predominantly sedentary routines. When continuous sequential updates are required without access to historical datasets, deep learning methods frequently succumb to catastrophic forgetting. This study introduces an ultra-lightweight (∼0.85 MB) Continual Learning (CL) architecture built upon a CNN-BiLSTM feature extractor, tailored to process multivariate Inertial Measurement Unit (IMU) streams. We exhaustively evaluated baseline Naïve Fine-Tuning against Elastic Weight Consolidation (EWC), Learning without Forgetting (LwF), and episodic Replay under three rigorous real-world paradigms: Class Incremental, Subject Incremental (domain shift), and Imbalanced Realistic scenarios. Our empirical findings expose the fragility of static paradigms: in Class Incremental expansions, Naïve Fine-Tuning collapsed to an Average Accuracy of 33.33%. Conversely, Experience Replay emerged as the most robust defense, achieving a statistically significant Average Accuracy of 74.64 ± 6.77% across multiple random seeds. Furthermore, LwF effectively mitigated structural variations across unseen animal domains (Subject Incremental) without requiring raw data buffers. Notably, under severe biological class imbalances (Imbalanced Cumulative), the architecture proved highly resilient, maintaining 98.46% Average Accuracy and retaining perfect minority class recall. This research validates the operational feasibility of deploying adaptive, privacy-preserving CL frameworks directly on low-power wearable devices for lifelong livestock monitoring.

AI doi: 10.3390/ai7070232

Authors: Muhammad Ali Shafique Imran Latif Hayat Ullah Alex C. Newkirk Arslan Munir

The rapid scaling of large language model (LLM) training has intensified demand for Graphics Processing Unit (GPU) clusters balancing throughput with energy efficiency. While NVIDIA’s H100 and B200 architectures are increasingly deployed in production datacenters, their comparative behavior under distributed training remains insufficiently characterized beyond vendor specifications, leaving datacenter operators without empirical guidance on metrics such as TFLOPs/kW and tokens-per-kilojoule. This work presents a system-level evaluation of single-node 8× H100 and 8× B200 configurations using Distributed Data Parallel (DDP) training across LLMs and vision–language models (VLMs) ranging from 7B to 32B parameters, spanning various real AI workload scenarios. We benchmark end-to-end throughput, utilization, power, energy, TFLOPs/kW, and tokens-per-kilojoule, complemented by architectural analysis explaining observed behavioral differences. Across LLM workloads, B200 achieves higher utilization (1–6%), faster training (up to 15%), and greater compute efficiency (up to 32% higher TFLOPs/GPU), attributable to higher memory bandwidth and large streaming multiprocessor (SM) count. However, B200 exhibits lower TFLOPs/kW and tokens-per-kilojoule, revealing a fundamental trade-off: throughput gains come at a measurable energy cost per useful token. VLM results further expose model-dependent asymmetries, with B200 consuming disproportionately more energy for lighter compute kernels due to elevated baseline power draw. These findings provide an empirical framework distinguishing compute efficiency from energy efficiency across next-generation GPU nodes, offering practical guidance for energy-aware AI datacenter design.

AI doi: 10.3390/ai7060231

Authors: Duc Nghia Vu Janos Demetrovics

The Q-Learning Relational Game (QLRG) framework provides a theoretically rigorous method for identifying minimal winning coalitions in online social networks (OSNs) under the restrictive assumption of global agent symmetry or uniform matroid structure. Real-world OSNs, however, exhibit significant asymmetry. This paper introduces the Cluster-Based Q-Learning Relational Game (C-QLRG), a practical extension that relaxes the global symmetry requirement by leveraging community structure. We partition the agent set into communities with bounded internal variation and represent the state solely by community membership counts of the seed set. Because the closure operator already captures all eventual influence spread, the problem reduces to a sequential seed selection task where the agent decides, at each step, from which community to add the next seed. We prove that the optimal Q-function of a suitably regularized reach-efficiency objective is Lipschitz continuous and derive a performance bound for the learned policy. The full algorithm is presented, and its complexity is analyzed. Empirical evaluations on a synthetic asymmetric network and Zachary’s Karate Club demonstrate that C-QLRG is highly sensitive to reward parameters, where default settings lead to premature stopping, but parameter tuning combined with a corrected minimality verification recovers high-efficiency coalitions by removing non-contributing agents. With tuned parameters, C-QLRG produces a near-winning coalition of size 11 and 99% reach on the synthetic network, surpassing the greedy baseline’s efficiency (size 12) despite a one-node coverage gap, while identifying the optimal winning coalition of size 1 on the Karate Club dataset, matching all baselines. The framework thus offers a principled trade-off between model fidelity and scalability, with the reward design choice being critical for practical deployment.

AI doi: 10.3390/ai7060230

Authors: Ahmed Ibrahim Ali AlSanousi Ahmed Serag

Agentic artificial intelligence (AI) offers new opportunities for intelligent clinical decision support, but deployment in emergency and crisis settings remains challenging because time-critical recommendations must often be generated under incomplete patient information and system constraints. Conventional clinical decision support systems rely on rule-based workflows that degrade when structured data are absent, while standalone language models lack coordination mechanisms to enforce mandatory safety checks. We present TriAgent, a multi-agent framework that unifies adaptive orchestration, iterative retrieval, embedded safety verification, and end-to-end auditability within a single crisis clinical decision support workflow. An Orchestrator Agent dynamically selects specialist modules for clinical assessment, retrieval, treatment planning, safety verification, and system coordination, with routing determined by model reasoning rather than fixed execution paths. A retrieval sub-agent performs iterative query refinement and relevance grading over 49,000 MIMIC-IV discharge notes, while medication-conflict screening and allergy-risk assessment are invoked in parallel only when clinically indicated. A Critique Agent reviews the full reasoning trace before recommendation finalization. In a retrospective evaluation on 1000 real emergency presentations under synthesized incomplete-information inputs, TriAgent achieved 85.0% critical-case recall and 65.7% overall triage accuracy, versus at most 14.7% and 43.4% for matched single-model and retrieval-only baselines, with safety checks executed on every continuation pathway and adaptive routing invoking only the modules each case required. These results support multi-agent orchestration as a promising design pattern for transparent and auditable AI in healthcare. These gains are internal system properties; clinical-safety benefit remains to be established through prospective, clinician-involved validation.

AI doi: 10.3390/ai7060227

Authors: Zewei Mi Minming Gu

Accurate detection of tea buds is a key technology for enabling automated tea harvesting. However, in natural environments, tea buds present challenges such as scale variation, dense distribution, and high similarity to the background, making it difficult for traditional methods to balance accuracy and efficiency. To address these issues, this paper proposes a lightweight detection framework, PCM-YOLO. The model introduces a cascaded gated feature modulation network into the YOLOv11 architecture, combining feedforward structures and gating mechanisms to selectively emphasize informative features, thereby improving tea bud detection performance. In addition, a feature-enhanced downsampling module is proposed, which employs a stepwise pooling-based feature enhancement mechanism to progressively expand the receptive field while preserving feature resolution, effectively incorporating multi-scale contextual information. Finally, a multi-scale feature enhancement module is designed to reduce the computational complexity of the model while maintaining detection performance as much as possible. Experimental results on public datasets demonstrate notable performance improvements over YOLOv11-N: Precision increases from 86.7% to 90.6% (an absolute increase of 3.9 percentage points), mAP50-95 increases by 1.6%, and the number of parameters is reduced by 20.6%. These results indicate that PCM-YOLO achieves a substantial reduction in model complexity while effectively improving detection accuracy, providing a feasible technical solution for deploying high-precision, real-time tea bud detection systems at the edge in tea plantation environments.

AI doi: 10.3390/ai7060229

Authors: Weifeng Zhai Jiahui Qiu Qingkuo Wang Binbin Li He Zhang

Industrial robot after-sales question answering often depends on multimodal evidence, such as error screenshots, interface displays, and wiring diagrams, which are difficult for conventional text-based retrieval-augmented generation (RAG) systems to exploit effectively. To address this issue, we design a dual-channel multimodal RAG system that converts image content into retrievable textual knowledge through the collaboration of optical character recognition (OCR) and structured semantic description. In the proposed system, OCR is used to extract explicit textual cues, such as error codes, parameter fields, and interface prompts, while expert-authored semantic descriptions complement implicit visual evidence, including device parts, fault phenomena, and contextual scene information. The transformed knowledge is further integrated into a hybrid retrieval pipeline that combines dense retrieval and BM25, followed by Reciprocal Rank Fusion (RRF) and Maximal Marginal Relevance (MMR) reordering to improve both relevance and contextual diversity. Experiments on a real-world industrial robot after-sales dataset show that the proposed method achieves an overall question-answering accuracy of 87.9%, outperforming the LLM-only baseline by 35.6 percentage points. For image-related questions, accuracy improves from 46.7% to 83.3%. These results indicate that the proposed framework provides a deployment-friendly and interpretable system-level alternative to end-to-end multimodal model fine-tuning for industrial after-sales question answering.

AI doi: 10.3390/ai7060228

Authors: Fnu Neha Deepshikha Bhati Deepak Kumar Shukla

Whole-slide imaging has transformed histopathology into a data-rich domain; however, many computational pathology models encode tissue morphology within latent representations, limiting interpretability, reproducibility, and generalization. This review positions histomics as an intermediate phenotype representation layer linking histological images with downstream clinical inference through structured descriptors of tissue morphology, spatial organization, and tissue architecture. Unlike prior reviews focused primarily on feature extraction or predictive performance, the study adopts a representation-centric perspective of histomics. A taxonomy of histomic features across biological scales is presented, and artificial intelligence frameworks, including machine learning, deep learning, weakly supervised learning, and multimodal approaches, are systematically examined. Key challenges, including segmentation dependence, feature instability, aggregation variability, and domain shift, are critically analyzed alongside emerging developments in foundation models, representation learning, and multimodal pathology. The review provides a unified framework for understanding histomic representations and identifies future directions for developing robust, interpretable, and generalizable computational pathology systems.

AI doi: 10.3390/ai7060226

Authors: Yi Zhang Shaoqi Dai Cheng Wang Xiuhe Li Jinhe Ran Guoqiang Zhu Wenbo Liu Shuyun Shi

The introduction of noise during image acquisition and transmission is inevitable, leading to a significant reduction in the accuracy of image processing tasks, such as target classification, localization, and recognition. To address this issue, this paper proposes a novel robustness-oriented pooling module called Gaussian adaptive pooling. Drawing on the principles of Gaussian filters, the method introduces a Gaussian weight for feature values in the pooling operation, thus integrating filtering and pooling in a novel manner. This approach is both lightweight and versatile, requiring no additional learnable parameters, and enables seamless integration into neural network architectures with pooling layers. Rigorous mathematical derivations and simulation experiments show that our proposed Gaussian adaptive pooling method surpasses conventional methods (average-pooling and max-pooling) in noise handling. Furthermore, its robustness is comparable to traditional pooling methods in addressing challenges such as rotations, scalings, and translations. Extensive evaluations across multiple computer vision tasks—including image classification (CIFAR-10/100), object detection (MS COCO and RTTS), and semantic segmentation (CamVid)—confirm its effectiveness. Specifically, under varying levels of noise and degraded conditions, Gaussian adaptive pooling achieves significant improvements in standard performance metrics compared to conventional pooling methods. For instance, it delivers notable quantitative gains across different tasks including up to a 12.67% increase in mean intersection over union on the CamVid dataset for semantic segmentation and a 1.1% mAP50 enhancement on the real-world RTTS dataset for object detection.

AI doi: 10.3390/ai7060225

Authors: George-Andrei Dima Ilie Cosmin Bilțan Mirabela-Melinda Medvei Luciana Morogan

Relation extraction is an important task for structuring information from unstructured text. However, the Romanian language still lacks dedicated datasets and benchmarks for this task. To address this gap, we introduce RoRED, a Romanian relation extraction dataset built by combining two complementary data construction strategies: translating existing high-quality English resources and applying distant supervision to native Romanian Wikipedia data. We leverage a powerful open-source large language model to automatically translate English examples into Romanian. For the native subset, we align Romanian Wikipedia entities with Wikidata relations to obtain naturally occurring Romanian examples. To better reflect real-world relation extraction scenarios, we also introduce synthetic negative examples generated using existing Romanian named entity recognition models. Finally, we validate the dataset by fine-tuning and evaluating multiple baseline models. Our strongest model, LUKE-RoRED, achieves a macro-F1 score of 0.8744 on the RoRED test set, demonstrating that the dataset can support relation extraction for Romanian. Overall, RoRED provides a strong first native benchmark for Romanian relation extraction.

AI doi: 10.3390/ai7060224

Authors: Hendrik Möller Alina Dima Benjamin Keinert-Weth Robert Graf Matan Atad Johannes Paetzold Friederike Jungmann Rickmer Braren Florian Kofler Bjoern Menze Daniel Rueckert Jan S. Kirschke Hanna Schön

Thoracolumbar stump ribs are one of the essential indicators of thoracolumbar transitional vertebrae or enumeration anomalies. While some studies manually assess these anomalies and describe the ribs qualitatively, this study aims to automate thoracolumbar stump rib detection and analyze their morphology quantitatively. To this end, we train a high-resolution deep learning model for rib segmentation using nnUNet and achieve significant improvements over existing models (Dice score 0.997 vs. 0.779, p-value < 0.01). In addition, we employ a novel iterative algorithm and piecewise linear interpolation to estimate rib length, achieving a success rate of 98.2%. When analyzing morphological features, we show that stump ribs articulate more posteriorly at the vertebrae (−19.2±3.8 vs. −13.8±2.5 mm, p-value < 0.01), are thinner (260.6±103.4 vs. 563.6±127.1mm2, p-value < 0.01), and are oriented more downwards and sideways within the first centimeters in contrast to full-length ribs. We show that with partially visible ribs, these features can achieve an F1-score of 0.84 and an AUC of 0.98 in differentiating stump ribs from regular ones. We publish the model weights and masks for public use.

AI doi: 10.3390/ai7060223

Authors: Ibrahim Al Janabi Tyler Bland

Generative artificial intelligence (genAI) enables educators to build custom learning tools, but the feasibility and impact of educator-driven, AI-assisted development (“vibe coding”) in medical education remain unclear. This study describes the rapid development of a custom ECG learning application using Gemini 3.1 Pro, evaluates its association with exam performance using difference-in-differences (DiD) and triple-difference (DDD) analyses, and assesses student perceptions with the user version of the Mobile App Rating Scale (uMARS). The app was implemented at one WWAMI site (intervention) with five sites as controls; aggregate performance from two first-year medical student cohorts (E24 vs. E25) was analyzed, comparing ECG-focused (focal) to non-ECG (baseline) exam items. DDD effects were inconsistent across exams, with no overall pooled effect on focal performance relative to baseline versus controls. In contrast, students rated the app highly (overall uMARS 4.57/5), particularly for quiz customization and waveform annotations. These findings support the feasibility of rapidly building and deploying tailored educational tools via genAI-assisted workflows and suggest strong perceived usability and acceptability among students. However, the study did not demonstrate a definitive short-term learning effectiveness effect on exam performance. Vibe coding is therefore positioned as a practical model for faculty-driven, context-specific educational innovation that requires further evaluation across broader implementations.

AI doi: 10.3390/ai7060222

Authors: Sana Amreen Ahmed Khairy Fakeha Masood Ngan Chu Anju Paudel Abdelrahman Aly Mohamed Ayantoyinbo Oluwabusayomi Yossef Alnasser

Background: Radiology underpins diagnosis and treatment across pediatrics, yet most artificial intelligence (AI) tools are developed for adults and validated on adult datasets only. Of more than 200 AI systems cleared by the United States (U.S.) Food and Drug Administration (FDA), only about 3% include pediatric validation. Because children differ from adults in anatomy, physiology, pathology, epidemiology, and imaging protocols, adult-trained models often perform sub-optimally in pediatric settings. Methods: A narrative review of peer-reviewed literature from 2000 to 2025 was conducted using PubMed, MEDLINE, Google Scholar, and Scopus. Studies involving AI applications in pediatric X-ray, ultrasound, computed tomography (CT), magnetic resonance imaging (MRI), echocardiography, and point-of-care ultrasound with quantitative performance metrics were included. Findings were synthesized by imaging modality, clinical task, and differences between high-income countries (HICs) and low- and middle-income countries (LMICs). Results: AI demonstrated strong performance across multiple pediatric imaging tasks. In X-ray interpretation, AI detected fractures with area under the curve (AUC) values up to 0.96 (sensitivity, 90.8%; specificity, 88.7%). Pneumonia classification achieved 76.5% accuracy, and foreign body aspiration detection showed 95.3% specificity in HICs. In ultrasound, AI improved junior sonographers’ detection of intussusception (AUC 0.857 to 0.966) and reduced scan time by more than 50%. AI-assisted bone age estimation achieved a mean error of 0.39 years. In echocardiography, AI-derived ejection fraction showed excellent agreement with experts’ interclass correlation coefficient (ICC 0.983), and AI support improved atrioventricular septal defect detection (84.4% to 86.5%). In MRI, the use of AI enhanced lesion detection and supported quantitative analysis. Deep-learning models trained on routine T1- and T2-weighted sequences predicted liver stiffness across multi-site datasets, while advanced neuroimaging pipelines improved the identification of subtle epileptogenic lesions that are often missed on conventional pediatric MRI. However, adult-trained models showed limited generalizability to children. Still, excluding children under the age of two years improved the reading accuracy of pediatric chest X-rays (CXRs) by adult-trained models from 88% to 97%. AI faces challenges beyond the development of age-specific models. Substantial heterogeneity, limited pediatric-specific datasets, and unresolved medicolegal responsibility further restrict adoption worldwide. Challenges are amplified in LMICs, where unstable electricity, limited radiology resources, weak digital infrastructure, and scarce pediatric providers limit implementation. Additionally, many large language models underperform and lack inclusive algorithms suitable for pediatric radiology in many LMICs. Conclusions: AI can enhance diagnostic accuracy, efficiency, and access to pediatric imaging, particularly in resource-limited settings, through task-shifting and decision support. However, it cannot replace pediatric radiologists as of today. Safe adoption requires pediatric-specific model development, standardized validation metrics, diverse datasets that include LMIC populations, stronger digital infrastructure, robust radiologist training in AI capabilities, and the establishment of clear guidelines and medicolegal policies.

AI doi: 10.3390/ai7060221

Authors: Veronica Manescu (Paltanea) Iosif-Vasile Nemoianu Gheorghe Paltanea Iulian Antoniac Aurora Antoniac Alexandru Streza Gabriel Cristescu Costel Paun Adrian-Vasile Dumitru

Background: Biodegradable implants manufactured from Mg-based alloys are one of the most commonly used in orthopedics. However, their overall clinical acceptance is influenced by their fast corrosion speed and hydrogen emission. Based on an innovative manufacturing route previously described, this study introduces a preliminary proof-of-concept for a Gemini-assisted Digital Twin (Gemini-DT),which is an AI-augmented in silico framework designed to consider a MgF2 conversion coating on the implant surface and to model the synchronization of the degradation process with new bone formation. Methods: Based on the integration of experimental data for Mg-Nd and Mg-Zn alloys and by considering the implant geometry and coating formation, we developed, in collaborative work with LLM Gemini 1.5 Flash (Google), a four-module cognitive framework (surface thermodynamic synergy (Module 1), degradation analysis and alloy extract concentration management (Module 2), micro-channel fluidics and mechanical stability (Module 3), and bio-mechanical synchronization and regenerative evaluation (Module 4)) to evaluate simulated implant behaviors). Results: Using a 10,000 iteration Monte Carlo stability simulation, the model demonstrated a potential 12% reduction in false-negative design screening errors compared to rigid rule-based systems, achieving strong internal decision consistency in sustaining the mandated parametric compliance window. Computational verification supports the projected biocompatibility trends of Mg-Zn alloys, as previously demonstrated in our in vivo studies. Conclusions: Our research leads to a consistent computational architecture dedicated to Mg-based implants and offers a robust platform for virtual design and optimization. These observations suggest that the developed model can recover viable designs, whereas traditional linear models may reject them.

AI doi: 10.3390/ai7060220

Authors: Miguel G. Juarez Jaime Cerda Alejandro Zamora-Mendez Jose Ortiz-Bejar Juan Carlos Silva-Chavez

Modern power systems are becoming increasingly complex due to the rapid integration of renewable energy sources, the widespread use of nonlinear power-electronic devices, and the deployment of microgrids operating in parallel with conventional power grids. These evolving conditions intensify the occurrence of diverse and highly complex power quality disturbances (PQDs), demanding accurate and computationally efficient monitoring strategies. This paper presents a novel multi-stage hierarchical framework for PQD detection and classification, comprising an initial training stage with a dedicated 1D Convolutional Neural Network (1D-CNN), a transfer learning stage, and a subsequent fine-tuning stage. The proposed approach operates directly on raw voltage waveforms, eliminating the need for any signal preprocessing, as the CNN performs internal feature extraction. The framework is evaluated using a comprehensive dataset that includes synthetic signals, Matlab/Simulink (version R2022a) time-domain simulations, and real voltage sag events. Additionally, up to 29 types of disturbances, including complex multi-event combinations defined by the IEEE-1159 Standard, are generated using the PQ-SyDa toolbox. The proposed model achieves an F1-score of 97.8% using a three-cycle analysis window and further improves to 98.86% when five cycles are used. These results highlight the robustness and generalization capability of the proposed approach for the real-time PQD monitoring task in modern electrical networks.

AI doi: 10.3390/ai7060219

Authors: Priyadarshini Raghavendra Manob Jyoti Saikia

This review examines the evolution and architectural foundations of agentic artificial intelligence (AI), with a focus on collaborative multi-agent systems for complex task execution. The paper analyzes the core components, agent architectures, coordination mechanisms, application domains, and deployment challenges that enable autonomous reasoning and decision-making in real-world environments. To complement the survey, a comparative cryptocurrency market analysis case study is conducted using CrewAI, LangChain, and LangGraph focusing on workflow orchestration characteristics such as tool invocation, task transitions, orchestration depth, and memory integration. The findings are further supported by evidence from real-world financial applications reported in the literature, indicating productivity gains of 50–80% in financial data tasks and up to 20% improvement in stock prediction accuracy, highlighting the growing impact of multi-agent AI systems in market intelligence. The study highlights how architectural design choices influence reasoning continuity, coordination behavior, scalability, and system reliability, providing practical guidance for the design and deployment of agentic AI systems in complex, data-intensive domains.

AI doi: 10.3390/ai7060218

Authors: Yusaku Fujii

Objectives: This paper presents a principle and framework for generating social trust in AI outputs as an institutional structure rather than an ethical declaration. Sound technical design alone does not guarantee the institutional trust required to establish social measurement. What is needed is not a declaration of trust but the construction of an infrastructure that supports it. Methods: First, the Extended Kelvin Principle is derived by prepending to Kelvin’s measurement–understanding–control chain the links “no social trust without trust infrastructure; no legitimate social measurement without social trust.” Infrastructure-scale trust requires not declarations but verifiability, recordability, and auditability. Just as GUM and calibration infrastructure underpin trust in measured values, AI output governance requires GLO, a common language for expressing output legitimacy, implemented by a VRAIO-type infrastructure. GLO treats an output candidate as a “claim” and declares the rule-conformity of its purpose and content as a legitimacy confidence L, derived from a fact-based argument accompanied by a legitimacy budget. Results: VRAIO integrates declaration, rule verification, tamper-resistant recording, and independent auditing. A sealed, deterministic verifier makes L reproducible: computational falsity is caught by re-computation, factual falsity by checking authoritative records, and severe sanctions render false declaration irrational. Conclusions: GLO is not a mere AI version of GUM but a common language for an underdeveloped domain, whose effectiveness depends on connection to an enforceable output-governance infrastructure.

AI doi: 10.3390/ai7060217

Authors: Feng Guo Xunhuang Wang Fumin Zou Lei Zou Tao Fang Xueming Wu Haocai Jiang Jianqing Weng

Accurate traffic flow prediction is fundamental to Intelligent Transportation Systems (ITSs), critical for transportation management and logistics. Despite advances in spatio-temporal prediction methods, existing approaches suffer from two key limitations: (i) multi-scale fusion methods inadequately capture hierarchical constraints between cross-scale features, and (ii) models rely on single spatio-temporal views, neglecting multi-source relationship complementarity. To address these issues, we propose ST-MAFNet, a spatio-temporal multi-scale adaptive fusion network comprising three key components, specifically, a Cross-Scale Hierarchical Anchoring strategy (CSHA) that anchors short-term predictions with multi-scale temporal patterns to mitigate noise; a Dual Spatial Perception Module (DSPM) that learns node heterogeneity and dynamic correlations through node embeddings and adaptive graph attention; and a Spatio-Temporal Adaptive Fusion Module (STAFM) that captures time-varying connectivity by integrating multi-scale temporal features with multi-source spatial relationships. Experiments on four real-world datasets demonstrate that ST-MAFNet is particularly effective for short-term traffic forecasting. Compared with the best previously reported MAE results, ST-MAFNet reduces MAE by 2.95%, 1.43%, 1.25%, and 0.37% on PEMS03, PEMS04, PEMS07, and PEMS08, respectively, and achieves the best or second-best performance on most evaluation metrics.

AI doi: 10.3390/ai7060216

Authors: Ainhoa Serna Jon Kepa Gerrikagoitia Juan de Oña

This study evaluates zero-shot domain transfer for multilingual sentiment analysis in sustainable urban mobility using XLM-RoBERTa, a transformer pre-trained on social media data and applied to transport reviews without task- or domain-specific fine-tuning. Starting from a manually annotated English corpus of 375 transport-related user reviews, we created sentence-aligned translations in Spanish, French, German, and Italian, yielding a multilingual evaluation dataset of 1875 instances. Results show that the model assigns consistently high confidence to polarized content (mean: 0.76–0.85) and lower confidence to neutral or ambiguous expressions (0.58–0.65), with visible but preliminary cross-lingual variations that require further linguistic validation. Confidence scores are treated as diagnostic indicators of model certainty, not as evidence of correctness or calibration. A qualitative analysis of 113 categorized low-confidence predictions identifies six recurring linguistic patterns associated with model uncertainty (led by translation drift, mixed sentiment, and idiomatic expressions) with substantial inter-annotator agreement (κ = 0.664). By releasing the annotated multilingual dataset and code publicly, this work provides a reproducible exploratory evaluation framework for annotation-scarce, domain-specific multilingual NLP.

AI doi: 10.3390/ai7060214

Authors: Tea Krčmar Dina Šabanović Mirko Köhler Ivica Lukić

Heterogeneous anomaly detection ensembles improve robustness by combining complementary detectors, yet existing approaches often rely on heuristic detector selection, fixed contamination assumptions, and equal weighting. We investigate whether compact ensembles of complementary detectors can outperform substantially larger heterogeneous configurations through diversity-aware weighting and adaptive contamination estimation. Experiments on 22 benchmark datasets show that a compact ensemble of four complementary classical detectors outperforms an eleven-detector ensemble containing deep learning components, while requiring only 13.8% of the computational cost. Across the benchmark, the proposed ensemble variants achieve strong rankings while remaining competitive with the strongest individual detectors (Friedman χ2=71.58, p<0.001). These findings suggest that detector diversity, rather than ensemble size or architectural complexity, is the primary driver of robust unsupervised anomaly detection performance in resource-constrained environments.

AI doi: 10.3390/ai7060215

Authors: Shunping Niu Kuo Chi Ting Su Yongqin Yang Jiabao Gao

Conventional recommender systems often rely on shallow collaborative signals, which limits their performance under sparse and popularity-skewed conditions. To address this, we propose a knowledge-aware framework that combines an item hypergraph induced by user interaction histories, a top-k user similarity graph, and one-hop, relation-aware knowledge-graph aggregation. The hypergraph branch learns high-order item co-occurrence representations, which are aggregated into initial user vectors and then refined through user similarity propagation. On the item side, user-conditioned relation attention aggregates one-hop KG neighbors to produce semantic item representations. User and item representations are fused by an MLP scorer, and a lightweight popularity-aware post-scoring adjustment can optionally be applied to moderate head-item dominance. Experiments on MovieLens-1M, Last.FM and Book-Crossing show strong performance among the compared baselines in AUC, ACC, and Recall@K.

AI doi: 10.3390/ai7060213

Authors: Alberto Gudiño-Ochoa Eduardo Ruiz-Velázquez Julio Alberto García-Rodríguez Raquel Ochoa-Ornelas Sofia Uribe-Toscano

Non-invasive blood glucose estimation from exhaled breath has been proposed as a painless alternative to repeated capillary measurements; however, performance evaluation remains challenging in small-sample settings. This study investigates the estimation of blood glucose from human breath using volatile organic compound (VOC) signals acquired with an electronic nose. Responses from three metal-oxide sensor channels sensitive to CO, alcohol, and acetone were collected from 58 individuals, with one measurement per subject, and analyzed using strictly patient-level five-fold cross-validation, in which test folds comprised only real subjects. Two experimental factors were examined. First, model performance was evaluated with and without an additional interpretable alcohol–acetone log-ratio capturing relative variation between compounds. Second, model training was performed using either real data only or fold-wise tabular synthetic augmentation generated via a Gaussian copula fitted exclusively on training subjects, while evaluation remained strictly real-only. Under real-only training, classical machine learning models achieved the lowest prediction errors (approximately 6–7 mg/dL), whereas under synthetic augmentation FTTransformer was the best-performing deep learning model. This findings should be understood as a constrained proof-of-concept analysis rather than as evidence of diagnostic capability or clinical readiness.

AI doi: 10.3390/ai7060212

Authors: Panagiotis Douros Konstantinos Kasaras Konstantinos Milioris

Background: The rapid advancement of generative artificial intelligence is transforming music composition from an exclusively human-centric activity into a hybrid human–algorithmic domain. Despite technological progress and growing commercial integration, consumer acceptance of AI-generated music remains empirically underexplored. Methods: This study formulates and empirically evaluates a multidimensional theoretical model integrating nine frameworks—including UTAUT2, parasocial interaction theory, anthropomorphism theory, authenticity theory, and innovation resistance theory—through a quantitative cross-sectional survey of 466 young adults aged 17–28. Confirmatory factor analysis and multiple regression analysis (with robust standard errors) were employed. Results: The model explained 63.6% of the variance in behavioral intention (R2 = 0.636). Five constructs emerged as significant predictors: hedonic motivation (β = 0.136, p = 0.017), parasocial relationships (β = 0.121, p = 0.002), social influence (β = 0.126, p = 0.002), performance expectancy (β = 0.102, p = 0.019), and innovation resistance (β = −0.089, p = 0.029). Authenticity concerns, ethical AI concerns, anthropomorphic perceptions, and technological substitution fears were non-significant in the multivariate model. Conclusions: Young consumers’ acceptance of AI-generated music is primarily driven by experiential, social, and relational factors rather than ethico-cultural concerns. These findings have substantive implications for creative industries navigating algorithmic cultural production.

AI doi: 10.3390/ai7060211

Authors: Gulnaz Zakariya Aiman Moldagulova Nor’ashikin Ali

Point-of-sale (POS) consumer credit represents the most rapidly expanding retail-lending channel within the emerging Eurasian markets, necessitating a stringent operational framework for the underwriting model: the decision must be rendered within a mere few hundred milliseconds during the in-store checkout process, while the inputs are constrained to what the application XML is capable of conveying. This research endeavors to develop, internally validate, and operationally delineate a hybrid, explainable artificial intelligence framework aimed at POS credit scoring within the production portfolio of Kazakhstan’s largest second-tier bank. The architectural framework is delineated along two orthogonal dimensions—client tenure and decision-making channel—resulting in the formulation of three distinct production models: two transparent Weight of Evidence–Logistic Regression scorecards tailored for the real-time channel, and one isotonically-calibrated stacked ensemble (comprising LightGBM, CatBoost, and a three-layer neural network) designated for the batch channel. The selection of hyperparameters was conducted utilising Bayesian optimization within the context of stratified five-fold cross-validation. The digital scorecards achieve an area under the receiver operating characteristic curve (AUROC) of 0.847 and 0.835, whereas the offline ensemble enhances performance to an AUROC of 0.918, accompanied by a Kolmogorov–Smirnov statistic of 0.682 and a Gini coefficient of 0.836. The population stability indices persist below the threshold of 0.07, while isotonic recalibration effectively reduces the Brier score by 18%. Furthermore, an extensive examination of fairness demonstrates variations in approval rates within a margin of ±1.2 percentage points—and equalised-odds gaps below 1.5 percentage points in the true-positive rate and 0.7 percentage points in the false-positive rate—across multiple demographic factors such as gender, age, and distinctions between urban and rural classifications, thus establishing an artificial intelligence framework that is both regulatorily compliant and interpretable, aligning with the directives set forth by the Agency of the Republic of Kazakhstan for Regulation and Development of the Financial Market.

AI doi: 10.3390/ai7060210

Authors: Ghaith Hussein Jeremy Smith Waleed Al-Nuaimy

Person Re-identification (Person Re-ID) is essential in surveillance and security. Traditional image processing methods often struggle to identify individuals accurately due to the sensitivity to occlusions and limited discriminative capability of the global feature representation. To address these challenges, this study proposes a deep-learning architecture for Person Re-ID, termed Dynamic Part-Based Fusion (DPBF), which integrates the Salient Part Discrimination (SPD) and the Adaptive Feature Integration and Contextual Fusion (AFICF) frameworks within a unified pipeline. The SPD module enhances representation learning by emphasizing discriminative body regions through an attention-guided part-based mechanism guided by human parsing information. The AFICF component performs the correlation-aware integration of localized part-specific features and global contextual features, reducing redundancy and improving discriminative feature representation. The proposed framework coordinates part-level feature extraction and correlation-aware integration within a unified pipeline to improve robustness under occlusion and appearance variations. Additional analyses demonstrate a stable performance across independent training runs, competitive computational complexity, and robustness under severe occlusion conditions through adaptive local–global feature integration. The method was evaluated on several Person Re-ID datasets, including Occluded-ReID, Market-1501, DukeMTMC-ReID, Occluded-Duke, P-DukeMTMC-ReID, and CUHK03-Labeled. The experimental results demonstrate a competitive performance compared with existing methods, while additional reproducibility, computational-complexity, and occlusion-stability analyses further validate the robustness and practical applicability of the proposed framework. Specifically, DPBF achieves a 10.6% increase in Rank-1 accuracy and a 16% improvement in mAP over the closest competitor on the Occluded-ReID dataset.

AI doi: 10.3390/ai7060209

Authors: Carmela Rita Balistreri Laura Asta Sabrina Nocerino Dario Tarantino Calogera Pisano Diego Gallo Salvatore Pasta

Aortic aneurysms (AAs), both abdominal and thoracic, remain one of the most lethal cardiovascular diseases, with increasing prevalence and incidence, especially in sporadic forms, in our populations, primarily represented by elderly individuals. The high mortality risk is primarily due to delayed management, although their management has shown progress, particularly regarding imaging techniques that facilitate diagnosis and otherwise complex surgical procedures. This is due to the clinical decision-making approach, which, unfortunately, is still based, according to guidelines, on the maximum aortic diameter. The maximum aortic diameter, as repeatedly emphasized, fails to capture the biological and biomechanical complexity of these pathological conditions, which are influenced, among other things, by highly individual factors (genetics, gender, lifestyle, etc.). Thanks to the advent of network medicine and omics sciences, diverse and complex clinical, imaging, and biomarker datasets are available. Artificial intelligence (AI) could process this data to facilitate the complex management of aneurysms and accurately predict risk. AI could prove an excellent tool for aneurysm management, improving risk prediction and radically transforming the way we understand, monitor, and manage aneurysm patients, despite some limitations, as well as improving its therapeutic applications towards personalized strategies. This narrative review provides an overview of these aspects based on current evidence.

AI doi: 10.3390/ai7060208

Authors: Haocheng Hu Tianzhen Wang Haoran Wang Yassine Amirat

With the increasing penetration of renewable energy systems, multilevel inverters have been widely adopted to meet the growing demand for high-power and high-quality energy conversion. Among various multilevel topologies, cascaded H-bridge multilevel inverters (CHMIs) are particularly attractive due to their modular structure and improved output voltage quality. However, the increased number of power semiconductor devices and switching states significantly complicates fault diagnosis under practical operating conditions. Currently, most existing neural networks for fault diagnosis are manually designed based on domain expertise. This may limit their adaptability to task-specific fault patterns as well as edge-side inference performance. To reduce the dependence on manually designed diagnostic networks, an edge-oriented fault diagnosis framework based on differentiable architecture search (DARTS) is proposed to automatically design task-specific diagnostic networks. A simplified special cell search strategy is adopted to improve search efficiency and facilitate practical deployment. The searched architectures are lightweight and suitable for deployment on edge platforms. The experiments show that the proposed method achieves an average diagnostic accuracy of 99.44% on the test set under the RL load of (7Ω,6mH). Furthermore, the searched model contains only 0.2417 M trainable parameters, and edge deployment experiments on the Jetson Orin Nano platform show low-latency inference capability.

AI doi: 10.3390/ai7060207

Authors: Ming Lei You Fu Ruyun Fu Shengliang Fang Youchen Fan

Accurate radio map (RM) construction is essential for 6G wireless network optimization, yet faces significant challenges owing to sparse real-world measurements and dynamic environmental obstacles. This paper presents RMF, a novel single-step generative model based on mean flow matching that enables direct mapping from a noise prior to the target radio map distribution in a single forward pass, eliminating the iterative inference required by diffusion-based approaches. The proposed model integrates a multi-feature U-Net backbone with four specialized branches that extract and fuse building-layout features—via dual-path frequency and spatial-domain processing—base station distance fields, graph neural network-encoded sparse measurements, and dynamic obstacle representations, all injected through multi-scale cross-attention. Evaluations on the RadioMapSeer benchmark show that RMF attains the best RMSE and PSNR among the compared methods, with RMSE between 0.0136 and 0.0162 and PSNR between 36.52 and 37.24 dB, SSIM within 0.012 of the leading diffusion baseline, and an order-of-magnitude reduction in per-sample inference time. In the challenging zero-measurement scenario, RMF achieves PSNR gains of 1.45–1.55 dB over competing methods in both static and dynamic environments. The single forward-pass design yields inference times of 0.05 s, making RMF a promising candidate for real-time 6G applications such as coverage optimization and dynamic spectrum management, subject to validation on field-measured data in future work.

AI doi: 10.3390/ai7060206

Authors: Pilar Herrero-Martín Álvaro García-Ballestero

Artificial intelligence systems increasingly mediate high-stakes human activities, yet urban navigation remains highly challenging for blind and visually impaired individuals. Although digital navigation technologies have significantly improved route planning and accessibility, many existing systems still rely on generic interaction paradigms that insufficiently account for cognitive load, contextual uncertainty, and the adaptive needs of vulnerable users. This challenge highlights the importance of Human-Centred AI approaches capable of supporting not only functional accessibility, but also cognitively sustainable and trustworthy interaction. This paper introduces LAZAR, a human-centred adaptive AI framework for accessible urban mobility grounded in a user-centred design methodology and formalised through a structured Software Requirements Specification. Rather than focusing exclusively on route optimisation, LAZAR approaches assistive navigation as an adaptive human–AI interaction problem in which instructional granularity, interaction frequency, and feedback mechanisms are designed to support user autonomy and situational awareness whilst limiting unnecessary cognitive burden. The proposed framework integrates high-fidelity prototyping, accessibility-oriented interaction modelling, and a modular multi-agent architecture intended to support adaptive and personalised guidance. Central to the approach is a cognitive load-aware interaction layer designed to regulate the presentation and timing of navigational assistance according to user needs and contextual conditions. The proposed multi-agent architecture is presented as a modular design framework whose interaction principles and interface logic were partially operationalised in the evaluated prototype. The complete integration of all adaptive coordination mechanisms, together with large-scale real-world validation, remains part of ongoing and future development work. This work contributes a structured methodology for the design of adaptive assistive AI systems that integrates accessibility requirements, human-centred interaction principles, and cognitively informed guidance strategies. A formative usability evaluation involving eleven visually impaired participants provides preliminary empirical evidence regarding usability, accessibility, and perceived usefulness of the proposed interaction model. The framework establishes a foundation for future research on inclusive and adaptive AI-based navigation systems in urban environments.

AI doi: 10.3390/ai7060205

Authors: Masoud Kamali Behnam Atazadeh Abbas Rajabifard Yiqun Chen

3D scene understanding in industrial environments is crucial for effective operation and maintenance (O&M) and asset monitoring. However, accurate asset detection and localisation face significant challenges due to asset diversity and scene complexity in these environments. Existing learning-based methods rely heavily on labelled training datasets, which are limited for industrial settings due to asset variability and intricate geometries. To address these challenges, this paper presents a novel framework for industrial asset detection and localisation without requiring labelled training datasets, using only point cloud data. Experimental results demonstrate the competitive performance of the proposed framework, achieving an average precision at 25% intersection over union (AP25) of 48.13% and an AP50 of 34.98%, significantly outperforming state-of-the-art (SOTA) methods. This framework can be employed to generate 3D digital models of brownfield industrial plants that lack up-to-date spatial information, serving as a foundational spatial layer for the development of digital twins within industrial environments.

AI doi: 10.3390/ai7060204

Authors: Maya Sultan ALGhafri Imran Khan Abdelhamid Abdesselam

The exponential growth of social media has revolutionized global communication, enabling instant idea exchange and transforming information sharing into a worldwide phenomenon while simultaneously accelerating the spread of abusive and hateful content that threatens online harmony and poses a serious risk to online community integrity and public trust. Although supervised deep learning approaches achieve impressive accuracy for hate speech detection, they remain fundamentally reliant on extensive annotated corpora, and their lack of interpretability makes them insufficient for transparent and scalable real-world hate speech detection. This study presents a category-oriented unsupervised architecture for English hate-speech detection and classification that substantially reduces reliance on large labeled datasets by requiring only minimal supervision (10% of labels for post hoc cluster interpretation), ensuring transparency and a high degree of semantic interpretability. We introduce an unsupervised Subspace Weighting Co-Clustering framework that uses HateBERT-driven contextual embeddings, enabling simultaneous interpretable feature weighting and semantic understanding for robust hate-speech detection. The obtained embeddings are further structured using the Subspace Weighting Co-Clustering approach, which enables the unsupervised discovery of latent subspaces and the organization of tweets into semantically coherent hate categories. The comprehensive evaluation shows that the framework achieves superior accuracy over existing methods, providing a more robust and effective mechanism for digital platforms to identify and mitigate hate speech and promote safer online interactions.

AI doi: 10.3390/ai7060203

Authors: Roxana Romero Hugo Ordoñez Carlos Cobos

This systematic literature review, guided by Kitchenham and Charters and following PRISMA 2020, analyzes explainable artificial intelligence (XAI) approaches for multiclass classification models, with an emphasis on explaining class differentiation and the relationship between feature contributions and changes in prediction probabilities. The protocol was defined in advance, but it was not preregistered. Searches were conducted in Scopus, Web of Science, SpringerLink, and ScienceDirect (2020–2025) using PICOC-based strings and explicit eligibility criteria. Following the PRISMA flow, 108 studies were included out of 8697 identified records. The most frequently reported approaches are based on feature contribution/attribution (e.g., SHAP, LIME, CAM, and Grad-CAM) and counterfactual explanations, with prominent applications in medicine, finance, and cybersecurity. Although several works analyze local contributions and, separately, probability variations, the synthesis reveals a methodological gap: there is a lack of a formal and explicit instance-level framework that quantitatively connects the differential contribution of a feature (e.g., SHAP values) with the probability variation between classes to explain class differentiation. In practical terms, such a linkage enables instance-level justification of why a model favors class A over a competing class B, improving traceability and decision support in high-stakes settings (e.g., differential diagnosis and risk assessment). These findings point to future directions toward more rigorous comparative local explanations in multiclass settings.

AI doi: 10.3390/ai7060202

Authors: Kyungman Cha Sohee Lee Jaekwang Shin Jee Yong Lim

Undertriage—the misclassification of acutely ill patients into low-acuity triage categories—is a persistent patient safety concern, and prior machine learning approaches restricted to vital signs have yielded modest predictive performance. We hypothesized that this ceiling reflects feature restriction rather than an inherent predictive barrier. In this retrospective cohort study of 10,792 adult patients (age ≥ 18) initially triaged as Korean Triage and Acuity Scale (KTAS) level 4 or 5 across two tertiary academic centers during 2025, the primary outcome was triage reclassification—change from initial KTAS 4/5 to final KTAS 1–3 (n = 941; 8.7%). Five nested feature sets of increasing breadth were compared using logistic regression (LR) and gradient-boosting classifiers (GBC). Calibration (slope, intercept, Brier score), sensitivity/specificity/positive and negative predictive values at operating thresholds of 3%, 5%, and 10%, and decision-curve net benefit were evaluated on a held-out test partition. NEWS alone yielded an AUROC of 0.58, whereas the full triage-time panel (Set E; 43 features) achieved a GBC AUROC of 0.72 (95% CI 0.68–0.76; 5-fold CV 0.73 ± 0.02) and an AUPRC of 0.23, approximately doubling the NEWS baseline (0.12). The model was well calibrated, with a Brier score of 0.075, a calibration slope of 0.85 (95% CI 0.70–1.01), and an intercept of −0.30 (95% CI −0.65 to 0.07); both intervals included the ideal values of 1 and 0, indicating that predicted probabilities can be interpreted as approximate absolute event likelihoods. At a 5% operating threshold, sensitivity was 0.79, capturing 79% of reclassifications while flagging 53% of the cohort. Decision curve analysis demonstrated positive net clinical benefit across thresholds of 3–20%, exceeding both a vital-signs-only model and the treat-all/treat-none baselines. Feature importance analysis identified pain score, onset-to-arrival time, heart rate, systolic blood pressure, and age as the dominant predictors. Contextual variables routinely documented at triage—particularly pain score and onset-to-arrival time—together with heart rate and systolic blood pressure form a discriminative composite that exceeds the performance of vital-signs-only models in the KTAS 4/5 subpopulation. The resulting model is well calibrated and provides positive net clinical benefit across the 3–20% threshold range, supporting its potential role as a secondary screening flag for low-acuity patients warranting clinician re-review. External validation in independent cohorts is needed before clinical deployment.

AI doi: 10.3390/ai7060201

Authors: Qianjin Wang Xiaoxu Xu Xin Li Luochenxi Mou Can Cui Liting Zhai Ronald Man Yeung Wong Wing Hoi Cheung Ning Zhang

Sarcopenia is a progressive, age-related skeletal muscle disorder that serves as a driver of frailty, falls, and mortality in older adults. Despite the recent paradigm shift introduced by the latest sarcopenia consensus, which emphasizes early, proactive detection, sarcopenia cases frequently evade traditional screening due to inherent diagnostic bottlenecks and resource limitations. Artificial intelligence has emerged as a transformative solution to dismantle these barriers across the entire continuum of sarcopenia care. This review explores the rapid evolution of artificial intelligence, beginning with automated opportunistic screening that extracts prognostic musculoskeletal data from routine imaging and electronic health records, advancing toward high-precision multimodal assessment architectures. Beyond initial assessment, artificial intelligence is actively restructuring longitudinal care by the integration of ubiquitous wearables, Large Language Models, and computer vision, enabling dynamic exercise prescriptions and real-time kinematic postural correction for sarcopenia rehabilitation. Realizing this potential requires the medical community to confront urgent clinical barriers, including multi-center validation, semantic interoperability, health economic justification, and the strict preservation of human-centric ethics. By addressing these challenges, this review provides a definitive roadmap for embedding artificial intelligence across the entire sarcopenia care pathway, transforming isolated instances of opportunistic screening into a unified ecosystem for intelligent, proactive rehabilitation.

AI doi: 10.3390/ai7060200

Authors: Suresh Rasappan Francis Saviour Devaraj Ahamed Nishath Syed Dilwar Islam Mazumder Wardah Abdullah Al Majrafi

Sign language understanding for automated vehicles sits at the intersection of accessibility, intelligent transportation, and safety-critical human–machine interaction. The existing sign-language recognition systems are largely confined to controlled environments, limiting their utility in mobility scenarios characterized by lighting variation, motion blur, and partial occlusion. This paper proposes STCM-HVNet, a safety-aware hybrid multimodal architecture integrating four components: a spatial visual encoder, a MediaPipe-based pose encoder, a bidirectional LSTM temporal encoder, and a context-aware fusion and safety decision module. The architecture is formulated as a multi-task system that jointly predicts sign category, interaction intent, and urgency level, and incorporates confidence-aware rejection and fail-safe action mapping. Experiments are conducted on two Arabic sign-language resources. On the RGBArS image benchmark (31 classes, 7856 images), the proposed pipeline achieves a Top-1 accuracy of 45.38%, Top-3 accuracy of 75.15%, and Macro-F1 of 0.4479, outperforming LinearECOC, kNN-5, and Bagged Trees baselines. On the Arabic sign-language video benchmark (12 classes, 479 clips), the BiLSTM temporal encoder achieves a Top-1 accuracy of 93.15% and Macro-F1 of 0.9383, outperforming frame-aggregation (87.67%) and CNN-LSTM (89.04%) baselines. Ablation results confirm complementary contributions from the visual and pose branches. A safety-threshold analysis and a Monte Carlo dropout comparison demonstrate that the proposed safety decision/gating layer provides a controllable trade-off between prediction coverage and reliability.

AI doi: 10.3390/ai7060199

Authors: Mohammad Bilal Firoz Ashir Ahmed

Evaluating AI-assisted tools in healthcare human–computer interaction (HCI) presents methodological challenges when practical constraints limit sample sizes. Standard pooled statistical analysis can then produce misleading results, including Simpson’s Paradox, where aggregate trends contradict patterns observed within subgroups. This paper introduces a conditional Gaussian model framework that models each experimental condition separately rather than pooling all observations. Through a within-subjects evaluation of an AI-assisted UI/UX design tool for medical software interfaces (n = 4 professional designers), we demonstrate how pooled analysis produced a misleading negative correlation between design time and IEC 62366 compliance (the medical device usability standard; pooled r=−0.76, p=0.029, n=8), even though every designer achieved both faster times and higher compliance with the AI tool. Within-condition correlations were non-significant and inconsistent in sign, confirming the pooled association as an aggregation artefact rather than a within-designer trade-off. The conditional analysis surfaces experience-indexed differences: the less UI-experienced designer showed the largest time reduction (up to 92%), while the two high-AI-experience designers showed the largest automated proxy-compliance gains (+25 to +29 percentage points). Sample standard deviations were also lower in the AI-assisted condition than in the traditional condition for both outcomes (time: 20.0→11.3 min; compliance: 10.6→7.6 percentage points); at n=4 per condition, however, this difference in variance can neither be confirmed nor falsified, and we make no inferential claim about variance compression. A follow-up phase (n = 3) that adapted the tool’s scaffolding to designer experience yielded a bidirectional response, with the two high-AI-experience designers further reducing time and the less UI-experienced designer engaging more deeply with the design output. Because all participants completed the traditional condition before the AI-assisted condition, the study is interpreted as a sequentially unbalanced exploratory comparison, not as a counterbalanced causal test of tool effectiveness. We provide guidelines for healthcare HCI researchers facing sample-size constraints endemic to specialised domains.

AI doi: 10.3390/ai7060198

Authors: Catalin Anghel Andreea Alexandra Anghel Marian Viorel Craciun Adina Cocu Diana-Elena Vulpe Constantin Adrian Andrei Calina Maier Cristian Scheau Serban Dragosloveanu Romica Cergan

Large Language Models (LLMs) are increasingly used for rubric-based assessment, but reliable automated grading requires more than a single prompt-response step. This study presents GradeAgentOps, a verification-first framework for evidence-anchored LLM exam grading that combines strict grading contracts, deterministic verification and canonicalization, bounded semantic repair, optional memory modules, and provenance-aware logging. The framework is evaluated on a university-level dataset of 1000 short open-ended exam responses from 100 students across 10 questions, with annotations from two independent expert human graders. A controlled ablation protocol compares six configurations, including a rubric-only baseline and progressively stronger variants with repair and memory augmentation. Human–human agreement provides the reference context, with overall ICC(2,1) = 0.678 and QWK = 0.678 between the two graders. Using one expert grader as the operational reference, the FULL configuration achieves the strongest model-human agreement (MAE = 1.935, RMSE = 2.500, QWK = 0.652, Within ±2 = 0.667), with the consistency memory configuration with repair (C1) emerging as the closest alternative. The gains are not uniform: improvements are more pronounced on argumentative items than on technical ones, and pairwise comparisons show that consistency memory contributes more strongly than rubric memory alone. The stronger configurations also produce cleaner accepted outputs, with lower rates of evidence-related postprocess issues, although at a moderate operational cost. Overall, within this controlled single-course evaluation setting, the results show that reliable automated grading benefits not only from model capability, but also from pipeline design choices that promote verification, evidential coherence, and stable grading behavior.

AI doi: 10.3390/ai7060197

Authors: Junkun Hong Zhipeng Zhou Shiyu Song Peng Lan Junfeng Man

Session-based recommendation seeks to deliver personalized suggestions by decoding transient interaction sequences generated by anonymous users. Although graph neural networks have advanced this field by modeling pairwise item transitions, they fundamentally struggle to capture the complex, high-order dependencies inherent in real-world user behavior modeling. Consequently, while hypergraphs offer a natural mathematical solution for representing these multi-item relationships, existing approaches frequently overlook the localized structural semantics necessary to ground these abstract relations in physical browsing logic. To address these limitations, we introduce MoHyNet, a novel motif-guided hypergraph framework explicitly designed to capture both inter- and intra-session dependencies. By extracting localized hypergraph motifs, MoHyNet effectively decodes the recurring topological sub-structures and latent intentions behind anonymous interactions. Rather than treating hypergraphs merely as static representations of item co-occurrence, our approach utilizes these motifs as dynamic semantic filters to extract stable behavioral signatures from pseudo-sequential noise. To complement this intra-session modeling, we construct an augmented line graph that maps multi-hop dependencies across different sessions, employing neighborhood-aware convolutions to aggregate global collaborative context. A dual-view contrastive learning optimization is subsequently integrated to semantically align these intra-session structural signatures with the inter-session global context, ensuring a robust and holistic understanding of user intent. Extensive empirical evaluations on three real-world e-commerce datasets demonstrate that MoHyNet consistently outperforms state-of-the-art methods in session-based recommendation performance.

AI doi: 10.3390/ai7060196

Authors: Walter Bonke Mahlangu Taurai Hungwe Nomasonto Rapulenyane Somandla Ncube

Organometallic chemistry deals with the synthesis, structure, reactivity, and applications of compounds containing metal–carbon covalent bonds. In recent years, there has been a growing interest in predicting the catalytic activity of organometallics using machine learning. However, the major drawback in developing algorithms that can be used in predicting organometallic reactions is the availability of organometallic reaction data and organometallic filtering tools. The main aim of the current study is to develop organometallic reaction-filtering tools that are crucial for building accurate and effective ML models in organometallic chemistry. Random Forest (RF), K-Nearest Neighbors (kNN), Support Vector Classifiers (SVC), and Multi-Layer Perceptrons (MLP) were employed, using feature subsets selected via Permutation Feature Importance from Morgan fingerprints and MACCS keys. The results demonstrate that the MACCS-based MLP architecture provides the most reliable filtering performance, achieving a superior F1 score of 0.85, a Recall of 0.85, and a high AUC-ROC of 0.837. Furthermore, the MACCS-MLP exhibited the highest predictive confidence, yielding the study’s lowest Log Loss of 0.312. In contrast, while Morgan fingerprints paired with kNN offered a specialized “strict” filter with absolute Precision (1.00), the sparse dimensionality of circular fingerprints generally resulted in lower calibration for probabilistic models. These findings underscore that dense, fragment-based descriptors refined by data-driven feature selection are most effective for identifying complex organometallic motifs. This study successfully provides a validated methodology for building precise filtering tools, establishing a critical foundation for automated catalyst discovery and the expansion of effective machine learning applications in organometallic chemistry. The study is limited to only identifying organometallic reactions and cannot filter based on organometallic reaction types. Future studies should also explore integrating multiple feature representations to classify or cluster the identified organometallic reactions based on the reaction types.

AI doi: 10.3390/ai7060195

Authors: Christos Hadjichristofi Michail Tsilimigkounakis Georgios Sotiropoulos Vassilios Vescoukis

Beyond code generation, large language models (LLMs) are increasingly explored in software architectural tasks. However, it remains unclear to what extent LLMs can apply explicitly requested architectural patterns when provided with user-defined requirements. In this paper, we empirically evaluate the ability of multiple LLMs to generate specific architectural styles under controlled conditions. Models are prompted with specific requirements expressed in different ways and are instructed to generate architectures in four typical styles using the same single prompt and execution strategy. The authors assess the generated architectures with respect to specific evaluation criteria. The results show that, while LLMs can correctly apply simpler architectural patterns, performance decreases as architectural complexity and problem size increase. Model size and requirement representation appear to influence pattern adherence, whereas retrieval-augmented generation (RAG) produces mixed effects. The findings contribute empirical evidence regarding prompting strategies, requirement representations, RAG configurations, and diagram-as-code (DaC) representations for LLM-generated software architectures. The study also introduces a reusable experimental workflow intended to support future benchmarking and comparative evaluation in software architecture generation tasks.

AI doi: 10.3390/ai7060194

Authors: Roberto C. Sotero Jose M. Sanchez-Bornot

The Transformer architecture drives modern Artificial Intelligence (AI), yet the physical principles that may constrain self-attention training remain poorly characterized. We develop a thermodynamic framework for attention training, drawing on the established Boltzmann correspondence between softmax attention and equilibrium statistical mechanics, and we propose a First Law analogue that decomposes the training energy budget into a heat term (the entropic cost of ordering attention) and a work term (the gain in mutual information about the target). From this framework we derive a Landauer-type bound on learning, which states that the loss reduction during training is bounded below by the entropic cost of structuring attention against thermal noise. The bound is satisfied across all configurations tested: 625 grid points spanning three datasets on a compact Vision Transformer trained from scratch (MNIST, CIFAR-10, and OrganAMNIST), and ten temperatures on a pretrained ViT-Small fine-tuned on Food-101. Reusing the same physical principles at inference time, we show that the thermodynamic work performed by each input patch provides a quantitative, energy-based measure of feature importance that outperforms standard attention weights and Integrated Gradients on ImageNet across pretrained ViT-Small, ViT-Base, and ViT-Large (22M to 304M parameters). The result is an integrated diagnostic framework that links phase structure, training-time bounds, and inference-time attribution within a single empirically falsifiable thermodynamic apparatus.

AI doi: 10.3390/ai7060193

Authors: Gali Chelouche-Dwek Peter Fonagy

Background: Adults closest to children, including parents and caregivers, teachers, and therapists, are major determinants of child mental health outcomes. However, access to high-quality psychological training for these groups remains severely limited and inequitable. Artificial intelligence (AI) tools may offer a scalable, accessible, and low-cost route to training delivery. This review aimed to provide the first systematic synthesis of evidence on AI tools used to train caregivers, educators, and therapists/practitioners in psychological approaches relevant to child and adolescent mental health. Methods: A systematic review was conducted in accordance with PRISMA guidelines (PROSPERO: CRD420261336167). Five databases, MEDLINE, PsycINFO, Embase, Web of Science, and ERIC, were searched from inception to March 2026, supplemented by reference hand-searching and forward citation tracking. Studies were eligible if they evaluated an AI-based training tool used with adults in caregiving, educational, or therapeutic roles involving children or adolescents aged 0–18 years, delivered a defined psychological approach, and reported at least one training outcome. Owing to substantial methodological and outcome heterogeneity, findings were synthesised narratively, and meta-analysis was not undertaken. Results: Twenty-four studies from nine countries, published between 2019 and 2026, met inclusion criteria. Studies were grouped into caregiver training (Group A, 5 papers), educator training (Group B, 3 papers), and therapist/practitioner training (Group C, 16 papers). Identified AI modalities included natural language processing (NLP)-based chatbots, generative AI/large language model (LLM) systems, AI-integrated virtual reality (VR), and AI-based feedback and analysis tools. Feasibility and acceptability findings were generally positive across groups. However, the evidence base was limited by pervasive methodological weaknesses, including small samples, with most studies enrolling fewer than 30 participants, reliance on unvalidated self-report outcomes, and the absence of follow-up data beyond one month. Conclusions: AI tools show early promise as scalable approaches to psychological training, particularly for procedural skill acquisition and enhancement of practitioner self-efficacy. However, the current evidence base is insufficient to support claims of effectiveness. A structural credibility–accessibility paradox characterises the field: tools with the strongest controlled evidence are the least scalable, while the most accessible tools have the weakest empirical support. Adequately powered, independent randomised controlled trials (RCTs) using validated outcomes, active comparators, and follow-up extending over multiple months are needed across all three population groups.

AI doi: 10.3390/ai7060192

Authors: Hugo Martínez Ángeles Cesar Augusto Navarro Rubio José Gabriel Ríos Moreno Margarita G. Garcia-Barajas Roberto Valentín Carrillo-Serrano Mariano Garduño Aparicio José Luis Reyes Araiza Mario Trejo Perea

This study presents a systematic review of Artificial Intelligence (AI) in vehicular bridge engineering, covering design, monitoring, and lifecycle decision support. The objective is to identify, classify, and critically analyze the main AI methods applied across the bridge lifecycle, including Machine Learning (ML), Deep Learning (DL), Artificial Neural Networks (ANNs), and Optimization Algorithms (OAs). The review follows the PRISMA 2020 framework to ensure transparency and reproducibility, considering publications from 2018 to 2026. The results show that AI applications span the entire bridge lifecycle; however, current research is predominantly concentrated in Structural Health Monitoring (SHM), damage detection, inspection, and predictive maintenance, while design-oriented applications—such as optimization, surrogate modeling, and structural analysis—remain comparatively less developed. Importantly, SHM data serve as a key input for data-driven modeling, enabling design optimization, reliability assessment, and lifecycle decision support. Classical ML methods remain effective for structured datasets, whereas DL models, particularly convolutional and recurrent neural networks, dominate image-based and time-series applications. In addition, hybrid physics-informed AI approaches are emerging to improve model reliability and interpretability. The review also identifies key challenges, including data quality limitations, lack of standardized methodologies, limited integration with engineering design codes, and barriers related to trust, expertise, and regulatory frameworks. Overall, the findings highlight a shift toward integrated digital frameworks, including digital twins and multimodal data fusion, to support more reliable monitoring and lifecycle decision-making. This study provides a comprehensive synthesis of current developments and outlines future research directions toward more resilient and intelligent bridge infrastructure systems.

AI doi: 10.3390/ai7060191

Authors: Remon Das Sajib Debnath Tarek Kandil Md Uzzal Mia

Accurate forecasting of municipal electric vehicle (EV) charging demand is increasingly important for distribution system planning, charging infrastructure management, and demand-side operation. This study proposes a weather-aware Transformer–LSTM hybrid framework for spatio-temporal forecasting of EV charging load across municipal public charging stations. The proposed approach integrates multi-source information within a unified pipeline, including cyclic temporal encodings, multi-lag autoregressive features, rolling statistics, behavioral aggregates, and meteorological variables, while combining a Transformer encoder to capture long-range temporal dependencies with an LSTM decoder to model local sequential dynamics and nonlinear load patterns. The framework was evaluated using 211,324 charging sessions collected from eight New York City municipal charging stations between July 2021 and December 2025. Under controlled benchmarking against Simple RNN, standalone LSTM, and encoder-only Transformer models using identical preprocessing, feature engineering, and training settings, the proposed hybrid model achieved R2 = 0.9731, MAE = 62.71 kWh, RMSE = 94.21 kWh, and MAPE = 19.62%. Relative to the standalone Transformer, the proposed model reduced RMSE by 32.6% and MAPE by 34.5%. In addition, the model maintained strong forecasting performance across stations with heterogeneous demand profiles without station-specific retraining and remained robust across seasonal variations. These results demonstrate that the proposed framework provides a reproducible and scalable solution for municipal EV charging load forecasting in real-world urban environments.

AI doi: 10.3390/ai7060190

Authors: Jinru Han Yunho Han Jiyoung Kim Woo-Chan Park

Atmospheric scattering caused by fog degrades image quality and significantly reduces the reliability of computer vision systems. Existing dehazing studies have mainly evaluated dehazing performance using pixel-level metrics such as PSNR and SSIM. However, these metrics do not fully reflect the actual impact of dehazing on downstream object detection performance. Therefore, this paper treats image dehazing as a preprocessing step for object detection in foggy environments and analyzes its effect using standard object detection evaluation metrics. The experimental results demonstrate that, under three fog-density conditions, β=0.005, 0.010, and 0.020, images processed by the DL-U-Net-based dehazing method achieved higher mAP@0.5 values than the corresponding original hazy images, with relative improvements of +0.39%, +6.60%, and +13.37%, respectively. Furthermore, under the dense fog condition of β=0.020, Recall improved more substantially than Precision. These results indicate that, as fog density increases, dehazing preprocessing becomes more effective in restoring object structural information, reducing missed detections, and enhancing downstream object detection performance.

AI doi: 10.3390/ai7060189

Authors: Elissar Abdul Khalek Jeffrey Macias Itamar Shabtai

Artificial intelligence (AI) increasingly permeates organizational work, yet research on AI–human collaboration remains fragmented and lacks a unified structure. This study provides a comprehensive bibliometric mapping of AI–human collaboration by examining its intellectual foundations and emerging research fronts across multiple disciplines. Using document co-citation and bibliographic coupling analysis, the study examines how research on AI–human collaboration has evolved and where it is heading. Data were collected from the Scopus database. A total of 2178 primary documents and 15,078 secondary documents were retrieved and analyzed using VOSviewer (1.6.20) software to visualize the thematic interconnectedness. Results from document co-citation revealed five significant research clusters underlying AI–human collaboration research, including psychological and social foundations of AI; organizational applications of AI in higher education; ethical–cognitive foundations of generative AI; AI literacy and educational transformation; and behavioral foundations of AI adoption. The bibliometric coupling results identified four active research fronts: AI governance, ethics, and humanization; AI–customer relationship management (CRM) adoption, capabilities, and organizational performance; anthropomorphic AI and consumer emotional response; and AI conversational agents and consumer experience dynamics. These findings suggest a thematic shift from technology-centered automation toward collaborative and human-centered integration. The study contributes theoretically by synthesizing insights across organizational behavior, psychology, and information systems to clarify the intellectual structure of this emerging domain. It also outlines implications for leaders designing AI-enabled workplaces that prioritize collaboration, ethical alignment, and adaptive capacity.

AI doi: 10.3390/ai7060188

Authors: Rigel Mahmood Sarosh Patel Khaled Elleithy

While Vision Transformers (ViTs) have achieved state-of-the-art (SOTA) results in visual recognition, their scalability remains fundamentally constrained by the quadratic complexity of global self-attention. To address this, we present a linear complexity attention design employing dual-stream information propagation to enhance representational efficiency and structured feature aggregation. Our proposed DS2 attention acts as a versatile replacement for standard attention in various SOTA designs, such as Tokens-to-Token (T2T) and FasterViT. In our design, half of the attention heads perform left-to-right segmented information propagation in a Perceiver-style manner, while the remaining half of the heads perform right-to-left propagation. This bidirectional structured attention enables efficient long-range dependency modeling without the overhead of full global attention. To improve classification performance, we introduce a segment-level classification strategy in which each segment is associated with a summary token. The final prediction is produced via cross-attention between image tokens and these summary tokens, enabling hierarchical semantic comprehension. Extensive experiments demonstrate that the proposed attention design achieves on average 0.3% higher accuracy on the ImageNet-1K dataset, while offering improved information flow and higher efficiency across SOTA Vision Transformer designs.

AI doi: 10.3390/ai7060187

Authors: Dimitra G. Papadopoulou Panagiotis D. Michailidis

The purpose of this review is to present an overview of artificial intelligence methods for classifying paintings into the artistic movement to which they belong. To achieve this goal, a literature review of research articles from the 2014–2024 period was carried out. The search for scientific articles was carried out in the Scopus database. The initial search yielded 492 publications and after successive stages of screening and full-text evaluation, 39 articles were finally selected for detailed analysis. The review presents (a) the datasets used in the works, (b) the range of artistic movements examined and (c) the computational methods from machine learning to deep neural networks and transfer learning. Methodological issues are highlighted, such as class imbalance of the samples, dataset bias and the limitations of commonly used evaluation metrics. The general finding is that a variety of methodologies were applied, with an increasing use of deep learning and transfer learning models, which in many cases are reported as effective within specific datasets and experimental protocols. Finally, the review offers a taxonomy of methodologies and maps trends and research gaps in research on painting style classification over the last decade, while at the same time making suggestions for future research.

AI doi: 10.3390/ai7060186

Authors: Jelena Mitić Velibor Ilić Uroš Durlević Milan Mitić

Background/Objectives: Accurate land cover mapping from multispectral Sentinel-2 imagery is fundamental for environmental monitoring, efficient natural resource management, and spatial planning. While deep learning has become the dominant approach for semantic segmentation, the combined impact of spectral input selection and network architecture on cross-regional robustness remains insufficiently explored. This study systematically investigates multispectral land cover segmentation in Serbia and evaluates its transferability to Western Balkan regions using a structured experimental framework. Methods: A comprehensive band-combination ablation analysis (3–10 spectral bands and index-only inputs) was first conducted using Attention U-Net, followed by a comparative evaluation of representative convolutional and transformer-based architectures, including ResNet-UNet-50, ConvNeXt-UNet, DeepLabV3+ (ResNet-50), and DINOv2-S/14. Model performance is evaluated on an internal Serbian test split (Test SR), an external Serbian dataset (Ext SR), and a cross-regional Balkan dataset (Ext WB). Results: The results demonstrate that compact multispectral configurations (6–9 bands) provide the most stable performance, achieving mIoU values of approximately 0.72–0.74 under in-domain evaluation and remaining robust under external testing. The inclusion of near-infrared and shortwave infrared bands proved critical for effective land cover discrimination, whereas increasing spectral dimensionality beyond this range did not yield systematic improvements in external robustness. Notably, the magnitude of performance degradation under pronounced geographic domain shift exceeds the performance differences observed between architectures under in-domain conditions, indicating that distribution shift exerts a stronger influence on segmentation accuracy than model choice alone. Class-wise analysis revealed agricultural areas as the most domain-sensitive category, while Shapley-based explainability analysis provides additional insight into class-specific spectral dependencies and their role in generalization behavior. Conclusions: Although transformer-based models demonstrated competitive robustness, attention-enhanced convolutional architectures achieved comparable stability across evaluation scenarios. Overall, the findings emphasize the importance of balanced spectral design, class-aware robustness analysis, and explicit out-of-domain evaluation for developing transferable land cover segmentation models in remote sensing applications.

AI doi: 10.3390/ai7060185

Authors: Yanjun Song Jianan Hou Lidong Zhu Yi Zheng

In Low Earth Orbit Beam-Hopping Satellite Systems (L-BHSS), co-channel interference among beams severely degrades communication quality. To address the inter-beam co-channel interference avoidance problem, this paper proposes a Parameter-Sharing Multi-Agent Deep Deterministic Policy Gradient-Based Beam Geometry Multi-Parameter Optimization Algorithm (PS-MADDPG-BGMPOA) for the joint optimization of satellite beam geometric parameters. The effects of free-space path loss, atmospheric impairments, and Rician fading are comprehensively considered, and a beam geometric multi-parameter optimization model is formulated with the objective of maximizing the long-term Signal-to-Interference-plus-Noise Ratio (SINR), incorporating beamwidth, beam center offset from the satellite nadir direction angle, inter-beam separation angle, and beam activation states. To tackle the resulting high-dimensional mixed action space, the proposed algorithm employs parameter sharing and grouped decision-making, which alleviates the dimensionality explosion problem and decouples the network scale from the number of beams, enabling efficient cooperative optimization with reduced training complexity. Simulation results show that, under various channel conditions and beam configurations, the proposed method effectively enhances communication quality and spectral efficiency while exhibiting superior real-time performance and stability.

AI doi: 10.3390/ai7060184

Authors: Eddiemar B. Lagua Hong-Seok Mun Md Sharifuzzaman Md Kamrul Hasan Ahsan Mehtab Jin-Gu Kang Hae-Rang Park Young-Hwa Kim Chul-Ju Yang

This study proposed a multi-class anomaly detection framework for group-housed pigs by integrating computer vision and machine learning. Nine classification algorithms were trained to identify five pig conditions—normal, heat stress, poor ventilation, infection, and recovery—using 10 combinations of feeding, drinking, and posture variables. The analysis revealed distinct behavioral patterns across stress conditions. Linear Discriminant Analysis (LDA) using all feeding and drinking variables achieved strong performance, with precision, recall, F1-score, and accuracy of 96.2% (95% confidence interval: 89.5–100%), 96.0% (91.5–100%), 96.0% (89.8–100%), and 96.0% (91.6–100%), respectively, and an AUC of 98.7% (88.2–95.5%). However, Random Forest and XGBoost trained on feeding and drinking variables achieved perfect classification on unseen data. With the present dataset, results indicate that feeding and drinking behaviors alone are sufficient for robust anomaly detection when paired with appropriate classifiers. Overall, this pilot study demonstrated that stressor-specific anomaly detection based on behavioral data is feasible and offers a practical, scalable approach for early stress detection, improved health and welfare monitoring, and more efficient precision livestock management. Future studies should utilize larger and more diverse datasets to further validate and strengthen the generalizability of the proposed framework.

AI doi: 10.3390/ai7060183

Authors: Anusha Ihalapathirana Piia Lavikainen Pekka Siirtola Satu Tamminen Gunjan Chandra Tiina Laatikainen Janne Martikainen Juha Röning

Traditional treatment strategies for Type 2 diabetes (T2D) adopt a “one-size-fits-all” approach, limiting individual effectiveness. This study presents an explainable, data-driven framework for multi-treatment and single-treatment selection of SGLT2 inhibitors (SGLT2-i) and DPP-4 inhibitors (DPP4-i) based on patient-specific health characteristics. Our approach evaluates treatment effectiveness across four outcomes—glycosylated hemoglobin (HbA1c), low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol, and body mass index (BMI)—to enable individualized treatment recommendations. The multi-treatment model, based on multi-output regression, achieved an R2 score of 0.44 and an RMSE of 5.58, identifying benefit subgroups for SGLT2-i and DPP4-i across all outcomes. Integrated with SHapley Additive exPlanations (SHAP) analysis, the model offers insights into the factors influencing treatment effects. The single-treatment selection algorithm achieved an accuracy of 0.47 and an F1 score of 0.46, showing a higher average treatment effect with SGLT2-i on all outcomes, notably in the reduction in HbA1c, LDL, and BMI and a modest increase in HDL. While DPP4-i demonstrated beneficial effects on HbA1c, LDL, and HDL, it was associated with an increase in BMI. These findings highlight the benefits of a multi-faceted, patient-centered precision medicine approach for T2D management, enabling treatment strategies that address individual health needs beyond HbA1c.

AI doi: 10.3390/ai7060182

Authors: Ehtisham Lodhi Lin Qiu

Background: Solar energy provides a sustainable and environmentally friendly alternative to fossil fuels, and photovoltaic (PV) systems are increasingly deployed worldwide. However, their operational reliability is often compromised by various fault conditions, which reduce power output and shorten system lifespan. Although automated image-based deep learning methods have shown promise for PV fault classification, their performance is often limited by severe class imbalance and subtle, low-contrast defect patterns. This study aims to address these challenges by proposing an improved deep learning framework for robust PV fault classification. Method: An attention-enhanced convolutional neural network framework, termed AEConvNeXt, is proposed for PV fault classification. The model is built on a ConvNeXt-Tiny backbone and incorporates a dropout-regularized Convolutional Block Attention Module (CBAM) to enhance localized feature refinement. To further improve learning under imbalanced data conditions, a hybrid loss function combining Cross-Entropy Loss and Focal Loss is employed. Results: Experimental evaluations demonstrate that AEConvNeXt achieves an overall accuracy of 94.37% and a macro F1-score of 94.43%, outperforming the strongest baseline model, ResNet-50, by more than 3%. Grad-CAM visualizations further confirm that the model effectively focuses on fault-relevant regions, improving interpretability. The proposed framework also shows consistent and robust performance across all six PV fault categories under varying conditions. Conclusions: The proposed AEConvNeXt framework provides an accurate and explainable solution for real-time PV fault detection, effectively addressing class imbalance and improving minority fault recognition.

AI doi: 10.3390/ai7060181

Authors: Kaspars Sudars Edgars Edelmers Arturs Nikulins Viktorija Cīrule Matīss Šņukuts Madara Ratniece Roberts Šamanskis Klinta Luīze Sprūdža Maija Radziņa

This study compares three methods for 2D bone metastasis segmentation on computed tomography slices-the self-configuring nnU-Net pipeline, a fine-tuned DINOv3 foundation model, and a prompt-free MedSAM foundation model adaptation-to assess their suitability for clinical-grade lesion delineation. Methods: A dataset of 2D CT slices from 88 patients (11,006 image–label pairs) was annotated by experts. The three models were trained and evaluated under comparable conditions, using model-specific native input resolutions and training schedules. Performance was evaluated using the Dice similarity coefficient (DSC) and Normalized Hausdorff distance (NHD) on a held-out test set, with a separate cohort of previously unseen patients. On a held-out test set, the MedSAM, DINOv3, and nnU-Net models achieved the following Dice scores: 0.6280, 0.4480, and 0.6849, respectively. Additionally, on a held-out test set, the MedSAM, DINOv3, and nnU-Net models achieved the following normalized Hausdorff distances: 0.1008, 0.1019, and 0.0473, respectively. In conclusion, the nnU-Net framework provides robust segmentation performance and serves as a strong baseline for 2D slice-wise bone metastasis delineation even with limited annotated data.

AI doi: 10.3390/ai7050180

Authors: Muhammad Minoar Hossain Md. Hasibul Hassan Himal Arslan Munir

This research introduces a quantum machine learning (QML) approach for predicting heart disease (HD). The method combines preprocessing of data with quantum feature map (QFM) and quantum classification techniques. In the method, clinical data of HD are preprocessed, and then features are optimized using principal component analysis (PCA). After that, the resulting features are encoded into quantum states with five different QFM methods, namely angle encoding (AE), amplitude encoding (AmE), basis encoding (BE), Pauli encoding (PE), and ZZ feature map (ZZFM). Finally, four quantum classifiers, such as quantum support vector machine (QSVM), quantum k-nearest neighbor (QKNN), quantum random forest (QRF), and variational quantum circuit (VQC), are evaluated to predict the HD from the encoded states. Experimental results show that QSVM with AE achieved the best performance, with an overall accuracy of 90.26%, specificity of 83.42%, sensitivity of 92.16%, precision of 88.89%, F1-score of 89.68%, and kappa value of 0.7608. These results are superior to those from classical state-of-the-art methods. This research finding suggests QML methods can capture complex nonlinear relationships in clinical data effectively and thus improve diagnostic reliability.

AI doi: 10.3390/ai7050179

Authors: Mohammed Houache Djallel Eddine Boubiche Homero Toral-Cruz Rafael Martínez-Peláez Rafael Sanchez-Lara

The Industrial Internet of Things (IIoT) is central to modern industrial automation, yet its growing connectivity exposes critical systems to evolving cyber threats. Traditional intrusion detection methods struggle in IIoT environments due to class imbalance and limited adaptability to zero-day attacks. This systematic review evaluates generative AI techniques for IIoT intrusion detection and identifies deployment requirements for industrial environments. We searched five databases (IEEE Xplore, ACM Digital Library, Springer, ScienceDirect, and arXiv) for studies published between January 2019 and December 2025, applying predefined inclusion criteria. Following a systematic selection process (identification plus three progressive screening stages) across 342 records, 42 primary studies were included for systematic synthesis. We examined four GenAI paradigms—Generative Adversarial Networks, Transformers, Diffusion Models, and Variational Autoencoders—analyzing nine state-of-the-art frameworks through comparative performance analysis. Hybrid Transformer architectures (e.g., Transformer-GAN-AE) achieve the most consistent detection performance, while diffusion-based models (e.g., Diff-IDS) provide computational advantages for edge deployments. However, substantial variability in evaluation methodologies and limited reporting of statistical rigor indicate important gaps in current research practices. These findings inform the development of GenAI-driven strategies tailored to industrial infrastructure constraints and highlight key directions for advancing IIoT cybersecurity.

AI doi: 10.3390/ai7050178

Authors: Luigi Ceccaroni Lyle Visa Iain Visa

Measuring the impact of citizen-science projects is hard because inputs are heterogeneous, mostly categorical, and sparse. We present Alquimics, a compact supervised neural network trained on one-hot project descriptors to predict impacts across five domains (Environment, Economy, Governance, Science, and Society). Each project is encoded as a binary vector of length 4460 (223 questions × 20 options, flattened). The network employs a 4460–42–5 topology with logistic activations throughout; labels consist of five continuous targets in [0, 1] obtained by scaling expert domain scores in [1, 42]. We implement L2-regularised training in Octave using fmincg with MaxIter = 10 and lambda = 0.07. Leave-one-out cross-validation (LOOCV) over nine projects yields an overall RMSE = 10 and R2 = 0.06 on the 1–42 scale, with Governance being the most predictable domain (RMSE = 6, R2 = 0.3). We document the entire data pipeline, objective, and implementation, provide a minimal reproducible script, and discuss limitations arising from the small dataset (n = 9 projects). This establishes a transparent baseline that complements rule-based scoring and can be expanded as more labelled projects become available.

AI doi: 10.3390/ai7050177

Authors: Anca-Elena Iordan Florin Covaciu Calin Vaida Iuliu Nadas Alexandru Banica Bogdan Gherman Ionut Ulinici Jose Machado Paul Tucan Doina Pisla

Robust exercise recognition is essential for robot-assisted lower-limb rehabilitation, where misclassifications of sensor-derived movements can degrade therapy execution and supervision. This study proposes a novel hybrid weighted stacking ensemble to increase the efficiency of the intelligent module of the LegUp parallel robotic system for lower limb rehabilitation. The approach combines a Residual Multilayer Perceptron (ResMLP) and an optimized Kernel Extreme Learning Machine (KELM), where model hyperparameters are tuned using Optuna and the base-model probability outputs are fused through optimized weighting and a meta-learner. Experiments were conducted on a five-class dataset built from nine IMU orientation features acquired from three sensors placed on the healthy limb. Four meta-learners were evaluated (Logistic Regression, Random Forest, Gradient Boosting, and AdaBoost), with AdaBoost providing the best overall performance. To further assess the robustness and generalization capability of the proposed approach, a 5-fold cross-validation procedure was performed for the ResMLP, KELM, and the hybrid ensemble models. The proposed stacking hybrid ensemble consistently surpassed the performance of the strongest individual classifiers as well as the original LegUp Multilayer Perceptron model. These results indicate that combining residual learning with kernel-based classification in a weighted stacking framework yields a stable and high-performing solution for multi-class rehabilitation exercise recognition.

AI doi: 10.3390/ai7050176

Authors: Nikolay Hinov

Modern power systems are undergoing a structural transformation driven by the rapid integration of renewable energy sources, distributed energy resources, electrification, and increasing operational uncertainty. These developments expose the limitations of traditional centralized energy management and rule-based automation in highly distributed, data-intensive, and dynamically coupled energy infrastructures. In response, recent advances in artificial intelligence offer new opportunities for improving prediction, coordination, and adaptive control. This paper develops a reference architecture for Autonomous Energy Systems based on the integration of generative AI, agentic AI, digital twins, and distributed cyber–physical energy infrastructures. Rather than treating forecasting, control, simulation, and market coordination as separate research tracks, the paper organizes them within a common architectural perspective. Generative AI is positioned as a source of scenario intelligence, synthetic data generation, and uncertainty-aware forecasting, while agentic AI is framed as a bounded decision layer for perception, reasoning, planning, and coordinated action under operational constraints. The paper further clarifies the distinction between agentic AI, conventional multi-agent systems, and multi-agent reinforcement learning in energy applications. Representative application domains are discussed, including self-healing power grids, autonomous energy markets, and digital twin training environments. Major open challenges are identified in relation to scalability, physical consistency, safety verification, sim-to-real transfer, cybersecurity, interoperability with legacy infrastructures, and governance. The paper concludes by outlining a research agenda for the staged and safe development of increasingly autonomous energy systems.

AI doi: 10.3390/ai7050175

Authors: Yanis Faquir Arnaldo Santos Henrique S. Mamede

Artificial intelligence (AI) is transforming how organizations develop human potential, offering scalable and data-driven support for coaching and capability building. This study proposes and validates a conceptual framework for integrating AI into organizational coaching processes to enhance competence development and strategic alignment. AI-supported coaching in this research is treated as an emerging organizational technology whose potential organizational value depends less on model capability and more on governance design, decision rights, and auditable evaluation outputs. Following a mixed-methods, multi-phase design, the research combined a Systematic Literature Review (SLR) with the construction of a layered design architecture in which OSCAR serves as the primary coaching-process scaffold, complemented by KSA for competency specification, Situational Leadership for adaptive guidance, and KPIs for monitoring and governance. The framework structures AI-supported coaching across 10 interrelated phases, from contextual anchoring to review and measurement, while preserving iterative re-entry to earlier phases whenever review evidence, contextual change, or insufficient progress makes adjustment necessary. Prototyping demonstrated feasibility and coherence across models, while the focus group provided qualitative expert feedback on the framework’s clarity, governance needs, and perceived usefulness for competence development. At this stage, however, the KPI structures generated by the framework and the descriptive comparison across AI tools should be interpreted as prototype-level outputs rather than as empirically validated performance measures or evidence of added value over baseline approaches. Because the evaluation relied on two fictional prototyping scenarios and a small expert-oriented focus group (n = 6), the findings should be interpreted as evidence of prototype demonstration and qualitative refinement rather than of real-world effectiveness or organizational impact. The study also does not include a control group or comparison with traditional human coaching, so the added value of the AI-supported framework over alternative coaching arrangements remains a question for future empirical testing. Findings suggest that AI can usefully support organizational coaching by personalizing dialogue, structuring reflection, and generating auditable development artefacts, provided ethical safeguards and human oversight remain integral. The research contributes a preliminarily validated, ethics-informed, and governance-aware framework for AI adoption in organizational coaching and offers practical insights for embedding AI-enabled development in learning organizations.

AI doi: 10.3390/ai7050174

Authors: Tad T. Brunyé

Evaluating large language models (LLMs) critically depends on human judgment. This article reviews and develops a conceptual framework for human-centered LLM evaluation, synthesizing research across evaluation methodology, psychometrics, cognitive science, and domain-specific applications. Four primary challenges are identified that limit current human evaluation practice: imperfect gold standards, evaluator fatigue and overload, shared and unique bias structures across humans and LLM judges, and the routine omission of uncertainty and dispersion estimates. To address these gaps, the STEP-V design framework is proposed: Stakes, Task-type, Evaluator availability, Purpose, and Volume, for selecting human and/or automated LLM evaluation methods under real-world constraints. An evaluator failure mode taxonomy is also proposed that analyzes human and LLM judges within a common error framework, clarifying where hybrid pipelines can compensate for weaknesses and where they might compound them. The framework motivates a more rigorous science of LLM evaluation, one that treats human judgment as a necessary but fallible measurement requiring explicit design, calibration, and uncertainty quantification.

AI doi: 10.3390/ai7050173

Authors: José Ramón Trillo Javanshir Mammadov Yusif Huseynov Matanat Ahmadova Aysel Eminova

In the context of rapid technological evolution and increasing market uncertainty, technoparks have emerged as critical ecosystems for bridging scientific research and high-tech industrial production; however, their effectiveness is often constrained by limited flexibility, fragmented control mechanisms, and delayed decision-making processes. Motivated by these challenges, this article investigates the automation of control processes in research-driven and flexible manufacturing environments within technopark infrastructures, positioning automation as a strategic lever for enhancing operational adaptability and innovation throughput. The study conceptualizes control process automation as a multi-stage framework encompassing data acquisition, processing, intelligent analysis, and real-time decision execution and examines the role of enabling technologies such as artificial intelligence, the Internet of Things (IoT), and cyber-physical systems in supporting this paradigm. The analysis demonstrates that the integration of these technologies significantly improves production flexibility, resource optimization, and responsiveness to dynamic conditions, while simultaneously accelerating the transformation of scientific and research outputs into measurable economic value. By combining theoretical foundations with illustrative practical applications, the article substantiates the effectiveness of automated control systems and highlights their strategic relevance for increasing the competitiveness of technoparks, fostering sustainable technological innovation, and shaping resilient long-term development strategies.

AI doi: 10.3390/ai7050172

Authors: Wenqin Zhao Carman Ka Man Lee Da Li Benny Chi Fai Cheung

Machine vision plays a pivotal role in precision engineering for high-precision measurement that relies on high-resolution images. The highly reflective nature of metal surfaces and the need for high-quality images pose significant challenges in image processing. Although existing research has made significant progress in enhancing the resolution of natural images, super-resolution methods specifically tailored for multi-view metal images remain unexplored areas. To fill this gap, this paper focuses on developing a deep learning-based super-resolution algorithm, focusing on detail recovery on under multi-view metal images. The proposed super-resolution model utilizes a hybrid-resolution input that combines light field super-resolution at the image level and reference-based super-resolution at the feature level, demonstrating the effectiveness for achieving a large-scale multi-view metal image super-resolution. An experiment using a public metal object image dataset is conducted, and a comparison has been carried out with Bicubic, LFhybridSR and ERVSR. The proposed method demonstrates superior SSIM and achieves average PSNR improvements of 4.45 dB and 1.18 dB on synthetic data and real-world data. The results demonstrate that the method can improve the resolution and detail representation of metal images in terms of PSNR/SSIM and address the problem of super-resolution in multi-view metal images. Furthermore, applying the proposed SR method as preprocessing reduces the absolute relative error in depth estimation from approximately 0.5 to 0.1.

AI doi: 10.3390/ai7050171

Authors: Andreea Dutulescu Stefan Ruseti Mihai Dascalu Danielle S. McNamara

Generating multiple text sequences and refining them through feedback is essential for improving the quality of outputs in many NLP tasks. While Large Language Models can leverage iterative feedback during inference, smaller models often lack this capability due to limited capacity and the absence of suitable training paradigms. In this paper, we propose a novel Feedback-Aware Inference approach that enables iterative sequence generation with integration of feedback signals. Our method allows models to generate multiple sequences, incorporate feedback from previous iterations, and refine outputs accordingly. This approach dynamically adjusts to different quality metrics, making it adaptable to various contexts and objectives. We evaluate our approach on two distinct tasks: Answer Selection for Question Generation and Keyword Generation, arguing for its generalizability and effectiveness. Results show that our method outperforms strong baselines, maintaining high performance across iterations and achieving superior results even with smaller, open-source models.

AI doi: 10.3390/ai7050170

Authors: Zhikai Yang Zhigang Meng Zhiqiang Wen

Large language models (LLMs) have demonstrated remarkable capabilities in learning complex tasks purely from sequential data. To explore whether such models can internalize strategic world representations, We investigate whether generative transformer models can learn structured world representations from sequential data. Using the domain of Go life-and-death problems as a controlled micro-world, we train a GPT-style generative model to predict moves from serialized board states. Focusing on localized life-and-death (tsumego) scenarios, we train the model to predict valid next moves from serialized board states without providing any explicit Go rules or strategic supervision. Probing the model’s internal activations reveals structured representations aligned with liberties, eyes, and tactical group status. To interpret these representations, we introduce the Multi-Aspect World Probe (MAWP), a modular probing framework that disentangles tactical concepts into orthogonal dimensions. We further apply interventional techniques to manipulate internal representations and causally evaluate their impact on model predictions. Our results show that the proposed model achieves 94.7% accuracy in sequence correctness and 92.1% in outcome validity on life-and-death tasks. This work extends interpretability research into spatially structured domains and offers tools for understanding decision-making in sequence models.

AI doi: 10.3390/ai7050169

Authors: Idriss El-Thalji Ali Usman Waqar Ali

Remaining Useful Life (RUL) forecasting models are essential to enable predictive maintenance strategies. However, selecting the most appropriate model based solely on conventional accuracy metrics may be insufficient for practical decision making, where an adequate prediction horizon is required to plan maintenance activities. This study investigates the impact of prediction horizon on model performance and its implications for maintenance decision making. A multi-horizon evaluation approach is applied to assess model accuracy across different predictive horizons. The results show the fluctuation of accuracy and prediction error over different prediction horizons. Across both datasets, predictive accuracy was generally lowest at the long horizon (11.64–86.62%), remained variable at the medium horizon (18.13–82.04%), and was highest at the short horizon (30.29–98.25%). The results demonstrate that model performance varies significantly with the prediction horizon, highlighting a trade-off between prediction accuracy and the time available for operational planning. These findings emphasize that models with high short-term accuracy may not necessarily support effective maintenance decisions if sufficient lead time is not provided. The findings show how prediction horizon considerations shall be integrated into a risk-based evaluation framework, in which model performance is interpreted in relation to the operational consequences of prediction errors. A complete risk-based predictive maintenance framework is proposed to support a shift toward comprehensive, risk-based evaluation as a prerequisite for reliable and effective RUL prediction in predictive maintenance systems.

AI doi: 10.3390/ai7050168

Authors: Daniel S. Berman Brian Merritt Stanley Ta Dana Udwin Amanda Ernlund Jeremy Ratcliff Vijay Narayan

At present, there are no easily understood explainable artificial intelligence (AI) methods for discrete token inputs, like text. Most explainable AI techniques do not extend well to token sequences, where both local and global features matter, because state-of-the-art models, like transformers, tend to focus on global connections. Therefore, existing explainable AI algorithms fail by (i) identifying disparate tokens of importance, or (ii) assigning a large number of tokens a low value of importance. This method for explainable AI for tokens-based classifiers generalizes a mask-based explainable AI algorithm designed originally for images. It starts with an Explainer neural network that is trained to create masks to hide information not relevant for classification. Then, the Hadamard product of the mask and the continuous values of the classifier’s embedding layer is taken and passed through the classifier, changing the magnitude of the embedding vector but keeping the orientation unchanged. The Explainer is trained for a taxonomic classifier for nucleotide sequences and it is shown that the masked segments are less relevant to classification than the unmasked ones. This method focused on the importance the token as a whole (i.e., a segment of the input sequence), producing a human-readable explanation.

AI doi: 10.3390/ai7050167

Authors: Tal Laor

Artificial intelligence (AI), particularly large language models, is increasingly shaping how people access and engage with news. Guided by a multi-level digital divide framework, this exploratory study examines patterns of AI use for news consumption (AI-access) and perceptions of AI influence on social and political attitudes (AI-influence). The analysis is based on a quantitative online survey conducted among a diverse national sample of 515 participants in Israel. Measures captured self-reported AI-enabled news practices, including consuming, summarizing, and identifying fake news, as well as perceived influence and concerns about bias. Demographic indicators included age, gender, education, and income. The findings indicate a nuanced pattern that diverges somewhat from conventional digital divide expectations. Bivariate analyses suggest that older individuals, women, and those with lower levels of education report somewhat higher levels of AI use for news-related practices. However, multivariable regression analyses show that only age, gender, and education remain significant predictors, while income does not show consistent independent effects. Overall, the observed associations are relatively limited, suggesting that demographic variables explain only a small portion of the variance. At the same time, perceived AI influence shows a limited association with demographic characteristics. These results provide empirical insight into digital divide processes in the AI context and suggest that future research should examine additional explanatory mechanisms, including AI literacy, trust, perceived usefulness, and digital skills.

AI doi: 10.3390/ai7050166

Authors: Calin Muntean Vasile Gaborean Razvan Constantin Vonica Sebastian Aurelian Stefaniga Alaviana Monique Faur Catalin Vladut Ionut Feier

Background and Objectives: Post-hepatectomy liver failure (PHLF) remains the leading cause of mortality following hepatic resection, with reported incidence rates ranging from 1.2% to 32%. Traditional scoring systems such as the Child–Pugh score, Model for End-Stage Liver Disease (MELD), and Albumin–Bilirubin (ALBI) grade have demonstrated limited predictive accuracy for PHLF. Machine learning (ML) algorithms have emerged as promising tools capable of integrating complex, multidimensional clinical data to improve predictive performance. This systematic review aims to evaluate the current evidence on ML-based prediction models for PHLF, assessing their predictive accuracy, methodological quality, clinical applicability, and the key variables utilized across models. Methods: A systematic literature search was conducted across PubMed, Embase, Web of Science, and the Cochrane Library from inception to January 2026. Studies that developed or validated ML models for predicting PHLF after hepatic resection were included. The Prediction Model Risk of Bias Assessment Tool (PROBAST) was used to evaluate the risk of bias. Data on model performance, algorithms employed, sample sizes, predictor variables, and validation strategies were extracted. The review was conducted in accordance with the PRISMA 2020 guidelines and registered in PROSPERO. Results: Twelve PubMed-verified studies involving 6913 patients were retained in the final analysis. Publication years ranged from 2020 to 2025, with five studies published in 2025. Gradient boosting approaches (LightGBM/XGBoost or phase-specific boosting models) were the most frequent best-performing architectures, while ANN/deep learning, radiomics-integrated, and ensemble approaches also showed clinically relevant discrimination. Best reported non-training AUCs ranged from 0.7927 to 0.981 (median, 0.873). The strongest generalization signals came from studies with temporal, external, or prospective validation structures. Common predictor domains included bilirubin-based liver function measures, coagulation variables, platelet count, volumetry or extent of resection, imaging-derived radiomics features, and perioperative dynamic data. Conclusions: Machine learning models remain promising for PHLF prediction, but the evidence base is smaller and more heterogeneous than the original draft suggested. Performance is highest in studies that combine clinical liver-reserve markers with imaging or perioperative temporal data; however, widespread clinical adoption is still limited by retrospective design predominance, inconsistent outcome definitions, and incomplete external validation.

AI doi: 10.3390/ai7050165

Authors: Chunyang Jiang Hengyu Song Baotong Ma Shiwen Wang Chulei Zhang Peng Zhao You Zhou

In recent years, with the advancement of production technology in the manufacturing industry, the scheduling problems that rely on modeling in real-world scenarios have gradually evolved into complex process flows. Aiming at the limited problem modeling capabilities of existing scheduling problems, this study proposes Multi-Stage Flexible Job Shop Scheduling Problem (MS-FJSP). MS-FJSP alters the fixed operation processing sequence of jobs in conventional scheduling problems and introduces staged processing to incorporate flexible constraints on operation selection. Furthermore, MS-FJSP modifies the constraint of unique machine compatibility, enabling arbitrary adjustments to machine combinations according to processing requirements. To address the complex flexibility and large-scale solution space of MS-FJSP, we propose a particle swarm optimization algorithm based on double neighborhood tabu search (TS-PSO). Specifically, the PSO algorithm determines a superior neighborhood structure for this problem, while the TS algorithm improves and optimizes the solution quality within the neighborhood of this solution structure. We verify the algorithm’s performance using a dataset consisting of 12,000 MS-FJSP instances and an MS-FJSP instance modeled from a real-world scheduling scenario. Experimental results demonstrate that TS-PSO can achieve excellent solution quality within a reasonable time, and MS-FJSP possesses efficient modeling capability for real-world scheduling scenarios.

AI doi: 10.3390/ai7050164

Authors: Maria V. Leyba-Mesa Buket D. Barkana

Optical coherence tomography plays a critical role in diagnosing retinal diseases, yet automated deep learning classification is hindered by severe class imbalance in which rare pathologies are underrepresented and frequently misclassified, a limitation rarely exposed by the aggregate metrics reported in most prior work. We investigate a targeted intermediate-supervision framework, in which a secondary classifier head is attached to mid-level backbone features and jointly optimized with the primary classifier using inverse-frequency weighted loss. Unlike conventional deep supervision, which is primarily aimed at optimizing stability, the proposed formulation is used here to improve minority-class representation under severe OCT class imbalance. The method is evaluated on ResNet-18, ResNet-50, EfficientNet-B0, and ViT-B/16 using a four-class OCT dataset, with full per-class metrics reported across a systematic ablation of the auxiliary weight λ. EfficientNet-B0 achieved the best performance at λ = 0.3, attaining 97.78% accuracy, an AUROC of 0.995, and a Drusen F1-score of 93.51%, a gain of 2.64 percentage points over the unweighted baseline. Vision Transformers showed greater sensitivity to background padding artifacts than convolutional models. Grad-CAM and Attention Rollout analyses confirm that auxiliary supervision improves the localization of clinically relevant retinal structures, supporting its potential for interpretable, class-balanced automated OCT diagnosis.

AI doi: 10.3390/ai7050163

Authors: Andrea Ortega-Guzmán Salvador Ibarra-Martínez José Antonio Castan-Rocha J. David Teran-Villanueva Mayra Guadalupe Treviño-Berrones Aurelio Alejandro Santigo-Pineda

Machine learning has a wide range of applications, including classification, which categorizes elements based on their characteristics. This paper addresses the challenge of optimizing attribute weighting while assessing two crossover operators on differential evolution optimization and increasing the performance of the k-nearest neighbors classification algorithm (KNN). We use a differential evolution optimization method and assess the performance of both the differential evolution and the harmony crossover operators. Finally, the optimization method uses the accuracy of a KNN classification algorithm as a fitness function. The results show that the proposed method significantly enhances the KNN performance while proposing an alternative for other classification models such as neural networks and Random Forest.

AI doi: 10.3390/ai7050162

Authors: Sergio Garcia Francisco Gomez-Donoso Miguel Cazorla

This paper presents an innovative methodology detecting out-of-distribution individuals based on a multimodal contrastive learning approach. The system combines voice and facial image data by projecting them into a shared representation in the embedding space, enable accurate identification of previously unseen individuals. This approach overcomes the limitations of traditional methods by providing more robust and consistent detection in dynamic scenarios, using advanced neural networks and optimized contrastive losses. Specifically, the main contribution of this work is the introduction of a multimodal contrastive framework that performs cross-modal consistency verification between facial and vocal representations, enabling reliable detection of out-of-distribution individuals without the need for identity gallery retrieval. Experimental results on multiple datasets highlight the effectiveness of the system, with accuracy above 90% in detecting in-distribution samples in all evaluated cases. Regarding the identification of out-of-distribution cases, the system maintains outstanding performance, achieving values close to 90% on average, with some datasets exceeding 95%. These results underscore its ability to recognize both known identities and handle unknown data, even under challenging conditions. This approach represents a significant advancement in the multimodal recognition of individuals, with potential applications in critical areas such as security, surveillance, and human–computer interaction.

AI doi: 10.3390/ai7050161

Authors: Seungjoon Lee Myungryul Choi

Technical documents differ from general text corpora in ways that complicate retrieval-augmented generation (RAG). Evidence for a single answer is often distributed across numbered clauses, tables, figures, captions, and ordered procedures rather than expressed in one passage. Standard RAG pipelines typically flatten these elements into independent chunks. This can break the document relations needed for exact evidence tracing. We introduce TechDocRAG, a relation-preserving framework for technical document question answering. The framework represents each document as a heterogeneous element graph and aligns three retrieval views for each element: technical identifiers, semantic summaries, and raw evidence. At query time, retrieval proceeds from identifier-aware recall to summary-level reranking and raw evidence bundling. We evaluate TechDocRAG on four benchmarks with more than 7500 evaluated question–answer pairs covering product manuals, engineering documents, and long multimodal PDFs. Across the suite, TechDocRAG improves the mean end-to-end score by 20.3 points over the strongest flat baseline and by 9.3 points over the strongest non-flat baseline. On the evidence-annotated subset, the strict raw evidence hit rate increases from 0.510 to 0.942. Resource profiling shows query time latency comparable to standard hybrid retrieval. Robustness tests show gradual degradation under relation loss, but clear sensitivity to severe identifier corruption. Overall, the results indicate that reliable RAG for technical documents depends less on retrieving more passages than on preserving the relations that make evidence interpretable.

AI doi: 10.3390/ai7050160

Authors: Fatmah Alqarni Nada Alhirabi Omer Rana Charith Perera

Artificial intelligence techniques have made notable progress in supporting learning processes, with increasing adoption across educational contexts. However, despite the increasing work on AI-assisted techniques, explicit and implicit learning mechanisms in AI educational assistants have not been systematically categorised. The study of how these techniques aid in and are implemented for learning remains underexplored. Therefore, a more systematic categorisation of how these techniques support learning through user interaction is needed. This paper presents a systematic review of 38 studies published between 2000 and 2024, spanning domains including programming education, cognitive skills, language learning, and the AI field. This review was conducted and reported in accordance with the PRISMA 2020 guidelines. In this review, we propose a taxonomy of explicit and implicit learning features. We analyse implementation aspects (e.g., knowledge representation, algorithms, and interaction modalities) and synthesise how prior work evaluates learning support capabilities. The findings show that (i) 79% of reviewed studies support explicit and 21% supported implicit learning through interaction; (ii) written interaction dominates (45%), followed by visualisation (34%), while voice-based interaction remains underrepresented (9%); (iii) some implementations lack details (e.g., knowledge bases and validation methods); and (iv) evaluation practices remain uneven, with most studies relying on experiment evaluation, highlighting the need for robust evaluation practices.

AI doi: 10.3390/ai7050159

Authors: Gabriel Marín Díaz Alvaro Manuel Rodriguez-Rodriguez Eva María Andrés Núñez

Galaxy morphology combines measurable structural properties with subjective visual interpretation, limiting strictly hard-label classifications. This study proposes a framework designed to compare physically derived and human-based galaxy classifications while explicitly accounting for uncertainty and interpretability. Using photometric and structural features from the Sloan Digital Sky Survey (SDSS), physical groupings are obtained through Fuzzy C-Means clustering, enabling gradual transitions via soft memberships. Human clusters are constructed from Galaxy Zoo 2 debiased vote fractions, capturing aggregated perceptual judgments. Supervised models are trained to predict both physical and human cluster assignments from the same set of physical variables, providing a quantitative assessment of structural coherence and perceptual–physical alignment. SHAP-based explainability identifies the relative influence of color and concentration parameters in each scheme. Results show that physical clustering is driven by structural concentration and bulge dominance, while human classification exhibits smoother decision boundaries and greater sensitivity to photometric appearance. Discrepancies concentrate in transitional and orientation-sensitive systems. An interactive visualization layer supports traceable qualitative inspection. The framework provides a reproducible methodology for analyzing classification consistency, uncertainty, and human–model alignment.

AI doi: 10.3390/ai7050158

Authors: Giuseppe Esposito Massimo Sanchez Federica Fratini Egidio Iorio Lucia Bertuccini Serena Cecchetti Valentina Tirelli Daniele Giansanti

Higher technical and professional education is increasingly discussed in relation to workforce readiness, innovation, and societal inclusion. In Italy, the PNRR-funded Rome Technopole operates as a multi-institutional ecosystem in which universities, research organizations, industry, and public bodies interact through a Hub & Spoke model to support training and innovation activities. Among its components, Spoke 4 addresses professional higher technical education through the co-development of modular learning initiatives involving multiple stakeholders. This commentary examines the role and activities of the Italian National Institute of Health (ISS) within this context, with particular reference to the development of two pilot modules: one on Artificial Intelligence and Algorethics, and one on Accessibility and Assistive Technologies, including applications supported by AI. The paper combines a conceptual discussion of the approach with selected empirical insights derived from pilot implementation, including stakeholder engagement processes, structured evaluations, and thematic prioritization exercises. The findings suggest the perceived relevance of multi-stakeholder co-design, the use of flexible and modular learning formats, and the integration of technical and ethical dimensions in higher technical education. At the same time, they point to challenges related to coordination, scalability, and alignment across institutional actors. Rather than proposing a definitive model, the Spoke 4 experience is discussed as a context-specific case that may offer insights contributing to ongoing debates on the design and implementation of higher technical education in complex, multi-institutional settings.

AI doi: 10.3390/ai7050157

Authors: Juan Manuel Tabares-Martinez Adriana Guzmán-López Micael Gerardo Bravo-Sánchez Francisco Villaseñor-Ortega Juan José Martínez-Nolasco Alejandro Israel Barranco-Gutierrez

This study aims to develop and experimentally evaluate an artificial neural network-based temperature control strategy for hot-air carrot drying within an IoT-enabled cyber-physical system. The experimental setup employs an Arduino Mega 2560 equipped with AM2302 (air temperature sensor), MLX90614 (infrared surface temperature sensor), and SHT35 (relative humidity sensor), an HX711 load cell, and a WS68 anemometer, with cloud communication provided by an ESP8266 module for remote monitoring via Wi-Fi. The neural controller, implemented using the Arduino Neurona library, regulates the dryer temperature in real time, enabling drying kinetics analysis under ANN-based thermal control to investigate its capability to maintain thermal stability. Three initial loads (2, 4, and 6 kg) were analyzed to determine the thermal efficiency. In the dehydration experiments, the 2 kg load reached a final moisture content of 10% in 4.4 h, consuming 1390 kJ with a thermal efficiency of 83%. The 4 kg load exhibited the best time–energy balance (6.6 h, 1850.0 kJ, 88%), while the 6 kg load achieved the highest efficiency (8.1 h, 2250.0 kJ, 91%). These results demonstrate the effectiveness of neural-network-based control implemented on low-cost microcontrollers to enhance thermal efficiency in food dehydration processes.

AI doi: 10.3390/ai7050156

Authors: Xingyi Zhong Junhao Liu Yiqiu Mao Yubin Zhong Guanquan Zhu

Semantic segmentation of high-resolution remote-sensing images is critical for urban planning, land-cover mapping, and ecological monitoring. However, existing methods face limitations in handling complex land-cover types, multi-scale objects, and modeling long-range dependencies. To address these challenges, we propose RFA2Net, an enhanced semantic segmentation model based on the DeepLabv3+ framework. The key innovations include the integration of the RFCSA-Conv module into the ResNet101 backbone to enhance feature representation and long-range dependency modeling, the design of the RFA-DASPP structure built upon the Dense ASPP framework with the novel RFCA-DConv dilated convolution module to reduce information loss during multi-scale feature fusion and enhance the model’s ability to perceive long-range directional structures, and the introduction of a Dual-Branch Fusion Network to improve segmentation accuracy for small-scale objects. Experimental results on the ISPRS Potsdam and LoveDA datasets demonstrate that RFA2Net outperforms several CNN and Transformer-based models, achieving 78.94% and 59.46% mean intersection over union (mIoU) on the ISPRS Potsdam and LoveDA datasets, respectively, with improvements of 3.19% and 3.08% over the original DeepLabv3+. Ablation studies and comparative experiments further confirm the model’s effectiveness, robustness, and practical applicability in high-resolution remote-sensing image segmentation, with particular relevance to environmental monitoring and sustainable energy applications.

AI doi: 10.3390/ai7050155

Authors: Enoc Tapia-Mendez Irving A. Cruz-Albarran Saul Tovar-Arriaga Dulce Gonzalez-Islas Arturo Orea-Tejeda Luis A. Morales-Hernandez

The integration of artificial intelligence (AI) into the diagnosis and prognosis of heart diseases is transforming cardiovascular and cardiac healthcare, improving predictive accuracy, and personalizing treatment plans. This review presents a novel contribution by providing a comprehensive overview of both diagnosis and prognosis in heart diseases through AI, covering ML and DL models. Following the PRISMA guidelines, a total of 84 recent research articles sourced from significant journals are reported. A bibliometric analysis using the VOSviewer tool was performed to map the impact of AI, enabling a detailed examination of academic connections and contributions. The findings reveal that DL models were employed 63% for diagnosis tasks, while ML models were utilized in 37% of the studies. Key recommendations include the incorporation of essential model evaluation metrics, as clinical validation indicators, integrating explainable artificial intelligence (XAI) to improve the transparency and interpretability of models, and adopting standardized frameworks to enable smooth clinical integration. This review highlights the potential of AI to improve cardiac and cardiovascular diagnosis and prognosis, providing an overview of its strengths, limitations, challenges and the possible application as AI-driven tools in patient monitoring and to support specialists in the decision-making process.

AI doi: 10.3390/ai7050154

Authors: P. Flint Morgan Amy Megowan Bonnie Sheehey Nikunj C. Oza Bradley M. Whitaker

In the past 20 years, Artificial Intelligence (AI) has made several advancements. Because of AI’s ability to process large datasets better than humans, it is thought to have a promising future in many fields. Despite their advantages, AI and specifically Machine Learning (ML) algorithms can have emergent behavior, which makes their adoption into safety-critical systems a challenge. Through an examination of the capabilities of AI systems at NASA, we see what AI is currently being used for and which algorithms are promising for future work. We also identify limitations in the potential impact of AI systems, noting that the majority of the reviewed papers focused on limitations in adopting AI systems rather than limitations in the technical abilities of AI systems. This review article provides insight into AI and ML algorithms, aviation, space and other AI-based platforms for automation.

AI doi: 10.3390/ai7050153

Authors: Daniel Gachulinec Viktoria Cvacho Maros Jakubec Radovan Madlenak

Building reliable vehicle detection models for intelligent transportation systems calls for large, well-annotated datasets—yet gathering and labelling real traffic data remains both costly and labour-intensive. This paper introduces Traffic Synth, an automated pipeline that generates synthetic training datasets by altering real traffic camera images rather than constructing entirely artificial scenes. The system begins by detecting vehicles through instance segmentation and removing them from the frame. It then places new vehicles directly into the cleared regions using diffusion-based inpainting, all while retaining the original road layout, lighting, and camera perspective. Doing so preserves the realistic scene context while broadening the visual variety of vehicles in the dataset. To ensure that the resulting traffic looks physically plausible, we incorporate a lane-aware prompting mechanism that matches each vehicle’s orientation to the direction of travel as seen from the camera. The system further draws on a weighted vehicle brand database that mirrors the makes and colours commonly found on European roads to better match actual deployment conditions. Class-specific mask processing—involving anisotropic scaling and relative dilation—rounds out the pipeline by improving generation quality across different vehicle size categories. The final output is a set of images with automatically generated annotations in a standard object detection format.

AI doi: 10.3390/ai7050152

Authors: Marko Pribisalić Sanda Martinčić-Ipšić

Large Language Models (LLMs) are increasingly deployed across professional and societal domains, introducing security, privacy, and governance challenges beyond traditional software vulnerabilities. Despite extensive research on individual risk categories, a unified lifecycle-oriented perspective connecting architectural properties, adversarial threats, and governance implications remains limited. This review examines security and privacy risks associated with LLMs through a lifecycle framework covering data acquisition, model training, alignment procedures, deployment, and post-deployment interaction. The study synthesizes prior research to construct a taxonomy of threats including prompt injection, jailbreaking, adversarial manipulation, training-stage attacks, privacy leakage, and socio-technical misuse. Ethical issues such as hallucination, bias amplification, and malicious use are analyzed alongside governance and regulatory frameworks. Results indicate that vulnerabilities in LLM systems arise primarily from probabilistic generation mechanisms, large-scale data ingestion, and complex deployment ecosystems rather than isolated implementation defects. Classical software vulnerability models therefore provide only partial coverage of risks associated with generative AI systems. The review is grounded in the concept of the alignment gap to explain how discrepancies between training objectives and real-world interaction contribute to persistent vulnerabilities. The findings highlight the need for lifecycle-oriented defense-in-depth strategies combining technical safeguards, privacy-preserving training, runtime monitoring, and governance mechanisms to support responsible deployment of LLM-based systems.

AI doi: 10.3390/ai7050151

Authors: Befekadu Bekuretsion Wolfgang Menzel Solomon Teferra

Multilingual learning is key in natural language processing, but is challenged by the transfer–interference trade-off, where positive transfer benefits certain languages, while negative interference affects others. Prior methods, including linguistic-based and embedding-based language clustering, have attempted to address this; yet, they remain constrained by their static design and lack of task-specific feedback. In this study, we propose a novel computational strategy inspired by molecular design that constructs molecules with targeted properties. Languages are modeled as nodes in an undirected graph, with edges representing the transfer strength. This language molecule is optimized via Reinforcement Learning to adjust edge connections and weights to enhance positive transfer and minimize interference, where graph clustering is applied, and clusters are then evaluated on the Named Entity Recognition and POS tagging tasks using 25 languages from the WikiANN dataset and 12 typologically diverse languages from the UDPOS dataset. Compared to linguistic and embedding-based language clustering baselines, our method yields substantial improvements, especially for low-resource languages, with some showing over 35% increase in F1 score, while high-resource languages benefit moderately, confirming reduced transfer–interference trade-off. Our atom–language model offers a novel path for multilingual learning, inspired by molecular principles from physical sciences.

AI doi: 10.3390/ai7050150

Authors: Renxiang Chen Ci Zhang

Objectives: To address the difficulty of data sharing under privacy constraints and the performance degradation of conventional federated models caused by pronounced inter-client data heterogeneity in rolling bearing remaining useful life prediction, an FDG-based framework is developed for this task. Methods: The proposed framework jointly optimizes client-side feature learning and server-side aggregation. On the client side, a domain-adversarial learning mechanism together with a gradient reversal strategy is introduced to suppress domain-related information in degradation representations and enhance domain-invariant feature learning. On the server side, a distribution-aware dynamic aggregation strategy is designed to adaptively assign aggregation weights by jointly considering client predictive performance and feature distribution discrepancies, thereby mitigating the adverse effects of non-IID data on model aggregation. Conclusions: A federated training scenario is constructed using the PHM 2012 and XJTU-SY datasets, which involve two different bearing types. Experimental results show that, without requiring raw data to leave local clients, the proposed framework improves the accuracy and generalization capability of rolling bearing remaining useful life prediction.