Search Results (1,606)

Large Language Models (LLMs) have emerged as transformative tools across various domains, exhibiting remarkable capabilities in natural language processing and generation. However, their reliance on static pre-training data limits their ability to access up-to-date and domain-specific information. The existing research often treats augmentation strategies in isolation, and limited efforts have been made to systematically compare them through the lens of information integrity. This review focuses specifically on Retrieval-Augmented Generation (RAG) and fine-tuning, identifying them as the two dominant paradigms for integrating external knowledge: RAG for retrieval-based context injection and fine-tuning for parametric knowledge adaptation. While existing surveys predominantly focus on performance metrics like accuracy or latency, this paper addresses the critical gap of data fidelity—the preservation of truthfulness, integrity, and fairness during augmentation. We systematically synthesize empirical findings from diverse methodologies to determine how each approach mitigates hallucinations and bias. By comparing the trade-offs between retrieval-based context injection and parametric knowledge adaptation, this survey provides unique value to readers by providing a structured taxonomy, a unified evaluation framework, and actionable insights to guide future research and practical deployment of robust, high-fidelity LLMs. Full article

The development of digital learning technologies has introduced innovative tools to enhance science and chemistry education, including PhET simulations. This study aims to evaluate the effectiveness of PhET simulations on students’ learning outcomes through a systematic literature review following the PRISMA 2020 guidelines. A systematic search of Scopus and Crossref databases was conducted (last search: January 2026) using predefined keywords. Eligible studies were empirical research published between 2020 and 2026 that investigated PhET simulations in science-related education and reported learning outcomes, while non-empirical studies and non-Scopus-indexed articles were excluded. Risk of bias was assessed using an adapted Joanna Briggs Institute critical appraisal tool. Due to heterogeneity in study designs and outcome measures, the results were synthesized using a narrative approach. A total of 14 studies across elementary to higher education levels were included. The findings indicate that PhET simulations consistently improve learning outcomes, particularly academic achievement and conceptual understanding, with effects generally favoring simulation-based instruction over traditional methods. However, higher-order skills and affective outcomes such as motivation and attitude remain less frequently investigated. The evidence is limited by variability in study designs, incomplete reporting of non-cognitive outcomes, and the absence of quantitative synthesis. Overall, PhET simulations demonstrate strong potential as an effective interactive learning medium, although their impact depends on instructional design, teacher facilitation, and technological accessibility. Full article

Synthetic aperture radar (SAR) object detection has long been affected by spatial speckle interference, spectral energy imbalance, and structural bias in cross-scale feature fusion. In this article, we propose the Multi-Domain Interference-Suppressed Detection Transformer (MDIS-DETR), a unified multi-domain interference-suppressed detection framework built on the Real-Time Detection Transformer (RT-DETR) architecture. Specifically, spatial-domain interference is suppressed by learnable fusion of complementary denoising responses at the input stage. Furthermore, frequency-domain interference is suppressed by polarization-guided attention together with adaptive frequency refinement within the encoder. In addition, structural-domain interference is suppressed by non-sequential cross-scale interaction to enhance multi-scale consistency. Extensive experiments on multiple SAR benchmarks demonstrate that MDIS-DETR establishes state-of-the-art (SOTA) performance across datasets. Notably, on SARDet-100K, currently the largest SAR detection dataset with a scale comparable to the Common Objects in Context (COCO) dataset, it achieves 58.82% mAP, surpassing the RT-DETR baseline by 4.58%. Full article

Background: AI-supported game-based learning (AI-GBL) integrates artificial intelligence mechanisms, including adaptive difficulty adjustment, large language model (LLM) scaffolding, intelligent non-player characters (NPCs), and stealth assessment, into game-based educational environments. Objective: This systematic review synthesizes the empirical evidence on AI-GBL effectiveness, adaptive mechanisms, and intelligent assessment approaches across diverse educational contexts. Method: Following PRISMA 2020 guidelines, 55 peer-reviewed empirical studies (2021–2026) were identified from Web of Science and Scopus databases. Two independent reviewers screened records (κ = 0.89; 100% consensus on disagreements), extracted data using a standardized coding scheme, and assessed methodological quality using a five-criterion rubric. A thematic synthesis approach was adopted due to the heterogeneity of the evidence base. Results: The reviewed studies generally suggest promising positive effects of AI-GBL on knowledge acquisition, intrinsic motivation, and affective engagement under a range of educational conditions. LLM-based scaffolding reduces cognitive load but risks fostering passive dependency; adaptive difficulty adjustment benefits depend critically on the direction and magnitude of adaptation; AI NPCs function as credible instructional partners in both EFL and STEM contexts; stealth assessment achieves AUCs of 0.848–0.913. Challenges include algorithmic bias in assessment models, LLM latency, over-reliance risks, and a near absence of longitudinal evidence. Conclusions: AI-GBL’s effectiveness rests on principled alignment between AI mechanisms and learning theory rather than algorithmic sophistication per se. Equity-by-design approaches and longitudinal evidence constitute the field’s priority research needs. Full article

Hybrid RF-FSO-THz communication systems are promising candidates for future Internet of Things (IoT) and 6G networks because they combine the robustness of radio frequency links, the high-capacity potential of Free-Space Optical communications, and the ultra-wideband capabilities of terahertz transmission. Adaptive channel selection in such systems depends on multiple correlated environmental and physical-layer variables, including distance, rain intensity, humidity, visibility, turbulence strength, signal-to-noise ratio, channel capacity, and energy-efficiency metrics. This paper presents a physics-supported benchmark framework for supervised adaptive channel selection in hybrid RF-FSO-THz systems and systematically investigates the impact of linear and nonlinear dimensionality-reduction techniques on predictive performance, statistical robustness, computational complexity, and physical interpretability. A multi-scenario dataset comprising 5000 samples was generated using calibrated RF, FSO, and THz propagation models under clear, rain, fog, and worst-case environmental conditions. Principal Component Analysis (PCA) and Kernel PCA were evaluated together with Random Forest, Support Vector Machines (SVMs), XGBoost, Gradient Boosting (GB), Multi-Layer Perceptron (MLP), Logistic Regression, and Decision Trees. The results demonstrate that PCA preserves nearly all predictive capabilities while reducing the original 33-dimensional feature space by approximately 81.8%, maintaining accuracies close to 97–98% with the best-performing classifiers. Statistical significance analysis confirms that PCA introduces only modest degradations, whereas Kernel PCA consistently reduces the predictive performance while increasing memory requirements and inference latency. Additional environmental-only validation experiments indicate that adaptive channel selection remains highly learnable even when only pre-selection environmental descriptors are available, partially mitigating concerns regarding self-consistency bias. Overall, the results suggest that PCA provides an advantageous compromise among predictive accuracy, computational efficiency, statistical robustness, and physical interpretability for supervised adaptive channel selection in physics-supported hybrid wireless communication systems. Full article

Background: Imaging-based logistic models are widely used for non-invasive risk stratification; however, their structural robustness and transportability across heterogeneous biological contexts remain insufficiently examined. Purpose: This study aimed to develop a simulation-based stress-testing framework to evaluate the structural robustness and transportability of a radiology-adapted logistic risk model across distinct cutaneous lesion phenotypes under both aligned and structurally perturbed conditions. Methods: A simulation-based methodological framework was implemented using three synthetic cohorts representing nodular, subcutaneous, and vascular lesion phenotypes (n = 2000 per cohort). Model performance was evaluated under naïve transfer, recalibration, and revision conditions. To address potential structural alignment bias, additional simulation scenarios incorporating coefficient perturbations, nonlinear transformations, and interaction effects were used to generate outcome processes partially independent from the original model structure. Model performance was assessed using discrimination (ROC-AUC, PR-AUC), calibration metrics, decision curve analysis, and Monte Carlo-based stability assessments. Results: Under naïve transfer, discrimination remained stable across phenotypes (ROC-AUC ≈ 0.78–0.84). Calibration shifts were observed but were effectively corrected through recalibration. Under structurally perturbed outcome generation, discrimination showed only modest reduction, while overall performance patterns remained consistent. Structural variables demonstrated high transferability, whereas vascular features exhibited phenotype-dependent variability. Decision curve analysis indicated consistent clinical utility across relevant thresholds. Conclusions: The radiology-adapted logistic model demonstrated structural robustness across heterogeneous phenotype conditions, with performance variations driven primarily by calibration differences rather than structural failure. Importantly, robustness was preserved under conditions of structural perturbation, supporting the model’s stability beyond idealized alignment assumptions. These findings suggest that simulation-based stress-testing frameworks provide a rigorous methodological approach for evaluating model transportability prior to large-scale clinical validation. Full article

Aconitum soongaricum, a medicinal plant endemic to the Tianshan Mountains in Xinjiang, China, produces numerous natural compounds with potential medicinal value. Mitochondria function as energy hubs and play critical roles in plant development and stress adaptation; thus, their genomic composition underpins biological functions. Here, we assembled the complete mitochondrial genome of A. soongaricum using next- and third-generation sequencing data and performed comparative analyses with related species. The mitochondrial genome exhibited a typical circular structure of 487,849 bp with a GC content of 46.80%. A total of 77 genes were annotated, including 41 protein-coding genes (PCGs), three rRNAs, 31 tRNAs, and two pseudogenes. The genome showed a strong A/U bias at the third codon position and displayed C-to-U RNA editing transitions, whereas no U-to-C transitions were estimated. Maximum-likelihood phylogenetic analysis supported a close relationship among A. soongaricum, A. carmichaelii, and A. kusnezoffii, confirming the utility of mitochondrial genomes for genetic relationship inference in genus Aconitum. Divergence time estimation placed the differentiation of A. soongaricum from the other two species at approximately 4.19 million years ago (Mya). Additionally, we evaluated the expression levels of NADH dehydrogenase (nad) genes across different tissues and under drought stress using real-time PCR, revealing diverse expression patterns. Collectively, this study provides a foundation for future investigations into the genetic mechanisms underlying evolution, energy metabolism, and environmental adaptation in A. soongaricum. Full article

This work proposes an Adaptive Neural Control for the asymptotic stabilization of a chain of integrators at the origin. The proposed approach addresses the stabilization of the integrator chain by means of a control law whose applied signal is structurally bounded to $(-1,1)$ by the hyperbolic tangent architecture, i.e., $u\left(t\right)=tanh\left(z\right)$, where z represents a weighted linear combination of the system states and a bias term. Furthermore, an adaptation law for the weights is proposed, based on the classical backpropagation algorithm for neural networks. The stability analysis is conducted using singular perturbation theory, demonstrating that, under a sufficiently high learning rate, the closed-loop system exhibits a Standard Singular Perturbation Form. This formulation allows for the analysis of the system across two distinct time scales: the adaptation dynamics (fast subsystem) and the state dynamics (slow subsystem). Based on this formulation, explicit conditions on the learning rate and the initial conditions are derived to guarantee local asymptotic stability using Tikhonov’s theorem. These conditions characterize the region of attraction and ensure that the adaptive neural controller stabilizes the system. Numerical simulations were carried out to evaluate the controller’s performance under three different scenarios: ideal conditions, initialization outside the region of attraction, and a low learning rate. These scenarios illustrate the closed-loop system behavior and validate the theoretical conditions required for asymptotic stability. Furthermore, comparative numerical simulations were conducted on an Inverted Pendulum on a Cart system to benchmark the proposed Adaptive Neural Control against Linear Quadratic Regulator, Sliding Mode Control, and Nested Saturation Function controllers. Based on the Integral of Time-weighted Squared Error performance index, the Adaptive Neural Control demonstrated a significant reduction in control effort, achieving performance improvements of up to 95.02% compared to the aforementioned strategies. Full article

To address the problems of closed-loop task organization, strong corridor constraints, and path failure after local disturbances in unmanned aerial vehicle (UAV) inspection of gas-insulated switchgear (GIS) rooms, this paper proposes a topology-and-corridor-guided bias-suppressed D* (TCG-BS-D*) method for closed-loop three-dimensional (3D) path planning and local replanning. The proposed method constructs a structured guidance model based on the inspection-corridor topology, generates local 3D path segments according to a predetermined inspection sequence, and forms a nominal closed-loop inspection path through bias suppression and path regularization. Meanwhile, for local maintenance blockage and dynamic disturbance scenarios, an alternative local replanning strategy is applied to the affected path segments. Simulation results show that, under the static closed-loop inspection condition, the proposed method achieves a total path length of 700.22 m, a total inspection time of 269.32 s, an average safety clearance of 8.18 m, 37 large-angle turns, a corridor adherence rate of 80.73%, and a task completion rate of 100%, showing superior performance in inspection efficiency, safety margin, trajectory regularity, and corridor consistency. Under the local blockage condition, the replanned path introduces path-length and time increments of 71.29 m and 25.88 s, respectively, while maintaining the minimum safety clearance at 1.52 m and increasing the corridor adherence rate to 83.91%. Under dynamic disturbance conditions, the minimum dynamic safety clearance is improved from −2.71 m to 17.84 m, effectively eliminating the local dynamic collision risk. The results demonstrate that the proposed method can balance closed-loop path-generation efficiency, corridor-structure consistency, safety margin, and adaptability to local disturbances, providing an effective solution for UAV inspection path planning in GIS rooms. Full article

Vessel port stay time is a key indicator for berth allocation, crane planning, and short-term operational coordination in container terminals. However, existing prediction approaches often rely mainly on numerical operational data and assume complete information availability, limiting their reliability when localized visibility constraints or incomplete sensing occur. This study develops and evaluates an availability-aware multimodal prediction framework for vessel port stay time estimation. The framework adapts cross-attention-based fusion to integrate structured operational variables, numerical marine weather observations, and image-derived visibility information extracted from port monitoring images under incomplete monitoring image availability. In the framework, operational and numerical weather variables form the structured predictive state, whereas image-derived visibility information is conditionally incorporated as an auxiliary visual signal only when a matched and usable monitoring image is available. The proposed approach was evaluated using long-term vessel call data from a major container terminal. Compared with commonly used machine learning and deep learning baselines, the proposed model improved prediction accuracy, while residual analyses indicated reduced systematic prediction bias. These findings suggest that the proposed framework can support more reliable short-term berth planning under practical data-collection constraints. Full article

Accurate and reliable localization is a fundamental requirement for autonomous mobile robots operating in structured indoor environments. Adaptive Monte Carlo Localization (AMCL), widely used due to its probabilistic flexibility, suffers from performance degradation in challenging situations such as low-motion, sensor degradation, symmetry ambiguity, and abrupt position changes (kidnapped robot). This study proposes the Confidence-Gated Virtual Scan Motion AMCL (CG-VSM-AMCL) approach, which extends the standard AMCL structure with a selective and confidence-based posterior enhancement mechanism to overcome these limitations. The proposed method integrates beam partitioning, cluster-based dominance analysis, observability-aware gating, and recovery-driven adaptive particle injection components within a holistic architecture. The method was evaluated on a structured department map under seven representative scenarios: cold-start, low-motion, kidnapped robot recovery, odometry bias, scan dropout, world–model mismatch, and symmetry ambiguity. Experimental results demonstrate that the proposed approach systematically reduces localization error, false-lock rate, and convergence time compared to basic AMCL variants, and improves stability under challenging conditions. The significant improvements achieved, particularly in low-motion and symmetry-containing environments, reveal that selectively activated correction strategies can substantially increase localization robustness without altering the fundamental probabilistic structure of AMCL. Full article

Forecasting cryptocurrency prices remains difficult because market dynamics are highly volatile, non-stationary, and regime-dependent. This study investigates whether combining a spiking-inspired recurrent architecture with the Grey Wolf Optimizer (GWO) can improve one-step-ahead Bitcoin forecasting within a controlled model family. We compare four configurations, LSTM, SLSTM, GWO-LSTM, and GWO-SLSTM, on 4039 daily BTC–USD closing prices from 17 September 2014 to 9 October 2025 using Min–Max normalization, strict chronological splitting, windowed regime-based robustness analysis across three distinct market regimes, and repeated-run testing. The proposed SLSTM replaces the conventional hidden-state recurrence with leaky integrate-and-fire-inspired synaptic, membrane, and adaptive-threshold dynamics, functioning as a spiking-inspired recurrent model with thresholded event gating (reset = `none’, learnable threshold). On the primary hold-out split, GWO-SLSTM achieved a test RMSE of 1840.97 and a test MAPE of 1.76%, compared with 2217.24 and 2.46% for GWO-LSTM, 3501.48 and 3.86% for SLSTM, and 4030.10 and 4.40% for LSTM. Both GWO-optimized models exhibited substantial improvements over their non-optimized counterparts, while the SLSTM baseline also outperformed the plain LSTM, indicating gains from both spiking-inspired recurrence and evolutionary hyperparameter optimization. Both optimized models exhibited near-zero bias (PBIAS 0.11% for GWO-LSTM and 0.36% for GWO-SLSTM). Within the present implementation, GWO-SLSTM also trained faster than GWO-LSTM (39.71 s vs. 137.28 s), although this runtime difference should be interpreted as setup-specific because the model families were implemented in different frameworks and stopped after different numbers of epochs. Overall, within the expanded univariate BTC–USD setting, the results support GWO-SLSTM as a strong within-family candidate for one-step-ahead forecasting under non-stationary conditions. Full article

The Abruzzo (AB) donkey is a mountain-adapted Italian population listed as a genetic resource at risk of extinction (census ≈ 600 animals; no studbook). We aimed to provide the first comprehensive morphometric description of the breed and to compare it with the Martina Franca (MF) donkey, its main progenitor. Sixty-nine adult donkeys (56 females, 13 males) from six farms were measured in 2024. Twenty-three linear traits plus body weight and body condition score were recorded three times by a single operator. Descriptive statistics, Welch’s t-test or Mann–Whitney U test with Benjamini–Hochberg correction, PCA and LDA with leave-one-out cross-validation were performed in R; comparison with MF was based on published summary statistics. Coefficients of variation for the three studbook-admission parameters were ≤0.10 in both sexes. Sixteen of 26 traits showed significant sex dimorphism, with the largest effect sizes for rump height, medial canthal distance and wither height. LDA correctly classified 94% of animals by sex. AB females were significantly smaller than MF in 22 of 23 shared traits but had a wider thorax (p = 0.012). The sexual dimorphism observed in the Abruzzo donkey is male-biased and predominantly size-based, with a minor and well-localised shape component in the head region. Males are significantly larger than females for all axial measurements (wither height A: +6.5 cm, +5.3%; rump height B: +6.4 cm, +5.0%; trunk length D: +12.1 cm, +9.5%), for thoracic circumference (M: +7.1 cm, +5.0%), for body weight (+49.9 kg, +20.6%) and for the main head traits (CM: +4.3 cm, +20.0%; G: +3.6 cm, +12.5%; H: +1.3 cm, +11.1%; E: +2.0 cm, +3.8%); no trait shows a significant female bias after BH-FDR correction. The AB donkey shows a uniform mesomorphic phenotype, smaller and stockier than MF, supporting the establishment of an official studbook. Full article

Background/Objectives: Incorporating parent training into cognitive-behavioral therapy for anxious youth has not been shown to significantly improve outcomes perhaps because these interventions have not addressed potential interfering psychological barriers to implementing parenting changes and rarely offer between-session support. There is growing evidence that Acceptance and Commitment Therapy (ACT) can target these psychological barriers and generate more flexible and adaptive behavioral repertoires in parents of children with a variety of presenting challenges. Methods: Following a pilot trial of “Acceptance and Commitment Therapy for Parents of Anxious Children (ACT-PAC)” a six-week group-based intervention focused on targeting psychological barriers to parenting change using mindfulness and acceptance approaches, we collected qualitative feedback from participants in two post-treatment phases by conducting individual interviews and a focus group with participants that completed the intervention. Results: Analysis of interview responses revealed that parents found ACT principles and processes to be helpful, and many also appreciated the ACT-PAC group setting that allowed parents to recognize their experiences were shared by others and to self-disclose in a non-judgmental space. Feedback from the focus group further provides preliminary evidence that ACT-PAC is acceptable to and feasible for parent participants and suggests modifications such as involving additional caregivers, making resources more readily available, and creative structural changes that may facilitate between-session practice. Conclusions: Results suggest that the group-based intervention can be both maintained and improved for future participants. Limitations to generalizability in light of possible selection bias and the small focus group sample size are addressed. Full article

The growing availability of genomic resources is changing how genetic diversity is studied in Vitis vinifera L. At the same time, it has become increasingly clear that a single reference genome cannot fully represent the complexity of a species characterised by high heterozygosity, clonal propagation and a long history of diversification. Recent grapevine pangenomes, super-pangenomes and graph-based resources have revealed forms of variation that are often overlooked in conventional reference-based analyses, including structural variants and gene presence–absence variation. Rather than providing another inventory of available datasets, this review examines how continued reliance on a single reference genome may influence the interpretation of grapevine diversity and what can be gained from a broader pangenomic perspective. Drawing on recent studies in grapevine and other crops, we discuss how these approaches are beginning to improve the representation of genetic diversity, uncover biologically relevant variation and strengthen links between genomic information and adaptive traits. We also examine the challenges that still limit their practical use, particularly the integration of genomic resources with functional studies and breeding programmes. In the end, the value of pangenomics will probably depend not only on generating additional genomic resources, but also on how effectively these can be translated into tools that support grapevine conservation, climate adaptation and varietal improvement. Full article