Search Results (502)

Building energy consumption accounts for 30% of global energy use, making building management pivotal to achieving global sustainability. Occupants have profound impacts on the building environment. Incorporating occupant-related factors into the environmental control process is essential for optimizing the efficiency of building management systems (BMSs), which thus gives rise to the concept of occupant-centric control (OCC). Conventional methods rely on simplified models and fixed schedules that fail to satisfy environmental control and occupant requirements, while constructing credible models places strict requirements on the dataset. In this paper, we propose a Model-Aware Predictive Control (MAPC) framework that can construct credible models with limited data and provide room-level control strategies to optimize the trade-off between occupant comfort and energy consumption. The technological innovations of this research are twofold. On the one hand, we design a model construction and fine-tuning method that combines data-driven subspace projection approach with physical priors that can construct credible thermal dynamic models with limited data. On the other hand, to balance the potential conflicts between enhancing occupant comfort and saving energy, we present a hierarchical decision-making mechanism that enables adaptive multi-objective room-level control considering dynamic occupant comfort requirements and energy usage. The experimental results obtained on an EnergyPlus-based simulation dataset and a publicly available dataset demonstrate that MAPC can provide room-level control strategies based on dynamic occupant requirements and user preferences and achieve superior trade-offs between occupant comfort and energy consumption. The ablation experiments also demonstrated the superiority of MAPC in constructing reliable models on limited datasets. MAPC provides pivotal support for the advancement of the intelligent buildings and sustainable indoor environment. Full article

Private Networks or Standalone Non-Public Networks (SNPNs) are essential for Industry 4.0 and enterprise connectivity. However, most existing studies rely on simulations, evaluate only a single radio access technology, or report raw key performance indicators (KPIs) without an interpretable quality assessment framework. In practical deployment, operators require measurement-driven evidence to assess the performance and feasibility of 4G LTE and 5G SNPN solutions. This study presents a controlled experimental comparison of software-defined radio (SDR)-based 4G LTE and 5G SNPNs using the same Universal Software Radio Peripheral (USRP) platform, Open5GS, srsRAN, and commercial off-the-shelf user equipment (COTS-UE). The evaluation was conducted in an indoor environment under line-of-sight (LOS) and non-line-of-sight (NLOS) conditions. Experimental iPerf3 results show that the SDR-based 5G SNPN achieves higher downlink and uplink throughput than the SDR-based 4G LTE SNPN across all tested scenarios. The 5G deployment reaches up to 55 Mbps downlink and 40.5 Mbps uplink under LOS conditions, while maintaining 42 Mbps downlink and 28 Mbps uplink under NLOS conditions. Furthermore, 5G achieves lower latency than 4G LTE, with average values ranging from 21 ms to 31 ms. To provide interpretable network quality assessment, a Mamdani fuzzy logic-based Network Quality Index (NQI) with 81 inference rules is proposed to map signal-to-interference-plus-noise ratio (SINR), throughput, latency, and jitter into linguistic quality levels. The proposed approach enables nonlinear integration of heterogeneous KPIs and provides a technology-agnostic framework for practical SNPN deployment. Full article

Recent advancements in Vision-Language Models (VLMs) have enabled robotic systems to leverage model-based understanding and reasoning over visual and linguistic inputs, offering a promising approach for interpreting user intent in human–robot interaction (HRI). In particular, deictic expressions commonly used in object handovers, such as “take this” and “give me that”, cannot be fully interpreted through language alone and require a comprehensive understanding of the speaker’s perspective and the environment. This study proposes a prompt-driven vision-language framework for deictic interpretation in human–robot handover. The system integrates a pre-trained VLM with a hierarchical prompt that decomposes reasoning into intent classification, spatio-temporal grounding, and output self-validation, enabling accurate identification of target objects and goal locations without model fine-tuning. Experimental results demonstrate 100% command interpretation accuracy across multiple interaction scenarios, including pick-and-place tasks, robot-to-human and human-to-robot handovers, and temporal deictic commands. Notably, the system operates under a prompt–command language mismatch, accurately interpreting Korean commands while being guided by English-based prompts. Analysis across progressive system configurations further demonstrates that structured prompting plays a critical role in reasoning performance. These results highlight the effectiveness of a prompt-driven approach for deictic interpretation and spatio-temporal grounding, providing a practical training-free framework for HRI. Full article

The recruitment headhunting process is time-intensive due to manual candidate searches across multiple job platforms, creating inefficiencies in identifying suitable candidates. Current AI-driven recruitment platforms frequently prioritize accuracy over explainability, limiting transparency for non-technical users such as recruiters. This study streamlines recruitment headhunting by (1) consolidating publicly available candidate data from multiple job portals using a professional data aggregation Application Programming Interface (API), and (2) implementing explainable machine learning for transparent candidate–job matching. We utilized the Coresignal API (v1) to aggregate and standardize candidate profiles (N = 587) sourced from LinkedIn and Indeed, including skills, experience, certifications, and education. Using Term Frequency–Inverse Document Frequency (TF-IDF) feature vectors and regression models (Ridge, Gradient Boosting, Random Forest), we matched and ranked candidates against a standardized Data Scientist job description. Shapash was incorporated to provide interpretable feature importance explanations accessible to non-technical users. Model performance was evaluated using stratified 5-fold cross-validation with statistical significance testing. Ridge Regression achieved superior performance (cross-validated R² = 0.935, bootstrap R² = 0.954, 95% confidence interval [0.939, 0.965], RMSE = 0.025) compared with Gradient Boosting (R² = 0.840) and Random Forest (R² = 0.733). Paired t-tests confirmed significant differences between all model pairs (all ps ≤ 0.001, Bonferroni corrected) with large effect sizes (Cohen’s d ≥ 1.992). Shapash analysis revealed that top-contributing features, such as “engineering”, “data science”, “machine learning”, and “python”, aligned precisely with job description requirements, validating the model’s feature-learning capability. This approach reduces repetitive manual searches across job portals while providing interpretable insights into candidate–job rankings. The methodology’s originality lies in combining professional data aggregation APIs that access publicly available profile data with interpretable models enhanced by user-friendly visualization tools, creating a practical, potentially transferable solution for transparent AI-driven recruitment. Full article

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. Full article

Large language models (LLMs) have significantly enhanced the fluency, consistency, and adaptability of social bots, raising new concerns about their ability to integrate into online communities. However, community integration requires more than text generation alone. Social bots often lack a systematic understanding of community culture, struggle to maintain consistency between persona settings and posting behavior, and have difficulty identifying users with higher interaction potential. To address these challenges, this paper proposes HSEF-CI, a human-like social bot enhancement framework for community integration. The framework constructs community profiles from target communities, reshapes bot identities and long-term memory, adopts a staged text generation workflow, and selects interaction targets via homophily-based matching. Experiments on multiple English-speaking communities show that the framework lowers detectability across several detectors, improves social bots’ ability to integrate into target communities, and increases users’ willingness to interact. These findings highlight the importance of jointly modeling community profiles, identity reshaping, adaptive text generation, and target selection in studying LLM-driven social bots for community integration. The proposed framework also helps reveal how social bots adapt to and integrate into online communities, and provides an empirical baseline for the future development of detectors targeting community integration behaviors. Full article

Nowadays, the market for wearable devices and biorobotics is growing rapidly. In active prostheses and exoskeletons, joints are typically driven by electric machines. The critical challenge is balancing the generated torque with the size and mass of the motor, ensuring the overall weight does not hinder the user’s mobility. This paper presents a comprehensive review of high-torque electric machines based on an analysis of 162 publications, primarily from the last decade. The study systematically compares geometric, electrical, and efficiency parameters across various electromechanical converters to identify the optimal limits for bionic applications. Data suggests that an electric machine aiming to align with typical biorobotic requirements would likely fall within an outer diameter of 150 mm, an axial length of 85 mm, and a mass of 1200 g. Furthermore, the required parameters for advanced applications include an efficiency above 95%, a safe nominal voltage of up to 48 V, and the ability to generate torques up to 65 Nm. The analysis highlights that while conventional motors (such as BLDC and PMSM) dominate the market, achieving a torque density exceeding 35–45 Nm/kg—necessary to approach biological muscle capabilities—often requires adopting emerging topologies, such as magnetic gears or Vernier machines. This review provides clear quantitative guidelines for engineers designing optimal drive systems for biorobotics and wearable devices. Full article

Academic research laboratories often exhibit gaps between formal safety policies and everyday practices, driven by variability in training, leadership engagement, and safety practices. Effective safety therefore depends not only on formal compliance programs but also on operational factors such as training quality, SOP use, audit practices, and reporting culture. This study examines how operational factors within a large research-intensive university, including laboratory role, access to and adequacy of safety training, use of standard operating procedures (SOPs), experience with audits, near-miss reporting practices, and laboratory workers’ perceptions of risk and safety culture, are related to one another. A cross-sectional anonymous survey was administered to 1340 individuals, of whom 245 self-identified as currently working in research laboratories. Categorical data were analyzed using likelihood ratio chi-square tests with false discovery rate adjustments. Respondents reported high overall use of SOP use (85%), but staff indicated significantly lower SOP use than graduate students (69% vs. 91%, p = 0.004), and staff were more likely than faculty to view audits as helpful (97% vs. 85%, p = 0.050). Only 68% of laboratories reported documenting near misses, and 25% of respondents reported difficulty locating required training, despite 88% of training users rating it as sufficient once accessed. Although 52% of respondents classified their laboratory as moderate or high risk, 96% nonetheless described their laboratory as safe, suggesting normalization of risk based on self-reported perceptions. No significant associations were observed between perceived laboratory safety and years of experience, hours worked in the laboratory, or extent of training completed. Overall, the findings highlight the importance of not only accessible training and consistent procedures but also institutional conditions that support reporting, learning, and shared responsibility for hazard mitigation in academic research laboratories. Full article

Low-voltage distribution networks are critical for supplying power to end-users, and electric shock safety is a key concern; however, the frequent incompleteness of topology information in practical operations makes it challenging to accurately assess electric shock risks. This paper proposes a two-stage approach for electric shock simulation and risk assessment in low-voltage distribution networks with completely unknown topology and absent phase-angle measurements, addressing the critical challenge of unavailable, incomplete, or outdated topology information using only conventional smart meter data. It innovatively investigates shock risks under TT, TN-C, and TN-S grounding systems without prior topology knowledge or synchronized phasors. The proposed methodology combines a phase-angle-agnostic data-driven stage and a model-driven stage: the data-driven stage uses an iterative algorithm for topology label matrix estimation and weighted Laplacian matrix reconstruction with hierarchical clustering to identify network structure and line parameters, requiring only active power, reactive power, voltage magnitude, and current magnitude. The model-driven stage adopts modified nodal analysis with the finite-difference time-domain (MNA-FDTD) method to evaluate transient leakage voltage distribution under single-phase-to-ground faults, thereby assessing electric shock risks in line with international safety standards. Key contributions include a practical phase-free topology identification framework, comparative risk analysis of three grounding systems, and an integrated data-model approach for real-world low-observability networks. Simulation results show accurate topology/parameter identification with a relative Frobenius-norm error of only 1.8% even without phase data. TN-S provides the highest safety complying with IEC standards, followed by TN-C and TT under specific conditions, offering a practical solution for utilities lacking detailed topology records. Full article

Smartphone operating systems must eventually evict resident apps when memory becomes scarce, yet prior work has focused more on reclaim mechanisms and next app prediction than on the ranking rule that chooses the victim. We study app eviction through relaunch distance and show that generalizing raw relaunch distance prediction is unsafe as a direct policy because small errors among short returns can easily reverse victim ordering, while some resident apps still require fallback handling. Therefore, we propose a calibrated relaunch distance framework that places predicted and fallback candidates on a common scale. In trace-driven fixed capacity app cache simulation on a multi-user smartphone trace, the proposed method remains above LRU from cache capacities $C=5$ to $C=13$ on the 279-user evaluation set and improves average hit ratio from 0.8900 to 0.8935. At low cache capacity $C=5$, it improves hit ratio from 0.7617 to 0.7691, recovering 21.2% of the remaining Oracle–LRU gap, whereas the raw prediction method is below LRU at 0.6283 for the all-user set. The gains are strongest for users with deeper histories, where the margin at $C=5$ reaches +0.0138 in q4. These results show that calibration is the step that turns relaunch distance prediction into a deployable app eviction policy. Full article

Trustworthy deployment of artificial intelligence in safety-critical systems requires accurate diagnosis of anticipated scenarios and reliable rejection of out-of-distribution (OOD) inputs that fall outside the modeled operational scope. Existing data-driven diagnostic models typically assume that test inputs are drawn from the training distribution or rely on heuristically tuned thresholds that lack enforceable safety guarantees. This article presents SCOPE (Safety-Calibrated Out-of-distribution Prediction via Contrastive Embeddings), a framework integrating supervised contrastive learning with split-conformal prediction to provide statistically grounded OOD rejection with finite-sample false-alarm control. SCOPE employs a causal residual convolutional encoder to map multivariate sensor streams into a hyperspherical embedding space with a compact, class-specific structure. A k-nearest-neighbor density nonconformity score, computed in the encoder embedding space, flags transients that occupy low-density regions relative to known accident manifolds; an ablation shows that this density score outperforms prototype distance, entropy, and conservative maximum fusion as well as a panel of standard OOD baselines (MSP, ODIN, energy, Mahalanobis, OpenMax, MC-dropout, and a reconstruction autoencoder). To support temporally evolving trajectories, SCOPE aggregates window-level scores under a monotone decision policy and performs trajectory-level conformal calibration, yielding distribution-free guarantees that bound the probability of falsely rejecting a known accident run. SCOPE is evaluated on the Nuclear Power Plant Accident Data (NPPAD) benchmark using high-openness splits that withhold entire accident families as unknowns, and all metrics are reported as mean ± standard deviation across multiple random seeds. Results demonstrate strong diagnostic accuracy on accepted trajectories, conservative false-alarm rates satisfying user-specified safety constraints across multiple operating points, and timely rejection of unseen accident mechanisms, making SCOPE suitable for deployment in safety-critical monitoring applications. Full article

Background: Short-form video platforms have become important channels for vaccine science communication, yet whether professionally produced vaccine content aligns with audience demand remains underexplored. Methods: We conducted a cross-sectional quantitative content analysis of 3752 publicly available vaccine-related videos retrieved from three major Chinese short-form video platforms between 21 November and 13 December 2024. A coding framework based on the Health Belief Model (HBM) and the World Health Organization (WHO) Behavioral and Social Drivers (BeSD) framework was used to identify key content themes. Multivariate Bayesian negative binomial regression and demand–avoidance analysis were used to examine engagement patterns and supply–demand alignment across account types. Results: Individual users produced the majority of videos (53.17%), whereas medical professionals received the highest level of engagement. Engagement was positively associated with themes related to disease severity (β ≈ 0.19–0.25) and side effects and management (β ≈ 0.31–0.67), but negatively associated with vaccine effectiveness (β ≈ −0.28 to −0.14) and vaccination precautions (β ≈ −0.28 to −0.27). Professional sources showed broader thematic coverage but also the greatest supply–demand mismatch, with mismatch indices of 0.377 for medical institution official media and 0.304 for medical professionals, primarily driven by overrepresentation of themes associated with audience avoidance. Conclusions: Significant structural mismatch exists between professionally produced vaccine content and audience engagement-based demand on short-form video platforms. Optimizing vaccine communication may require prioritizing audience-concerned risk-related information and dynamically adjusting content strategies based on engagement feedback to enhance the effectiveness of vaccine education. Full article

The release of datasets containing sensitive user information requires a careful balance between data utility and privacy preservation. To address this challenge, numerous privacy preservation models have been proposed, including k-Anonymity, l-Diversity, t-Closeness, and Differential privacy. However, these models are largely designed for simple datasets in which each attribute is represented by a single (atomic) value, limiting their effectiveness in more complex data environments. Specifically, k-Anonymity and its variants have been widely adopted to mitigate privacy risks arising from quasi-identifier-based inference attacks. While l-Diversity and t-Closeness are extended from k-Anonymity to address the disclosure of sensitive attributes. However, they are primarily effective when sensitive attributes are singular and well defined, which restricts their applicability in scenarios involving complex or content-based data. Another prominent approach is Differential privacy and its variants, which rely on probabilistic mechanisms and the introduction of random noise into query outputs. It provides strong theoretical guarantees and is well suited for numerical data and computation-driven applications. However, it is also less effective for content-based datasets, where semantic meaning and contextual integrity are essential and cannot be preserved through randomization. To overcome these limitations, this study proposes a new privacy preservation model, $(d,c,l)$-Privacy, specifically designed for content-based datasets. The proposed model ensures that released datasets satisfy the constraints defined by parameters d, c, and l, thereby mitigating potential privacy violations. To enforce these constraints, three algorithms are introduced, i.e., FCFS, greedy, and optimal $(d,c,l)$-privacy algorithms. The FCFS algorithm prioritizes computational efficiency while maintaining acceptable privacy guarantees. The greedy algorithm balances execution time and data utility. While the optimal algorithm focuses on maximizing semantic preservation and overall data usefulness, albeit at a higher computational cost. Experimental results show that the proposed algorithms effectively mitigate privacy risks in released datasets under $(d,c,l)$-privacy constraints. Among the evaluated algorithms, FCFS achieves the highest computational efficiency, while the greedy algorithm provides a favorable trade-off between efficiency and data utility. The optimal algorithm consistently delivers the highest level of data quality, despite increased computational overhead. These findings indicate that the proposed model and algorithms provide an effective and practical solution for privacy preservation data publishing in real-world, content-based data environments. Full article

This study develops and applies a climate-based, user-centred and data-informed framework to assess lighting performance in educational buildings through the integrated use of daylight, high-efficiency LED systems and smart lighting controls. The research was conducted as a case study in university classrooms in Madrid, Spain, using a mixed-methods approach that combined in situ illuminance measurements, climate-based simulations with DIALux Evo 12.1, lighting energy assessment and structured user-perception surveys. The main objective was to quantify the dynamic interaction between daylight availability, electric lighting demand and perceived visual comfort, while assessing the energy-saving potential of daylight-responsive control strategies. Results show that the existing LED systems meet current illuminance requirements, with calculated lighting power density values ranging from 4.38 to 12.47 W/m². However, the analysis also reveals that high daylight availability does not necessarily guarantee better lighting performance, since excessive or uneven daylight can generate spatial imbalance, glare risk, and reduced visual stability. Survey results confirmed a strong student preference for daylight and exterior views but also showed that visual task clarity and glare control remain essential for user-centred lighting design. Overall, the findings demonstrate that effective classroom lighting retrofits should move beyond LED replacement alone towards adaptive, daylight-driven and user-centred control strategies capable of reducing energy use while maintaining visual comfort in educational buildings under Mediterranean climatic conditions. Full article

In contemporary software development, user satisfaction plays a central role in guiding software evolution, particularly in domains driven by continuous user feedback. Such feedback contains valuable signals about functional priorities, quality concerns, and emerging risks; however, its unstructured nature makes requirements prioritization for releasing planning complex and error-prone. This challenge is further amplified by differences in perceived value, risk severity, and semantic interdependencies among requirements. This paper proposes a Context-Aware Semantic Genetic Algorithm (CA-SGA) that formulates requirements prioritization and release planning as a unified optimization problem that jointly considers user satisfaction and dependency-based cohesion. The approach is built upon a predefined Context-Aware Semantic Requirement Matrix (CASRM), which encodes Kano-based functional importance, risk impact levels inferred from user feedback, and semantic dependencies derived from an ontology-driven Dependency structure matrix. These components are integrated into a composite fitness function that maximizes satisfaction while accounting for contextual dependencies and feasibility constraints. The proposed method is evaluated on a large-scale real-world mobile banking dataset and compared against multiple baseline approaches. The results demonstrate that CA-SGA consistently generates more coherent and effective release plans, achieving higher fitness values and improved convergence stability. Full article