Search Results (4,691)

Large language model (LLM)-based document translation systems typically treat each segment independently, discarding terminology decisions, human corrections, and discourse cues after each generation step. This stateless approach causes terminology inconsistency across segments, failure to propagate approved post-edits downstream, and redundant prompt-token consumption. Existing solutions—document-level MT, retrieval-augmented generation, and computer-assisted translation (CAT) tools as a general category—address individual aspects but lack a unified, state-aware architecture with provenance, update rules, and rollback semantics. We propose Language Twin, a shared-state architecture that organizes translation projects into seven versioned layers (L0–L6), supporting selective context loading, scoped human-edit propagation, and reversible updates. A pilot study translated three curated English-to-Korean document bundles (17 segments) using GPT-4o with a temperature of 0.3. The Language Twin condition (P1) achieved numerically higher preferred-term accuracy than the strongest baseline (17/21 vs. 14/21; not statistically significant at this sample size) and showed no repeated downstream errors in the monitored set (0/5 vs. 5/5 against the propagation-disabled ablation; Fisher’s exact test: p = 0.008), while reducing prompt tokens by 39.2% relative to full-context loading (A4). In blinded human evaluation (quadratic-weighted κ = 0.71–0.78), P1 achieved the highest terminology rating (4.38/5 vs. 3.97/5) and lowest post-editing time (16.9 s vs. 19.1 s per segment). These pilot-scale results indicate that governed shared state can improve terminology consistency and editing efficiency. Full article

Forest structural complexity underpins habitat quality, microclimate buffering, and resilience, yet it remains poorly characterized across the Hindu Kush Himalaya (HKH) where field inventories and airborne LiDAR are difficult to scale across rugged terrain. Conservation planning and protected-area evaluation in the HKH therefore often rely on canopy height or cover proxies that do not directly represent vertical structural organization. Here we develop a repeatable, uncertainty-aware indicator of forest structural complexity from GEDI waveform LiDAR using the Waveform Structural Complexity Index (WSCI) and its prediction intervals. We first define a conservative analysis footprint (“trustable pixels”) by combining a woody-vegetation screen with minimum GEDI sampling support and canopy-stature plausibility, and by excluding the highest-uncertainty tail using a relative prediction-interval criterion. To separate complexity from canopy height, we model the HKH-wide expected WSCI–RH98 relationship and map height-normalized excess complexity (observed minus expected), identifying structural complexity hotspots and coldspots as the upper and lower tails of the excess distribution. Anomaly patterns are strongly organized along elevation and treeline-relevant belts and show coherent departures among ecoregions that persist after stratified adjustment for elevation and mean annual precipitation, indicating additional controls beyond broad environmental gradients. Protected areas exhibit systematically lower hotspot prevalence than surrounding landscapes, and within-elevation comparisons suggest this association is not explained by elevation alone, highlighting the need to interpret protected-area signals in the context of placement and land-use pressure. Overall, the anomaly atlas provides an operational indicator framework to stratify monitoring, prioritize field validation, and support the landscape-scale assessment of structural conditions beyond canopy height across one of the world’s most critical mountain forest systems. Full article

Water surface object detection in shipborne remote sensing is challenged by unstable wave-induced backgrounds, illumination variations, extreme scale changes with tiny objects, and limited annotations. Multimodal RGB–infrared (RGB–IR) sensing leverages complementary visible and infrared cues to enhance robustness. However, most existing RGB–IR methods rely on backbones pretrained on limited-scale data, which constrain their performance for complex water surface scenes. In this work, we propose the WS-R-IR Adapter, a parameter-efficient vision foundation model (VFM)-based framework for shipborne RGB–IR object detection. Instead of full fine-tuning, it adapts frozen VFM representations via lightweight task-specific designs. the WS-R-IR Adapter includes (1) a water scene domain-aware modal adapter that progressively guides frozen backbone features with evolving semantic cues, (2) a parallel multi-scale structural perception module for fine-grained, scale-sensitive modeling, (3) an adaptive RGB–IR feature modulation fusion strategy, and (4) a resolution-aligned context semantic and structural detail fusion module. Moreover, we introduce an object-guided global-to-local registration framework to address dynamic cross-modal misalignment, and construct modality-aligned PoLaRIS-DET and ASV-RI-DET datasets that cover diverse water surface scenes. On the two datasets, the proposed method achieves mAP@0.5:0.95 scores of 74.2% and 50.2%, respectively, significantly outperforming existing methods with only 11.9M additional parameters. These results demonstrate the effectiveness of parameter-efficient VFM adaptation for multimodal water surface remote sensing. Full article

Lung nodule detection remains a critical challenge in clinical diagnostics due to the small size, weak contrast, and high background interference of nodules in CT scans. To address these issues, a novel deep neural network architecture, termed Daisy-Net, is proposed. This model incorporates dual attention mechanisms and inter-scale feature perception, consisting of two primary components: the Parallelized Patch and Spatial Context Aware (PPSCA) module and the Omni-domain Multistage Fusion (OMF) module. The PPSCA module enhances the extraction of fine-grained textures and boundary information through multi-branch patch perception and spatial attention. The OMF module employs omni-domain feature fusion and progressive stage-wise supervision to improve robustness and discrimination under complex conditions. The lung nodule detection task is formulated as a two-dimensional segmentation problem and evaluated on the LUNA16 dataset. In the post-binarization comparative evaluation, Daisy-Net achieves the best overall performance among all compared methods, with an Intersection over Union (IoU) of 81.41, a Dice coefficient of 89.75, a precision of 95.34, a sensitivity of 84.78, and a specificity of 99.9974. These findings indicate the model’s strong capability in detecting small pulmonary nodules accurately and reliably. Full article

This study examines the effects of a three-month pedagogical intervention that integrated artificial intelligence (AI), social media, and web-based tools to strengthen digital literacy, creativity, and cultural participation among secondary education students in Ecuador. The intervention was theoretically grounded in perspectives of inclusive digital education and Universal Design for Learning (UDL), emphasizing participation, accessibility, and collaborative knowledge construction. The intervention involved 61 students supported by 31 university facilitators and was developed under a mixed-methods action research design with a pre–post (quasi-experimental) approach. Pre- and post-test surveys were administered to assess changes in digital competencies and creativity, while semi-structured interviews explored students’ perceptions of creative expression and their engagement with the cultural and technological ecosystem. Quantitative results showed statistically significant improvements in digital literacy and creativity (p < 0.001), while qualitative findings evidenced increased student empowerment, critical awareness of algorithms, and active cultural participation. The integration of AI and social media promoted an inclusive, student-centered learning environment that enhanced autonomy, reflective thinking, and media engagement. These results suggest that hybrid and culturally contextualized AI-mediated interventions may foster 21st-century competencies, strengthen digital equity, and promote creative agency in educational contexts of the Global South, particularly within emerging digital learning environments in Ecuador. Full article

In this study, we examined cross-cultural interest in sustainable brands in Saudi Arabia and the United Kingdom in 2015–2025 as a dynamic proxy of consumer behavior and public awareness through the Google Trends tool. Due to the increasing significance of sustainability marketing as a part of environmental, social, and governance (ESG) strategies, the research focuses on the evolution of sustainability discourse as influenced by cultural environment, language, and policy frameworks. Sustainability-related search terms in English and Arabic were gathered in Google Trends monthly and aggregated to eliminate short-term variability and were compared through nonparametric Wilcoxon rank-sum tests and time-stability tests. The findings demonstrate that there is more stable and higher public interest in issues that relate to sustainability in the UK, which is an indicator of a well-established ESG and regulatory climate. Contrarily, Saudi Arabia shows a significant upsurge in search transactions on sustainability-related topics that occurred after 2018, when the country started implementing reforms to Vision 2030 and launched more environmental programs. The results indicate the importance of policy context and language in the formation of sustainability awareness and show the usefulness of Google Trends as a useful cross-national sustainability and marketing research tool. Full article

Clothing classification by occasion is an important area in computer vision and artificial intelligence (AI). This task is particularly challenging because of the subtle visual similarities among clothing categories such as formal, party, and casual attire. Variations in color, fabric, patterns, and lighting further increase the complexity of this task. To address this challenge, we used the Fashionpedia dataset to create a balanced subset of 15,000 images. Specifically, we adopted two different methods for labeling these images: automated classification, which relies on category identifications (IDs) and components, and manual labeling performed by human annotators. We then implemented our preprocessing pipeline, which includes several steps: resizing, image normalization, background removal using segmentation masks, and class balancing. We benchmarked traditional models, including artificial neural networks (ANNs), support vector machines (SVMs), and k-nearest neighbors (KNNs), which use a histogram of oriented gradient (HOG) features, as well as deep learning models such as convolutional neural networks (CNNs), the Visual Geometry Group 16 (VGG16) model utilizing transfer learning, and the vision transformer (ViT) model, all evaluated using identical data splits and preprocessing procedures. The traditional models achieved moderate accuracy, ranging from 54% to 66%. In contrast, the ViT model achieved an accuracy of 81.78% with automated classification and 98.09% with manual labeling. This indicates that a higher label accuracy, along with the preprocessing steps used, significantly enhances the performance. Together, these factors improve the effectiveness of ViT in context-aware apparel classification and establish a reliable baseline for future research. Full article

Energy transition is most often examined through the lens of technological development and integration, including renewable energy sources, energy storage systems, and digital energy management solutions. In practice, however, the actual performance of energy systems—understood as both energy efficiency and environmental impact across the life cycle—depends not only on technical parameters but also on decision-making processes, operational practices, and management capabilities. This paper aims to conceptualize energy and environmental awareness as a determinant influencing energy system performance at organizational and system levels. The study is based on a structured review of the literature from energy engineering, life cycle assessment, and energy management, complemented by a comparative analysis of how similar energy technologies are utilized under different decision-making contexts. On this basis, an integrated analytical framework is proposed that combines conventional energy and environmental performance indicators with awareness-related dimensions, including energy knowledge, perception of environmental risk, and managerial competence. The analysis demonstrates that insufficient energy awareness leads to systematic gaps between the technological potential of energy systems and their actual performance, resulting in increased environmental burdens despite high nominal technical efficiency. The proposed framework helps to explain performance variability in energy systems operating under comparable technical conditions and highlights the importance of incorporating managerial and competency-related factors into life cycle assessments and energy transition policies. The paper contributes to the literature by extending energy system evaluation beyond purely technical criteria and offers practical implications for the design of energy systems, industrial energy management, and policy instruments supporting sustainable energy transitions. Full article

Occupational sedentary behavior presents a public health risk, yet current interventions often rely on subjective self-reports or context-blind prompts. This study validates a privacy-preserving, edge-computing time-of-flight (ToF) sensor that detects postural states and quantifies therapeutic exercise gestures in real time. The dual-sensor architecture distinguishes between sitting, standing, and absence, while capturing rapid sit-to-stand repetitions suitable for active-break interventions. In this paper, a laboratory study (N = 7) evaluated the system against ground truth comprising activPAL3 accelerometry and video analysis. Across 378 postural events, the sensor achieved high temporal fidelity (mean absolute error < 1.6 s) and 100% sensitivity in counting exercise repetitions. The system differentiated workstation occupancy from physical absence. These findings demonstrate that ToF sensing matches the accuracy of video analysis without privacy concerns while offering the contextual awareness required for just-in-time, adaptive workplace interventions. Full article

The evolution toward 6G communications requires integrating Tiny Machine Learning (TinyML) principles to enable intelligent, energy-efficient, and adaptable signal processing at the network edge. However, current receiver architectures face a fundamental trade-off: classical model-driven designs, while naturally efficient due to their basis in communication theory, lack the flexibility to adapt to varying channel conditions. Meanwhile, fully data-driven deep-learning-based approaches break the stringent resource constraints of TinyML. This paper introduces the tiny neural receiver (TNR), a pioneering architecture that bridges these paradigms by integrating model-based signal processing with lightweight neural optimization to overcome this challenge. The TNR’s primary contribution is its unique hybrid design, which combines the efficiency and interpretability of traditional theory-based receivers with the ability to adapt to different contexts using trainable neural components. This integration occurs within resource budgets that align with TinyML specifications. Experimental results show that the TNR achieves a 5 dB SNR reduction at a target block error rate of ${10}^{-4}$. The reported 5 dB SNR gain is a direct result of our resource-aware design framework, which selectively applies lightweight neural optimization to only the most impactful receiver blocks (channel estimation and decoding) to maximize gain without exceeding TinyML complexity limits. This achievement is further supported by an end-to-end training protocol that uses 15,000 iterations of over-the-air data to fine-tune these parameters for the specific static 3.5 GHz propagation channel and OFDM configuration evaluated. Furthermore, the TNR’s modular design enables flexible deployment across a range of 6G scenarios, from mobile broadband to mission-critical IoT. This establishes the TNR as a promising framework for AI-native 6G receivers. Full article

Background/Objectives: Cancer-related terminology is not merely descriptive and plays a critical role in shaping emotional responses, personal identity, and communication across clinical, social, and public spheres. Despite growing interest in the psychosocial dimensions of illness language, few studies have centered the lived experiences of individuals navigating cancer through the lens of terminology. This study explores how people living with and beyond cancer perceive, interpret, and emotionally respond to cancer-related language, focusing on the way terminology influences identity, stigma, and communicative interaction. Methods: A sequential mixed-methods design was employed. The quantitative phase involved 146 participants with a cancer diagnosis completing a structured questionnaire on preferred terminology and emotional impact. The qualitative phase followed, using open-ended questionnaires with 11 participants to deepen understanding of linguistic experiences. Thematic content analysis was used to identify patterns across narratives. Results: These findings reveal that labels such as “cancer patient” evoke strong negative emotional reactions, associated with stigma, fear, and identity reduction. Person-first and context-sensitive language was perceived as more respectful and empowering. Emotional responses to language varied widely, from fear to neutrality, shaped by speaker role, context, and time since diagnosis. Media representations were often seen as dramatizing or moralizing, reinforcing the need for communicative clarity, empathy, and education in both clinical and public discourse. Conclusions: Cancer-related language is a powerful psychosocial force. It shapes how individuals are seen and see themselves and can either reinforce stigma or foster dignity and resilience. This study highlights the urgent need for person-centered, context-aware communication practices across healthcare, media, and society. Full article

Learning Analytics (LA) can support student success through dashboards and early-support interventions, but adoption depends on students’ willingness to allow educational data use under privacy and data-protection requirements. This study examines predictors of students’ willingness to allow educational data use for LA in higher education, focusing on perceived benefits, perceived risks, control and transparency expectations, and institutional trust. A cross-sectional survey was administered to engineering students (N = 109); after an instructed-response attention check, N = 102 valid responses were retained. Composite Likert constructs (BENEFIT, RISK, CONTROL, TRANSPARENCY, TRUST) and two willingness outcomes were analyzed: academic-support LA (WILL_ACAD) and broader aggregated institutional reporting under safeguards (WILL_BROAD). Willingness was high in both scenarios, and the paired difference did not reach statistical significance. Regression models showed that institutional trust was the strongest predictor of willingness across both use cases; perceived benefits additionally predicted willingness for academic-support LA, while perceived risk was a positive predictor in the broader-use model. Descriptive results indicated that students prioritize human review before any action affecting a student and strong security measures as key safeguards. These provide initial evidence to inform privacy-aware learning analytics governance in similar technical-university contexts; broader generalization across higher education requires replication across disciplines and institutions. Full article

Social media platforms have become critical communication environments during disasters, where individuals express emotions, share information, and engage in public discourse. These platforms also reflect heterogeneous communication patterns shaped by different actor groups. However, existing studies predominantly focus on emotion classification and often overlook the combined role of actor identity and conflict dynamics. To address this gap, this study proposes an integrated AI-based analytical framework for actor-aware emotion and conflict analysis in post-disaster social media. An expert-annotated Turkish tweet dataset was constructed based on Ekman’s emotion model, including anger, fear, sadness, happiness, and surprise, along with an additional irrelevant/off-topic category and conflict-level labels. A Transformer-based model (BERTurk) was fine-tuned for multi-class emotion classification. Experimental results show that the proposed model achieves strong classification performance, with an accuracy of 0.931 and an F1-score of 0.912, outperforming conventional machine learning and deep learning baselines. Actor-based analysis reveals systematic differences in emotional and conflict patterns across groups. Scientists, journalists, and individual users exhibit higher levels of conflict and more pronounced negative emotional expressions, whereas institutionally oriented actors display comparatively balanced and supportive communication patterns. In addition, a web-based decision support system was developed to enable interactive visualization and actor-level exploration of emotional and conflict dynamics. Overall, the proposed framework provides a scalable, analytically robust approach to understanding social media discourse in disaster contexts and offers practical implications for AI-driven crisis communication and decision-support systems. Full article

The increasing risk of air pollution in closed areas like passenger vehicles requires smart and real-time air quality reading solutions. Gases such as carbon monoxide (CO)—which is colorless and odorless and is produced by exhaust systems—air conditioners, and combustion sources are very dangerous to health because they can cause respiratory distress and poisoning at high levels. Traditional in-vehicle CO monitoring systems use a single-point sensor and a fixed threshold, which are insufficient in a dynamic cabin environment subject to factors such as vehicle size, ventilation rate, number of occupants, and incoming traffic. To address these drawbacks, this paper proposes a new E-Nose system with Virtual Sensor-Augmented Embedded Intelligence to estimate the CO concentration in vehicle cabins in real time. The system combines data from cheap gas sensors and improves it using virtual sensor machine learning models trained to predict or enhance sensor responses in real time. Embedded intelligence, deployed locally on edge hardware, supports low-latency processing, dynamic calibration, and noise filtering to respond to fluctuating environmental conditions adaptively. This architecture enables more accurate, robust, and context-aware estimation of CO levels compared to traditional threshold-based methods. Experimental validation across varied vehicular scenarios demonstrates superior precision and responsiveness, providing timely warnings even under complex dispersion patterns. Classifier Gradient Boosting, which builds an ensemble of weak learners sequentially, matched the Random Forest with 99.94% training and 98.59% model accuracy, confirming its strong predictive capability. The system is designed to be cost-effective, scalable, and easily integrable into modern automotive platforms. This study also contributes to the field of smart ecological recording and demonstrates the effectiveness of the virtual sensor-enhanced embedded system as an effective way to improve passenger safety by providing pre-emptive on-board air quality monitoring. Full article

Multi-label deep learning models are widely used in real-world applications where predictions depend on the joint presence of several semantically correlated labels. However, existing adversarial attacks largely overlook these inter-label dependencies, often perturbing outputs indiscriminately and producing structurally implausible or easily detectable changes. This paper presents CAPG (Context-Aware Perturbation Generation), a white-box, label-space targeted adversarial framework for generating selective and contextually consistent perturbations in multi-label settings. CAPG incorporates correlation-weighted regularization into the adversarial objective, enabling targeted manipulation of specific labels while preserving the contextual integrity of non-target outputs. Using the Pascal VOC 2012 dataset and a ResNet-101 multi-label classifier, we show that CAPG achieves higher Attack Success Rates (ASR) and substantially improved Contextual Consistency Scores (CCSs) than FGSM, PGD, CW, and DeepFool under identical perturbation budgets. CAPG also produces lower perceptual distortion, yielding adversarial examples that better preserve contextual structure. These results highlight the importance of correlation-aware adversarial evaluation for assessing the robustness of modern multi-label deep learning systems. Full article