Search Results (9,537)

Underwater vessel localization using passive magnetic anomaly sensing is a challenging problem due to the variability in vessel magnetic signatures and operational conditions. Data-based approaches may fail to generalize even to slightly different conditions. Thus, we propose an Attentive Neural Process (ANP) approach, in order to take advantage of its few-shot capabilities to generalize, for robust localization of underwater vessels based on magnetic anomaly measurements. Our ANP models the mapping from multi-sensor magnetic readings to position as a stochastic function: it cross-attends to a variable-size set of context points and fuses these with a global latent code that captures trajectory-level factors. The decoder outputs a Gaussian over coordinates, providing both point estimates and well-calibrated predictive variance. We validate our approach using a comprehensive dataset of magnetic disturbance fields, covering 64 distinct vessel configurations (combinations of varying hull sizes, submersion depths (water-column height over a seabed array), and total numbers of available sensors). Six magnetometer sensors in a fixed circular arrangement record the magnetic field perturbations as a vessel traverses sinusoidal trajectories. We compare the ANP against baseline multilayer perceptron (MLP) models: (1) base MLPs trained separately on each vessel configuration, and (2) a domain-randomized search (DRS) MLP trained on the aggregate of all configurations to evaluate generalization across domains. The results demonstrate that the ANP achieves superior generalization to new vessel conditions, matching the accuracy of configuration-specific MLPs while providing well-calibrated uncertainty quantification. This uncertainty-aware prediction capability is crucial for real-world deployments, as it can inform adaptive sensing and decision-making. Across various in-distribution scenarios, the ANP halves the mean absolute error versus a domain-randomized MLP (0.43 m vs. 0.84 m). The model is even able to generalize to out-of-distribution data, which means that our approach has the potential to facilitate transferability from offline training to real-world conditions. Full article

Currently, the development of adeno-associated virus (AAV)-based gene therapy is a promising method for treating various diseases and is gaining increasing popularity. However, the use of AAV has certain drawbacks and faces limitations such as immune responses and an increased risk of insertional mutagenesis, which have not always been adequately considered in the context of AAV therapy. Moreover, a significant limitation for the application of AAV lies in the challenge of producing it in large quantities. This article discusses the use of AAV in treating various diseases, reviews AAV production approaches, highlights challenges with insufficient viral titers during production, and explores potential solutions at key stages of AAV drug production. Full article

Focal adhesion kinase (FAK) is a crucial protein component of focal adhesions (FAs) and belongs to the cytoplasmic non-receptor protein tyrosine kinase family. FAK primarily regulates adhesion signaling and cell migration and is highly expressed in various tumors, including lung, liver, gastric, and colorectal cancers, as well as in conditions such as acute lung injury (ALI) and pulmonary fibrosis (PF). Recent research on FAK and its small-molecule inhibitors has revealed that targeting FAK provides a novel approach for treating various lung diseases. FAK inhibitors can obstruct signaling pathways, demonstrating anti-tumor, anti-inflammatory, and anti-fibrotic effects. In lung cancer, FAK inhibitors suppress tumor growth and metastasis; in ALI, they exert protective effects by alleviating inflammatory responses and oxidative stress; and in pulmonary fibrosis, FAK inhibitors reduce fibroblast activation and inhibit collagen deposition. The findings demonstrate promising efficacy and an acceptable safety profile in preclinical models. However, these early-stage results require further validation through clinical studies. Additionally, the underlying mechanisms, as well as the toxic effects and side effects, necessitate further in-depth investigation. Some have progressed to clinical trials (Defactinib (Phase II), PF-562271 (Phase I), CEP-37440 (Phase I), PND-1186 (Phase I), GSK-2256098 (Phase II), BI-853520 (Phase I)), offering potential therapeutic targets for lung diseases. Collectively, these findings establish a foundational basis for the advancement of FAK inhibitor discovery. Emerging methodologies, such as PROTAC degraders and combination regimens, demonstrate significant potential for future research. Based on a comprehensive analysis of the relevant literature from 2015 to the present, this review briefly introduces the structure and function of FAK and discusses recent research advancements regarding FAK and its inhibitors in the context of pulmonary diseases. Full article

This study presents a validated drone-based methodology for inspecting fire protection belts in Galicia, Spain, with a focus on secondary protection belts surrounding settlements. Current manual inspection methods are limited by resource constraints and inefficiency, especially given Galicia’s steep slopes and fragmented, vegetated terrain. Our integrated approach combines high-resolution drone imagery, RTK positioning, GIS tools, and the Time2Parcel algorithm, enabling synchronized, parcel-level documentation at cadastral scale and allowing office-based technicians to directly review automatically generated video segments specific to each parcel for inspection verification. The methodology employs a hybrid classification system: automated assessments via orthophoto and LiDAR analysis and manual verification for cases with low confidence scores. Government technicians can perform office-based reviews without GIS expertise; the system automatically matches video to cadastral records, eliminating manual video review. Key results include the Time2Parcel algorithm for automatic video-to-parcel correlation, completion of inspections for 4934 parcels, and an operational efficiency increase of 68–70% reduction in inspection time compared with traditional methods. This workflow enables faster, safer, and more accurate inspections in highly fragmented rural contexts, improving legal compliance and environmental management. Full article

Sound source localization (SSL) adds a spatial dimension to auditory perception, allowing a system to pinpoint the origin of speech, machinery noise, warning tones, or other acoustic events, capabilities that facilitate robot navigation, human–machine dialogue, and condition monitoring. While existing surveys provide valuable historical context, they typically address general audio applications and do not fully account for robotic constraints or the latest advancements in deep learning. This review addresses these gaps by offering a robotics-focused synthesis, emphasizing recent progress in deep learning methodologies. We start by reviewing classical methods such as time difference of arrival (TDOA), beamforming, steered-response power (SRP), and subspace analysis. Subsequently, we delve into modern machine learning (ML) and deep learning (DL) approaches, discussing traditional ML and neural networks (NNs), convolutional neural networks (CNNs), convolutional recurrent neural networks (CRNNs), and emerging attention-based architectures. The data and training strategy that are the two cornerstones of DL-based SSL are explored. Studies are further categorized by robot types and application domains to facilitate researchers in identifying relevant work for their specific contexts. Finally, we highlight the current challenges in SSL works in general, regarding environmental robustness, sound source multiplicity, and specific implementation constraints in robotics, as well as data and learning strategies in DL-based SSL. Also, we sketch promising directions to offer an actionable roadmap toward robust, adaptable, efficient, and explainable DL-based SSL for next-generation robots. Full article

This study investigated the dynamic resolution of lexical–semantic ambiguity during sentence comprehension, focusing on how uncertainty evolves as contextual information accumulates. Using time-resolved eye-tracking and a novel entropy-based measure derived from group-level semantic choice distributions, we quantified semantic uncertainty at a fine-grained temporal resolution for ambiguous words. By parametrically manipulating the semantic bias strength of the sentence context, we examined how context guides disambiguation over time. The results showed that semantic uncertainty declined gradually over temporal segments and dropped sharply following the onset of ambiguous words, reflecting both incremental integration and syntactic anchoring. A stronger contextual bias led to faster reductions in uncertainty, with effects following a near-linear trend. These findings support dynamic semantic processing models that assume continuous, context-sensitive convergence toward intended meanings. In contrast, a pretrained Chinese BERT model (RoBERTa-wwm-ext) showed similar overall trends in uncertainty reduction but lacked sensitivity to contextual bias. This discrepancy suggests that, while language models can approximate human-level disambiguation broadly, they fail to capture fine-grained semantic modulation driven by context. These findings provide a novel empirical characterization of disambiguation dynamics and offer a new methodological approach to capturing real-time semantic uncertainty. The observed divergence between human and model performance may inform future improvements to language models and contributes to our understanding of possible architectural differences between human and artificial semantic systems. Full article

Game commentary enhances viewer immersion and understanding, particularly in football video games, where dynamic gameplay offers ideal conditions for automated commentary. The existing methods often rely on predefined templates and game state inputs combined with an LLM, such as GPT-3.5. However, they frequently suffer from repetitive phrasing and delayed responses. Recent studies have attempted to mitigate the response delays by employing traditional machine learning models, such as SVM and ANN, for event prediction. Nonetheless, these models fail to capture the temporal dependencies in gameplay sequences, thereby limiting their predictive performance. To address these limitations, an integrated framework is proposed, combining a lightweight convolutional model with multi-scale temporal filters (OS-CNN) for real-time event prediction and an open-source LLM (LLaMA 3.3) for dynamic commentary generation. Our method incorporates prompt engineering techniques by embedding predicted events into contextualized instruction templates, which enables the LLM to produce fluent and diverse commentary tailored to ongoing gameplay. Evaluated in the Google Research Football environment, the proposed method achieved an F1-score of 0.7470 in the balanced setting, closely matching the best-performing GRU model (0.7547) while outperforming SVM (0.5271) and Transformer (0.7344). In the more realistic Balanced–Imbalanced setting, it attained the highest F1-score of 0.8503, substantially exceeding SVM (0.4708), GRU (0.7376), and Transformer (0.5085). Additionally, it enhances the lexical diversity (Distinct-2: +32.1%) and reduces the phrase repetition by 42.3% (Self-BLEU), compared with template-based generation. These results demonstrate the effectiveness of our approach in generating context-aware, low-latency, and natural commentary suitable for real-time deployment in football video games. Full article

In recent years, the increasing complexity of production systems driven by technological development has created new opportunities in the industrial world but has also brought challenges in the practical use of these systems by operators. One of the biggest changes is data existence and its accessibility. This work proposes an IoT architecture specifically designed for real-world industrial environments. The goal is to present a system that can be effectively implemented to monitor operations and production processes in real time. This solution improves fault detection and identification, giving the operators the critical information needed to make informed decisions. The IoT architecture is implemented in two different industrial applications, demonstrating the flexibility of the architecture across various industrial contexts. It highlights how the system is monitored to reduce downtime when a fault occurs, making clear the loss in performance and the fault that causes this loss. Additionally, this approach supports human operators in a deeper understanding of their working environment, enabling them to make decisions based on real-time data. Full article

Artificial intelligence (AI) plays a pivotal role in Industry 4.0, with predictive maintenance (PdM) emerging as a core application for improving operational efficiency by reducing unplanned downtime and extending asset life. Despite these advancements, the black-box nature of AI models remains a significant barrier to adoption, as industry stakeholders require systems that are both transparent and trustworthy. This study presents a systematic literature review examining how human-in-the-loop explainable AI (HITL-XAI) approaches can enhance the effectiveness and adoption of AI systems in PdM contexts. This review followed the PRISMA methodology, employing predefined search strings across Scopus, ProQuest, and EBSCO databases. Sixty-three peer-reviewed journal articles, published between 2019 and early 2025, were included in the final analysis. The selected studies span various domains, including industrial manufacturing, energy, and transportation, with findings synthesized through both descriptive and thematic analyses. A key gap identified is the limited empirical exploration of generative AI (GenAI) in improving the usability, interpretability, and trustworthiness of HITL-XAI systems in PdM applications. This review outlines actionable insights for integrating explainability and GenAI into existing rule-based PdM systems to support more adaptive and reliable maintenance strategies. Ultimately, the findings underscore the importance of designing HITL-XAI systems that not only demonstrate high model performance but are also effectively aligned with operational workflows and the cognitive needs of maintenance personnel. Full article

Objectives: In localized upper-tract urothelial carcinoma (UTUC), radical nephroureterectomy (RNU) represents the surgical gold standard, but kidney-sparing surgery (KSS) offers an alternative. The surgical perspective, including complications, remains understudied in this context. This study aimed to compare KSS and RNU, assess kidney function and survival, and identify the surgical risk factors. Methods: This retrospective analysis included UTUC patients undergoing KSS (n = 46) or RNU (n = 46) at a single center from 2016 to April 2024, matched by propensity scores. The primary endpoint was Clavien–Dindo complications. Other endpoints included Days Alive and Out of the Hospital within 30 days (DAOH30), changes in the eGFR, cancer-specific survival (CSS), and disease-free survival (DFS). A UTUC Surgery Risk Score was developed to identify the surgical risk factors for severe complications. Results: KSS was significantly associated with higher rates of Clavien–Dindo grades ≥ 3 (KSS: 14; RNU: 3). DAOH30 was significantly longer following RNU. The UTUC Surgery Risk Score, based on a non-endoscopic KSS approach, an ASA score ≥ 3, and preoperative creatinine > 0.9 mg/dL, was significantly associated with overall and severe complications and DAOH30 (both p < 0.001). KSS showed significantly better early postoperative eGFR preservation (+0.55 mL/min vs. −4.3 mL/min for RNU, p = 0.015). No significant differences were observed in the median CSS or DFS between the groups. Conclusions: KSS is associated with a higher rate of certain postoperative complications, but offers superior kidney function preservation, with comparable oncological outcomes to RNU. The novel UTUC Surgery Risk Score can aid in patient counseling and personalized decision-making prior to surgery. Full article

Metaphor detection is challenging in natural language processing (NLP) because it requires recognizing nuanced semantic shifts beyond literal meaning, and conventional models often falter when contextual cues are limited. We propose a method to enhance metaphor detection by augmenting input sentences with auxiliary context generated by ChatGPT. In our approach, ChatGPT produces semantically relevant sentences that are inserted before, after, or on both sides of a target sentence, allowing us to analyze the impact of context position and length on classification. Experiments on three benchmark datasets (MOH-X, VUA_All, VUA_Verb) show that this context-enriched input consistently outperforms the no-context baseline across accuracy, precision, recall, and F1-score, with the MOH-X dataset achieving the largest F1 gain. These improvements are statistically significant based on two-tailed t-tests. Our findings demonstrate that generative models can effectively enrich context for metaphor understanding, highlighting context placement and quantity as critical factors. Finally, we outline future directions, including advanced prompt engineering, optimizing context lengths, and extending this approach to multilingual metaphor detection. Full article

Due to their close connection to the working world, digital transformation is particularly important for vocational schools. To ensure the sustainable integration of digital media into everyday school life, a holistic school improvement approach is necessary. In this context, school leadership plays a key role as the initiator and driver of relevant development processes. This study first examines the current development state of the digital transformation in vocational schools in Switzerland. Building on this, it investigates whether there are relations between the digital status quo and innovative school leadership practices. The data were collected in spring 2023 and the sample consists of 320 school management members from 135 vocational schools. The findings indicate that the digital development status of vocational schools in Switzerland is generally assessed positively. Based on the assessments of their schools’ digital development status, three distinct profiles of school management members emerge: those perceiving their schools as digitally advanced, digitally average, or having digital development potential. Innovative leadership practices are more common among school management members who perceive their schools as more digitally advanced. The study also reveals differences between language regions and financial resources depending on the stage of digitalization-related development. The results highlight the crucial role of school leadership in promoting digital transformation. Finally, education policy measures—such as language-region-specific support programs—are discussed. Full article

Human biomonitoring (HBM) is a vital tool for assessing chemical exposure in populations and informing evidence-based public health policy. For smaller countries such as Latvia, establishing a national HBM program presents specific challenges, including limited prior experience, national data gaps, and resource constraints. This study explores the expert experiences and reflections gathered during the development of Latvia’s national HBM chemical prioritization process. Semi-structured interviews were conducted with eight experts who were directly involved in evaluating and selecting substances for inclusion in the program. The focus of this study is not on the outcomes of the prioritization itself—published elsewhere—but rather on the strategies applied, challenges encountered, and lessons learned in navigating the prioritization process. A qualitative content analysis identified several key themes, including limitations in data availability, institutional coordination challenges, differences in expert opinion, and the complexity of adapting international methodologies to the national context. Despite these obstacles, the process benefitted from interdisciplinary collaboration, iterative methodological refinement, and the strategic use of international frameworks. The findings offer practical insights for countries with limited resources that are initiating or refining their national HBM programs. This study highlights the importance of national data infrastructure, stakeholder engagement, and tailored methodological approaches to ensure an effective and context-sensitive prioritization process. Full article

This study presents a comprehensive comparison between classical sampling-based motion planners from the Open Motion Planning Library (OMPL) and a learning-based planner based on Soft Actor–Critic (SAC) for motion planning in industrial robotic arms. Using a UR3e robot equipped with an RG2 gripper, we constructed a large-scale dataset of over 100,000 collision-free trajectories generated with MoveIt-integrated OMPL planners. These trajectories were used to train a DRL agent via curriculum learning and expert demonstrations. Both approaches were evaluated on key metrics such as planning time, success rate, and trajectory smoothness. Results show that the DRL-based planner achieves higher success rates and significantly lower planning times, producing more compact and deterministic trajectories. Time-optimal parameterization using TOPPRA ensured the dynamic feasibility of all trajectories. While classical planners retain advantages in zero-shot adaptability and environmental generality, our findings highlight the potential of DRL for real-time and high-throughput motion planning in industrial contexts. This work provides practical insights into the trade-offs between traditional and learning-based planning paradigms, paving the way for hybrid architectures that combine their strengths. Full article

This study conducts a simulation-based assessment of a recently commissioned office building in the Republic of Korea, representing a typical public office facility. The building was modeled using EnergyPlus 23.1.0 after construction, although no validation was performed due to the absence of metered consumption data. Previous approaches relying on simplified methods such as the Radiant Time Series (RTS), which neglect dynamic building behavior, have often led to overestimated cooling and heating loads. This has emerged as a major obstacle in designing energy-efficient buildings within the context of compact and smart cities pursuing carbon neutrality. Consequently, the trend in building performance analysis is shifting toward dynamic simulations and digital twin-based design methodologies. Furthermore, electrification of buildings without adequate thermal load assessment may also contribute to overdesign, irrespective of urban environmental characteristics. From an urban planning standpoint, there is a growing need for performance criteria that reflect occupant behavior and actual usage patterns. However, dynamics-based building studies remain scarce in the Republic of Korea. In this context, the present study demonstrates that passive design strategies, implemented through systematic changes in envelope materials, HVAC operational standards, and compliance with ASHRAE 90.1 criteria, can significantly enhance thermal comfort and indoor air quality. The simulation results show that energy consumption can be reduced by over 36.21% without compromising occupant health or comfort. These findings underscore the importance of thermal load understanding prior to electrification and highlight the potential of LEED-aligned passive strategies for achieving high-performance, low-energy buildings. Full article