Search Results (3,129)

Robotic search of people in human-centered environments, including healthcare settings, is challenging, as autonomous robots need to locate people without complete or any prior knowledge of their schedules, plans, or locations. Furthermore, robots need to be able to adapt to real-time events that can influence a person’s plan in an environment. In this paper, we present MLLM-Search, a novel zero-shot person search architecture that leverages multimodal large language models (MLLM) to address the mobile robot problem of searching for a person under event-driven scenarios with varying user schedules. Our approach introduces a novel visual prompting method to provide robots with spatial understanding of the environment by generating a spatially grounded waypoint map, representing navigable waypoints using a topological graph and regions by semantic labels. This is incorporated into an MLLM with a region planner that selects the next search region based on the semantic relevance to the search scenario and a waypoint planner that generates a search path by considering the semantically relevant objects and the local spatial context through our unique spatial chain-of-thought prompting approach. Extensive 3D photorealistic experiments were conducted to validate the performance of MLLM-Search in searching for a person with a changing schedule in different environments. An ablation study was also conducted to validate the main design choices of MLLM-Search. Furthermore, a comparison study with state-of-the-art search methods demonstrated that MLLM-Search outperforms existing methods with respect to search efficiency. Real-world experiments with a mobile robot in a multi-room floor of a building showed that MLLM-Search was able to generalize to new and unseen environments. Full article

Federated semi-supervised learning (Fed-SSL) has emerged as a powerful framework that leverages both labeled and unlabeled data distributed across clients. To reduce communication overhead, real-world deployments often adopt partial client participation, where only a subset of clients is selected in each round. However, under non-i.i.d. data distributions, the choice of client sampling strategy becomes critical, as it significantly affects training stability and final model performance. To address this challenge, we propose a novel federated averaging semi-supervised learning algorithm, called FedAvg-SSL, that considers two sampling approaches, uniform random sampling (standard Monte Carlo) and a structured lattice-based sampling, inspired by quasi-Monte Carlo (QMC) techniques, which ensures more balanced client participation through structured deterministic selection. On the client side, each selected participant alternates between updating the global model and refining the pseudo-label model using local data. We provide a rigorous convergence analysis, showing that FedAvg-SSL achieves a sublinear convergence rate with linear speedup. Extensive experiments not only validate our theoretical findings but also demonstrate the advantages of lattice-based sampling in federated learning, offering insights into the interplay among algorithm performance, client participation rates, local update steps, and sampling strategies. Full article

Accurate estimation of available rooftop areas for PV power generation at the city scale is critical for sustainable energy planning and policy development. In this study, using publicly available high-resolution satellite imagery, rooftop solar energy potential in urban, rural, and industrial areas is estimated using deep learning models. In order to identify roof areas, high-resolution open-source images were manually labeled, and the training dataset was trained with DeepLabv3+ architecture. The developed model performed roof area detection with high accuracy. Model outputs are integrated with a user-friendly interface for economic analysis such as cost, profitability, and amortization period. This interface automatically detects roof regions in the bird’s-eye -view images uploaded by users, calculates the total roof area, and classifies according to the potential of the area. The system, which is applied in 81 provinces of Turkey, provides sustainable energy projections such as PV installed capacity, installation cost, annual energy production, energy sales revenue, and amortization period depending on the panel type and region selection. This integrated system consists of a deep learning model that can extract the rooftop area with high accuracy and a user interface that automatically calculates all parameters related to PV installation for energy users. The results show that the DeepLabv3+ architecture and the Adam optimization algorithm provide superior performance in roof area estimation with accuracy between 67.21% and 99.27% and loss rates between 0.6% and 0.025%. Tests on 100 different regions yielded a maximum roof estimation accuracy IoU of 84.84% and an average of 77.11%. In the economic analysis, the amortization period reaches the lowest value of 4.5 years in high-density roof regions where polycrystalline panels are used, while this period increases up to 7.8 years for thin-film panels. In conclusion, this study presents an interactive user interface integrated with a deep learning model capable of high-accuracy rooftop area detection, enabling the assessment of sustainable PV energy potential at the city scale and easy economic analysis. This approach is a valuable tool for planning and decision support systems in the integration of renewable energy sources. Full article

Social network platforms have a big impact on the development of companies by influencing clients’ behaviors and sentiments, which directly affect corporate reputations. Analyzing this feedback has become an essential component of business intelligence, supporting the improvement of long-term marketing strategies on a larger scale. The implementation of powerful sentiment analysis models requires a comprehensive and in-depth examination of each stage of the process. In this study, we present a new comparative approach for several feature extraction techniques, including TF-IDF, Word2Vec, FastText, and BERT embeddings. These methods are applied to three multilingual datasets collected from hotel review platforms in the tourism sector in English, French, and Arabic languages. Those datasets were preprocessed through cleaning, normalization, labeling, and balancing before being trained on various machine learning and deep learning algorithms. The effectiveness of each feature extraction method was evaluated using metrics such as accuracy, F1-score, precision, recall, ROC AUC curve, and a new metric that measures the execution time for generating word representations. Our extensive experiments demonstrate significant and excellent results, achieving accuracy rates of approximately 99% for the English dataset, 94% for the Arabic dataset, and 89% for the French dataset. These findings confirm the important impact of vectorization techniques on the performance of sentiment analysis models. They also highlight the important relationship between balanced datasets, effective feature extraction methods, and the choice of classification algorithms. So, this study aims to simplify the selection of feature extraction methods and appropriate classifiers for each language, thereby contributing to advancements in sentiment analysis. Full article

Studying consumers’ willingness to pay (WTP) for sustainable certification labels and preferences in consumption is a relevant issue for policymakers. Several studies have revealed a positive WTP a premium price for many certified products. The aim of this paper is to assess an overview of Italian consumers’ WTP for eight different sustainable certification labels and to collect information about their consumption preferences and perceptions in consumption. Participants were selected by stratified simple random sampling, using regional distribution, gender, and age as stratification criteria, to obtain a representative sample of n = 3600. Eight ordered logit models were estimated to understand how consumer sociodemographic characteristics influence the price premium. The results show important differences in WTP among different certification labels; a higher WTP emerged for ethical certifications than for environmentally focused labels. Younger individuals; women; and those with higher education, income and life satisfaction, as well as consumers in southern regions, were significantly more willing to pay premiums for certified products. However, a key finding for policymakers is that the stated price premium consumers are willing to pay falls significantly short of the actual higher costs of these products in supermarkets. Furthermore, insights into consumer perceptions and preferences revealed that quality and origin are perceived as key price drivers, while method of production holds less importance. It also emerged that consumers primarily seek a balance between quality and price, with only a small segment prioritizing certified products. Full article

The standard genetic code (SGC) plays a fundamental role in encoding biological information, but its evolutionary origins remain unresolved and widely debated. Thus, we used a methodology based on the evolutionary algorithm to investigate the emergence of stable coding systems. The simulation began with a population of varied primitive genetic codes that ambiguously encoded only a limited set of amino acids (labels). These codes underwent mutation, modeled by dynamic reassignment of labels to codons, gradual incorporation of new amino acids, and information exchange between themselves. Then, the best codes were selected using a specific fitness function F that measured the accuracy of reading genetic information and coding potential. The evolution converged towards stable and unambiguous coding systems with a higher coding capacity facilitating the production of more diversified proteins. A crucial factor in this process was the exchange of encoded information among evolving codes, which significantly accelerated the emergence of genetic systems capable of encoding 21 labels. The findings shed light on key factors that may have influenced the development of the current genetic code structure. Full article

Phishing attacks pose significant risks to security, drawing considerable attention from both security professionals and customers. Despite extensive research, the current phishing website detection mechanisms often fail to efficiently diagnose unknown attacks due to their poor performances in the feature selection stage. Many techniques suffer from overfitting when working with huge datasets. To address this issue, we propose a feature selection strategy based on a convolutional graph network, which utilizes a dataset containing both labels and features, along with hyperparameters for a Support Vector Machine (SVM) and a graph neural network (GNN). Our technique consists of three main stages: (1) preprocessing the data by dividing them into testing and training sets, (2) constructing a graph from pairwise feature distances using the Manhattan distance and adding self-loops to nodes, and (3) implementing a GraphSAGE model with node embeddings and training the GNN by updating the node embeddings through message passing from neighbors, calculating the hinge loss, applying the softmax function, and updating weights via backpropagation. Additionally, we compute the neighborhood random walk (NRW) distance using a random walk with restart to create an adjacency matrix that captures the node relationships. The node features are ranked based on gradient significance to select the top k features, and the SVM is trained using the selected features, with the hyperparameters tuned through cross-validation. We evaluated our model on a test set, calculating the performance metrics and validating the effectiveness of the PhishGNN dataset. Our model achieved a precision of 90.78%, an F1-score of 93.79%, a recall of 97%, and an accuracy of 93.53%, outperforming the existing techniques. Full article

Background and Objectives: Peptide receptor radionuclide therapy (PRRT) using radiopharmaceuticals labelled with Lutetium-177 is currently a therapeutic option for patients with advanced neuroendocrine neoplasms overexpressing somatostatin receptors (SSTRs). One promising option that has gained interest for PRRT is using alpha-emitting radioisotopes such as Actinium-225. The aim of this study was to perform a systematic review and meta-analysis on the efficacy and safety of radioligand therapy with Actinium-225 DOTATATE in advanced, metastatic or inoperable neuroendocrine neoplasms. Materials and Methods: A comprehensive literature search of studies on radioligand therapy with Actinium-225 DOTATATE in neuroendocrine neoplasms was carried out. Three different bibliographic databases (Cochrane Library, Embase, and PubMed/MEDLINE) were screened up to May 2025. Eligible articles were selected, relevant data were extracted, and the main findings on efficacy and safety are summarized through a systematic review. Furthermore, proportional meta-analyses on the disease response rate and disease control rate were performed. Results: Five studies (153 patients) published from 2020 were included in the systematic review. The pooled disease response rate and disease control rate of radioligand therapy using Actinium-225 DOTATATE were 51.6% and 88%, respectively. This treatment was well-tolerated in most patients with advanced, metastatic or inoperable neuroendocrine neoplasms. Conclusions: Radioligand therapy with Actinium-225 DOTATATE in advanced, metastatic or inoperable neuroendocrine neoplasms is effective with an acceptable toxicity profile and potential advantages compared with SSTR-ligands labelled with Lutetium-177. Currently, the number of published studies on this treatment is still limited, and results from multicenter randomized controlled trials are needed to translate this therapeutic option into clinical practice. Full article

Glycated hemoglobin (HbA1c) is an important biomarker applied for the diagnosis, evaluation, and management of diabetes; therefore, its accurate determination is crucial. In this study, an innovative nanoplatform was developed, integrating carbon nanotubes (CNTs) with enhanced hydrophilicity achieved through cyclodextrin (CD) functionalization, and combined with gold nanoparticles (AuNPs) electrochemically deposited onto a screen-printed carbon electrode. The nanomaterials significantly improved the analytical performance of the sensor due to their increased surface area and high electrical conductivity. This nanoplatform was employed as a substrate for the covalent attachment of thiolated ferrocene-labeled HbA1c specific aptamer through Au-S binding. The electrochemical signal of ferrocene was covered by a stronger oxidation peak of Fe²⁺ from the HbA1c structure, leading to the elaboration of a nanostructured aptasensor capable of the direct detection of HbA1c. The electrochemical aptasensor presented a very wide linear range (0.688–11.5%), an acceptable limit of detection (0.098%), and good selectivity and stability, being successfully applied on real samples. This miniaturized, simple, easy-to-use, and fast-responding aptasensor, requiring only a small sample volume, can be considered as a promising candidate for the efficient on-site determination of HbA1c. Full article

Ergonomic risk classification during manual lifting tasks is crucial for the prevention of workplace injuries. This study addresses the challenge of classifying lifting task risk levels (low, medium, and high risk, labeled as 0, 1, and 2) using multi-modal time-series data comprising raw facial landmarks and bio-signals (electrocardiography [ECG] and electrodermal activity [EDA]). Classifying such data presents inherent challenges due to multi-source information, temporal dynamics, and class imbalance. To overcome these challenges, this paper proposes a Multi-Adaptive Functional Neural Network (Multi-AdaFNN), a novel method that integrates functional data analysis with deep learning techniques. The proposed model introduces a novel adaptive basis layer composed of micro-networks tailored to each individual time-series feature, enabling end-to-end learning of discriminative temporal patterns directly from raw data. The Multi-AdaFNN approach was evaluated across five distinct dataset configurations: (1) facial landmarks only, (2) bio-signals only, (3) full fusion of all available features, (4) a reduced-dimensionality set of 12 selected facial landmark trajectories, and (5) the same reduced set combined with bio-signals. Performance was rigorously assessed using 100 independent stratified splits (70% training and 30% testing) and optimized via a weighted cross-entropy loss function to manage class imbalance effectively. The results demonstrated that the integrated approach, fusing facial landmarks and bio-signals, achieved the highest classification accuracy and robustness. Furthermore, the adaptive basis functions revealed specific phases within lifting tasks critical for risk prediction. These findings underscore the efficacy and transparency of the Multi-AdaFNN framework for multi-modal ergonomic risk assessment, highlighting its potential for real-time monitoring and proactive injury prevention in industrial environments. Full article

Estradiol (E2) plays an important role in cell proliferation and certain brain functions. To reveal its mechanism of action, its detectability is essential. Only a few fluorescent-labeled hormonally active E2s exist in the literature, and their mechanism of action usually remains unclear. It would be of particular interest to develop novel labeled estradiol derivatives with retained biological activity and improved optical properties. Due to their superior optical characteristics, aza-BODIPY dyes are frequently used labeling agents in biomedical applications. E2 was labeled with the aza-BODIPY dye at its phenolic hydroxy function via an alkyl linker and a triazole coupling moiety. The estrogenic activity of the newly synthesized fluorescent conjugate was evaluated via transcriptional luciferase assay. Docking calculations were performed for the classical and alternative binding sites (CBS and ABS) of human estrogen receptor α. The terminal alkyne function was introduced into the tetraphenyl aza-BODIPY core via selective formylation, oxidation, and subsequent amidation with propargyl amine. The conjugation was achieved via Cu(I)-catalyzed azide–alkyne click reaction of the aza-BODIPY-alkyne with the 3-O-(4-azidobut-1-yl) derivative of E2. The labeled estrogen induced a dose-dependent transcriptional activity of human estrogen receptor α with a submicromolar EC₅₀ value. Docking calculations revealed that the steroid part has a perfect overlap with E2 in ABS. In CBS, however, a head-tail binding deviation was observed. A facile, fluorescent labeling methodology has been elaborated for the development of a novel red-emitting E2 conjugate with substantial estrogenic activity. Docking experiments uncovered the binding mode of the conjugate in both ABS and CBS. Full article

In remote sensing building detection tasks, data acquisition remains a critical bottleneck that limits both model performance and large-scale deployment. Due to the high cost of manual annotation, limited geographic coverage, and constraints of image acquisition conditions, obtaining large-scale, high-quality labeled datasets remains a significant challenge. To address this issue, this study proposes an automatic semantic labeling framework for remote sensing imagery. The framework leverages geospatial vector data provided by OpenStreetMap, precisely aligns it with high-resolution satellite imagery from Bing Maps through projection transformation, and incorporates a quality-aware sample filtering strategy to automatically generate accurate annotations for building detection. The resulting dataset comprises 36,647 samples, covering buildings in both urban and suburban areas across multiple cities. To evaluate its effectiveness, we selected three publicly available datasets—WHU, INRIA, and DZU—and conducted three types of experiments using the following four representative object detection models: SSD, Faster R-CNN, DETR, and YOLOv11s. The experiments include benchmark performance evaluation, input perturbation robustness testing, and cross-dataset generalization analysis. Results show that our dataset achieved a mAP at 0.5 intersection over union of up to 93.2%, with a precision of 89.4% and a recall of 90.6%, outperforming the open-source benchmarks across all four models. Furthermore, when simulating real-world noise in satellite image acquisition—such as motion blur and brightness variation—our dataset maintained a mean average precision of 90.4% under the most severe perturbation, indicating strong robustness. In addition, it demonstrated superior cross-dataset stability compared to the benchmarks. Finally, comparative experiments conducted on public test areas further validated the effectiveness and reliability of the proposed annotation framework. Full article

This study analyzes the use of a lightweight image-based deep learning model to classify dairy cows into low-, medium-, and high-milk-yield categories by automatically detecting the udder region of the cow. The implemented model was based on the YOLOv11 architecture, which enables efficient object detection and classification with real-time performance. The model is trained on a public dataset of cow images labeled with 305-day milk yield records. Thresholds were established to define the three yield classes, and a balanced subset of labeled images was selected for training, validation, and testing purposes. To assess the robustness and consistency of the proposed approach, the model was trained 30 times following the same experimental protocol. The system achieves precision, recall, and mean Average Precision (mAP@50) of 0.408 ± 0.044, 0.739 ± 0.095, and 0.492 ± 0.031, respectively, across all classes. The highest precision (0.445 ± 0.055), recall (0.766 ± 0.107), and mAP@50 (0.558 ± 0.036) were observed in the low-yield class. Qualitative analysis revealed that misclassifications mainly occurred near class boundaries, emphasizing the importance of consistent image acquisition conditions. The resulting model was deployed in a mobile application designed to support field-level assessment by non-specialist users. These findings demonstrate the practical feasibility of applying vision-based models to support decision-making in dairy production systems, particularly in settings where traditional data collection methods are unavailable or impractical. Full article

The increasing interest in fibre-enriched and functional bakery products has led to the exploration of novel plant-based ingredients with both technological functionality and consumer acceptance. This study evaluates the effects of incorporating flours derived from apple (Malus domestica cv. Oberländer Himbeerapfel), sweet chestnut (Castanea sativa), horse chestnut (Aesculus hippocastanum), and red, sessile, and pedunculate oak (Quercus rubra, Q. petraea, and Q. robur) into wheat bread at 5%, 10%, and 15% substitution levels. The impact on crumb structure, crust colour, textural parameters (hardness, adhesiveness, springiness), and sensory attributes was assessed. The inclusion of apple and sweet chestnut flours resulted in a softer crumb, lower adhesiveness, and higher sensory scores related to flavour, aroma, and crust appearance. In contrast, higher levels of oak- and horse-chestnut-derived flours increased crumb hardness and reduced overall acceptability due to bitterness or excessive density. Apple flour preserved crumb brightness and contributed to warm tones, while oak flours caused more intense crust darkening. These findings suggest that selected non-traditional flours, especially apple and sweet chestnut, can enhance the sensory and physical properties of wheat bread, supporting the development of fibre-rich, clean-label formulations aligned with consumer trends in sustainable and functional baking. Full article

Background: Semaglutide and Liraglutide are medications in the Glucagon-like peptide-1 agonists (GLP-1 RAs) class used to manage type 2 diabetes mellitus and obesity in Saudi Arabia. Although the 1.0 mg once weekly dosage of Semaglutide does not have a labeled indication for the management of obesity, many believe that this dosage is more effective than the 3.0 mg once daily Liraglutide dosage for the management of both diabetes and obesity. Objective: To compare the effectiveness of the dosage of 1.0 mg of Semaglutide administered once weekly versus 3.0 mg of Liraglutide administered once daily in controlling HbA1c levels, promoting weight loss, and evaluating their financial implications among obese patients in Saudi Arabia using real-world data. Methods: A retrospective review of Electronic Medical Records (EMRs) from January 2021 to June 2024 was conducted on patients prescribed Semaglutide or Liraglutide for at least 12 months. Exclusion criteria included pre-existing severe conditions (e.g., cardiovascular disease, stroke, or cancer) and missing baseline data. The primary outcomes assessed were changes in HbA1c, weight, and direct medical costs. Results: Two hundred patients (100 patients on the 1.0 mg once weekly dose of Semaglutide and 100 patients on the 3.0 mg once daily dose of Liraglutide) of those randomly selected from the EMRs met the inclusion criteria and were included in the analysis. Of the 200 eligible patients (65.5% female, mean age 48.54 years), weight loss was greater with Semaglutide (−8.09 kg) than Liraglutide (−5.884 kg). HbA1c reduction was also greater with Semaglutide (−1.073%) than Liraglutide (−0.298%). The use of Semaglutide resulted in lower costs of USD −1264.76 (95% CI: −1826.82 to 33.76) and greater reductions in weight of −2.22 KG (95% CI: −7.68 to −2.784), as well as lower costs of USD −1264.76 (95% CI: (−2368.16 to −239.686) and greater reductions in HbA1c of −0.77% (95% CI: −0.923 to −0.0971) in more than 95% of the cost effectiveness bootstrap distributions. Conclusions: Semaglutide 1.0 mg weekly seems to be more effective and cost-saving in managing prediabetes, diabetes, and obesity compared to Liraglutide 3.0 mg daily. Future studies should examine these findings using a more representative sample and a robust study design. Full article