Search Results (4,960)

Staining-based assays are widely used for cell analysis but are invasive, alter physiology, and prevent longitudinal monitoring. Label-free, morphology-based approaches could enable real-time, non-invasive drug testing, yet detection of subtle and dynamic changes has remained difficult. We developed a deep learning framework for stain-free monitoring of leukemia cell cultures using automated bright-field microscopy in a semi-automated culture system (AICE3, LABMaiTE, Augsburg, Germany). YOLOv8 models were trained on images from K562, HL-60, and Kasumi-1 cells, using an NVIDIA DGX A100 GPU for training and tested on GPU and CPU environments for real-time performance. Comparative benchmarking with RT-DETR and interpretability analyses using Eigen-CAM and radiomics (RedTell) was performed. YOLOv8 achieved high accuracy (mAP@0.5 > 98%, precision/sensitivity > 97%), with reproducibility confirmed on an independent dataset from a second laboratory and an AICE3 setup. The model distinguished between morphologically similar leukemia lines and reliably classified untreated versus differentiated K562 cells (hemin-induced erythroid and PMA-induced megakaryocytic; >95% accuracy). Incorporation of decitabine-treated cells demonstrated applicability to drug testing, revealing treatment-specific and intermediate phenotypes. Longitudinal monitoring captured culture- and time-dependent drift, enabling separation of temporal from drug-induced changes. Radiomics highlighted interpretable features such as size, elongation, and texture, but with lower accuracy than the deep learning approach. To our knowledge, this is the first demonstration that deep learning resolves subtle, drug-induced, and time-dependent morphological changes in unstained leukemia cells in real time. This approach provides a robust, accessible framework for label-free longitudinal drug testing and establishes a foundation for future autonomous, feedback-driven platforms in precision oncology. Ultimately, this approach may also contribute to more precise and adaptive clinical decision-making, advancing the field of personalized medicine. Full article

In a digital landscape marked by the exponential growth of cyber threats, the development of automated domain reputation systems is extremely important. Emerging technologies such as artificial intelligence and machine learning now enable proactive and scalable approaches to early identification of malicious or suspicious domains. This paper presents an adaptive domain name reputation system that integrates advanced machine learning to enhance cybersecurity resilience. The proposed framework uses domain data from .ro domain Registry and several other sources (blacklists, whitelists, DNS, SSL certificate), detects anomalies using machine learning techniques, and scores domain security risk levels. A supervised XGBoost model is trained and assessed through five-fold stratified cross-validation and a held-out 80/20 split. On an example dataset of 25,000 domains, the system attains accuracy 0.993 and F1 0.993 and is exposed through a lightweight Flask service that performs asynchronous feature collection for near real-time scoring. The contribution is a blueprint that links list supervision with registry/DNS/TLS features and deployable inference to support proactive domain abuse mitigation in ccTLD environments. Full article

During the peak load period, there is a high level of imbalance between power supply and demand, which has become a critical challenge, leading to higher operational costs for power grids. To improve the accuracy of peak load forecasting, this study introduces a novel approach based on Extreme Gradient Boosting Trees (XGBoost) and Multiple Linear Regression (MLR) for daily peak load prediction. The proposed methodology first employs an improved version of the Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (ICEEMDAN) algorithm to decompose the raw load data, subsequently reconstructing each Intrinsic Mode Function (IMF) into high-frequency and stationary components. For the high-frequency components, XGBoost serves as the base predictor within a Bagging-based ensemble structure, while the Sparrow Search Algorithm (SSA) is employed to optimize hyperparameters automatically, ensuring efficient learning and accurate representation of complex peak load fluctuations. Meanwhile, the stationary components are modeled using MLR to provide fast and reliable estimations. The proposed framework was evaluated using actual daily peak load data from Western Inner Mongolia, China. The results indicate that the proposed method successfully captures the peak characteristics of the power grid, delivering both robust and precise predictions. When compared to the baseline model, the RMSE and MAPE are reduced by 54.4% and 87.3%, respectively, underscoring its significant potential for practical applications in power system operation and planning. Full article

Wire Arc Additive Manufacturing (WAAM) has been recognized as an efficient and cost-effective metal additive manufacturing technique due to its high deposition rate and scalability for large components. However, the quality and repeatability of WAAM parts are highly sensitive to process parameters such as arc voltage, current, wire feed rate, and torch travel speed, requiring advanced monitoring and adaptive control strategies. In this study, a vision-based monitoring system integrated with a reinforcement learning framework was developed to enable intelligent in situ control of WAAM. A custom optical assembly employing mirrors and a bandpass filter allowed simultaneous top and side views of the melt pool, enabling real-time measurement of layer height and width. These geometric features provide feedback to a tabular Q-learning algorithm, which adaptively adjusts voltage and wire feed rate through direct hardware-level control of stepper motors. Experimental validation across multiple builds with varying initial conditions demonstrated that the RL controller stabilized layer geometry, autonomously recovered from process disturbances, and maintained bounded oscillations around target values. While systematic offsets between digital measurements and physical dimensions highlight calibration challenges inherent to vision-based systems, the controller consistently prevented uncontrolled drift and corrected large deviations in deposition quality. The computational efficiency of tabular Q-learning enabled real-time operation on standard hardware without specialized equipment, demonstrating an accessible approach to intelligent process control. These results establish the feasibility of reinforcement learning as a robust, data-efficient control technique for WAAM, capable of real-time adaptation with minimal prior process knowledge. With improved calibration methods and expanded multi-physics sensing, this framework can advance toward precise geometric accuracy and support broader adoption of machine learning-based process monitoring and control in metal additive manufacturing. Full article

Originating from the practical demands of digital irrigation district construction, this study aims to provide support for precise irrigation management. This study developed a reinforcement learning-based intelligent irrigation decision-making model for districts employing traditional surface flood irrigation methods. Grounded in the theoretical framework of water cycle processes within the Soil–Crop–Atmosphere Continuum (SPAC) system and incorporating district-specific irrigation management experience, the model achieves intelligent and precise irrigation decision-making through agent–environment interactive learning. Simulation results show that in the selected typical area of the irrigation district, during the 10-year validation period from 2014 to 2023, the model triggered a total of 22 irrigation events with an average annual irrigation volume of 251 mm. Among these, the model triggered irrigation 18 times during the winter wheat growing season and 4 times during the corn growing season. The intelligent irrigation decision-making model effectively captures the coupling relationship between crop water requirements during critical periods and the temporal distribution of precipitation, and achieves preset objectives through adaptive decisions such as peak-shifting preemptive irrigation in spring, limited irrigation under low-temperature conditions, no irrigation during non-irrigation periods, delayed irrigation during the rainy season, and timely irrigation during crop planting periods. These outcomes validate the model’s scientific rigor and operational adaptability, providing both a scientific water management tool for irrigation districts and a new technical pathway for the intelligent development of irrigation decision-making systems. Full article

Flood events, intensified by climate change, pose significant threats to both human settlements and ecological systems. This study presents an integrated approach to evaluate flood impacts on ecosystem carbon dynamics using remote sensing and machine learning techniques. The case of the Emilia-Romagna region in Italy is presented, which experienced intense flooding in 2023. To understand flood-induced changes in the short term, we quantified the differences in net primary productivity (NPP) and above-ground biomass (AGB) before and after flood events. Short-term analysis of NPP and AGB revealed substantial localized losses within flood-affected areas. NPP showed a net deficit of 7.0 × 10³ g C yr⁻¹, and AGB a net deficit of 0.5 × 10³ Mg C. While the wider region gained NPP (6.7 × 10⁵ g C yr⁻¹), it suffered a major AGB loss (3.3 × 10⁵ Mg C), indicating widespread biomass decline beyond the flood zone. Long-term ecological assessment using the Remote Sensing Ecological Index (RSEI) showed accelerating degradation, with the “Fair” ecological class shrinking from 90% in 2014 to just over 50% in 2024, and the “Poor” class expanding. “Good” and “Very Good” classes nearly disappeared after 2019. High-hazard flood zones were found to contain 9.0 × 10⁶ Mg C in AGB and 1.1 × 10⁷ Mg C in soil organic carbon, highlighting the vulnerability of carbon stocks. This study underscores the importance of integrating flood modeling with ecosystem monitoring to inform climate-adaptive land management and carbon conservation strategies. It represents a clear, quantifiable carbon loss that should be factored into regional carbon budgets and post-flood ecosystem assessments. Full article

Underwater Optical Camera Communication (UOCC) has emerged as a promising paradigm for short-range, high-bandwidth, and secure data exchange in autonomous underwater vehicles (AUVs). UOCC performance strongly depends on exposure time and International Standards Organization (ISO) sensitivity—two parameters that govern photon capture, contrast, and bit detection fidelity. However, optical propagation in aquatic environments is highly susceptible to turbidity, scattering, and illumination variability, which severely degrade image clarity and signal-to-noise ratio (SNR). Conventional systems with fixed imaging settings cannot adapt to time-varying conditions, limiting communication reliability. While validating the feasibility of deep learning for exposure prediction, this baseline lacked environmental awareness and generalization to dynamic scenarios. To overcome these limitations, we introduce a Real-to-Sim-to-Deployment framework that couples a physically calibrated emulation platform with a Hybrid CNN-MLP Model (HCMM). By fusing optical images, environmental states, and camera configurations, the HCMM achieves substantially improved parameter prediction accuracy, reducing RMSE to 0.23–0.33. When deployed on embedded hardware, it enables real-time adaptive reconfiguration and delivers up to 8.5 dB SNR gain, surpassing both static-parameter systems and the prior CNN baseline. These results demonstrate that environment-aware multimodal learning, supported by reproducible optical channel emulation, provides a scalable and robust solution for practical UOCC deployment in positioning, inspection, and laser-based underwater communication. Full article

An integrated approach for enhancing microparticle separation efficiency in acoustofluidic lab-on-a-chip systems is presented, combining numerical modeling in COMSOL 6.2 Multiphysics^® with reinforcement learning techniques implemented in Python 3.10.14. The proposed method addresses the limitations of traditional parameter tuning, which is time-consuming and computationally intensive. A simulation framework based on LiveLink™ for COMSOL–Python integration enables the automatic generation, execution, and evaluation of particle separation scenarios. Reinforcement learning algorithms, trained on both successful and failed experiments, are employed to optimize control parameters such as flow velocity and acoustic frequency. Experimental data from over 100 numerical simulations were used to train a neural network, which demonstrated the ability to accurately predict and improve sorting efficiency. The results confirm that incorporating failed outcomes into the reward structure significantly improves learning convergence and model accuracy. This work contributes to the development of intelligent microfluidic systems capable of autonomous adaptation and optimization for biomedical and analytical applications, such as label-free separation of microplastics from biological fluids, selective sorting of soot and ash particles for environmental monitoring, and high-precision manipulation of cells or extracellular vesicles for diagnostic assays. Full article

This study examines how large-scale school infrastructure reforms shape teaching practice, using Australia’s Building the Education Revolution (BER) initiative as a case example. Guided by Cultural-Historical Activity Theory (CHAT), the research explores how redesigned learning environments act as mediating tools that influence pedagogy, collaboration, and teacher wellbeing. An explanatory sequential mixed-methods design was employed, combining survey data from 34 teachers with focus group interviews involving 13 participants in a redeveloped Victorian Primary School, Australia. Quantitative results showed that 70.5% of teachers reported changes in their teaching practices directly linked to the new infrastructure, with 100% affirming that they had enhanced collaboration opportunities. Qualitative findings revealed that features such as breakout rooms, shared learning zones, and transparent sightlines enabled differentiated instruction, co-teaching, and improved supervision, while also fostering professional pride and collegial support. Contradictions emerged around automated lighting systems, limited display space, and partial teacher consultation during the design process. CHAT analysis demonstrated how physical spaces interact with rules, community, and division of labour within the school activity system, producing both enabling conditions and systemic tensions. The study underscores the need for infrastructure planning to be pedagogically informed, inclusive of teacher voice, and designed to support adaptive, collaborative, and inclusive teaching practices. Full article

The sine cosine algorithm (SCA) is a population-based stochastic optimization method that updates the position of each search agent using the oscillating properties of the sine and cosine functions to balance exploration and exploitation. While flexible and widely applied, the SCA often suffers from premature convergence and getting trapped in local optima due to weak exploration–exploitation balance. To overcome these issues, this study proposes a multi-faceted SCA (MFSCA) incorporating several improvements. The initial population is generated using dynamic opposition (DO) to increase diversity and global search capability. Chaotic logistic maps generate random coefficients to enhance exploration, while an elite-learning strategy allows agents to learn from multiple top-performing solutions. Adaptive parameters, including inertia weight, jumping rate, and local search strength, are applied to guide the search more effectively. In addition, Lévy flights and adaptive Gaussian local search with elitist selection strengthen exploration and exploitation, while reinitialization of stagnating agents maintains diversity. The developed MFSCA was tested against 23 benchmark optimization functions and assessed using the Wilcoxon rank-sum and Friedman rank tests. Results showed that MFSCA outperformed the original SCA and other variants. To further validate its applicability, this study developed a fuzzy c-means MFSCA-based adaptive neuro-fuzzy inference system to forecast energy consumption in student residences, using student apartments at a university in South Africa as a case study. The MFSCA-ANFIS achieved superior performance with respect to RMSE (1.9374), MAD (1.5483), MAE (1.5457), CVRMSE (42.8463), and SD (1.9373). These results highlight MFSCA’s effectiveness as a robust optimizer for both general optimization tasks and energy management applications. Full article

Physics-Informed Neural Networks (PINNs) have emerged as a promising paradigm for solving partial differential equations (PDEs) by embedding physical laws into the learning process. However, standard PINNs often suffer from training instabilities and unbalanced optimization when handling multi-term loss functions, especially in problems involving singular perturbations, fractional operators, or multi-scale behaviors. To address these limitations, we propose a novel gradient variance weighting physics-informed neural network (GVW-PINN), which adaptively adjusts the loss weights based on the variance of gradient magnitudes during training. This mechanism balances the optimization dynamics across different loss terms, thereby enhancing both convergence stability and solution accuracy. We evaluate GVW-PINN on three representative PDE models and numerical experiments demonstrate that GVW-PINN consistently outperforms the conventional PINN in terms of training efficiency, loss convergence, and predictive accuracy. In particular, GVW-PINN achieves smoother and faster loss reduction, reduces relative errors by one to two orders of magnitude, and exhibits superior generalization to unseen domains. The proposed framework provides a robust and flexible strategy for applying PINNs to a wide range of integer- and fractional-order PDEs, highlighting its potential for advancing data-driven scientific computing in complex physical systems. Full article

This study investigates the applicability of the Japanese forest education model, exemplified by the Whole Earth Nature School (WENS), to environmental education programs within urban parks and botanical gardens. Focusing on WENS’s approach—characterized by immersive natural experiences, innovative facilitation techniques, and support from interdisciplinary educator teams to foster independent thinking through experiential learning—the research identifies three transferable insights: (1) Reorienting educational philosophy to prioritize nature-based learning for holistic development, particularly within urban green spaces; (2) Developing localized curricula that leverage the unique ecological resources of specific park and garden environments; (3) Establishing dedicated professional development systems for educators focused on site-based pedagogy. The findings demonstrate that adapting the forest education model can significantly enhance environmental education initiatives in urban parks and botanical gardens, which offers actionable strategies for integrating experiential nature learning into their educational planning. Full article

In increasingly networked environments, artificial agents are required to operate not with fixed roles but with identities that adapt, evolve, and emerge through interaction. Traditional identity modeling approaches, whether symbolic or statistical, fail to capture this dynamic, relational nature. This paper proposes a network-based framework for constructing and analyzing AI identity by modeling interaction, representation, and emergence within complex networks. The goal is to uncover how agent identity can be inferred and explained through structural roles, temporal behaviors, and community dynamics. The approach begins by transforming raw data from three benchmark domain, Reddit, the Interaction Network dataset, and AMine, into temporal interaction graphs. These graphs are structurally enriched via motif extraction, centrality scoring, and community detection. Graph Neural Networks (GNNs), including GCNs, GATs, and GraphSAGE, are applied to learn identity embeddings across time slices. Extensive evaluations include identity coherence, role classification accuracy, and temporal embedding consistency. Ablation studies assess the contribution of motif and temporal layers. The proposed model achieves strong performance across all metrics. On the AMiner dataset, identity coherence reaches 0.854, with a role classification accuracy of 80.2%. GAT demonstrates the highest temporal consistency and resilience to noise. Role trajectories and motif patterns confirm the emergence of stable and transient identities over time. The results validate the fact that the framework is not only associated with healthy quantitative performance but also offers information on behavioral development. The model will be expanded with semantic representations and be more concerned with ethical considerations, such as privacy, fairness, and transparency, to make identity modeling in artificial intelligence systems responsible and trustworthy. Full article

The global food crises and growing population necessitate efficient agricultural land use. Weeds cause up to 40% yield loss in major crops, resulting in over USD 100 billion in annual economic losses. Camera-equipped UAVs offer a solution for automatic weed detection, but the high computational and energy demands of deep learning models limit their use to expensive, high-end UAVs. In this paper, we present a low-cost UAV system built from off-the-shelf components, featuring a custom-designed on-board computing system based on the NVIDIA Jetson Nano. This system efficiently manages real-time image acquisition and inference using the energy-efficient Squeeze U-Net neural network for weed detection. Our approach ensures the pipeline operates in real time without affecting the drone’s flight autonomy. We also introduce the AgriAdapt dataset, a novel collection of 643 high-resolution aerial images of salad crops with weeds, which fills a key gap by providing realistic UAV data for benchmarking segmentation models under field conditions. Several deep learning models are trained and validated on the newly introduced AgriAdapt dataset, demonstrating its suitability for effective weed segmentation in UAV imagery. Quantitative results show that the dataset supports a range of architectures, from larger models such as DeepLabV3 to smaller, lightweight networks like Squeeze U-Net (with only 2.5 M parameters), achieving high accuracy (around 90%) across the board. These contributions distinguish our work from earlier UAV-based weed detection systems by combining a novel dataset with a comprehensive evaluation of accuracy, latency, and energy efficiency, thus directly targeting deep learning applications for real-time UAV deployment. Our results demonstrate the feasibility of deploying a low-cost, energy-efficient UAV system for real-time weed detection, making advanced agricultural technology more accessible and practical for widespread use. Full article

In today’s digital ecosystem, where millions of users interact with diverse online services and generate vast amounts of textual, transactional, and behavioral data, ensuring the trustworthiness of this information has become a critical challenge. Low-quality data—manifesting as incompleteness, inconsistency, duplication, or noise—not only undermines analytics and machine learning models but also exposes unsuspecting users to unreliable services, compromised authentication mechanisms, and biased decision-making processes. Traditional data quality assessment methods, largely based on manual inspection or rigid rule-based validation, cannot cope with the scale, heterogeneity, and velocity of modern data streams. To address this gap, we propose DQMAF (Data Quality Modeling and Assessment Framework), a generalized machine learning–driven approach that systematically profiles, evaluates, and classifies data quality to protect end-users and enhance the reliability of Internet services. DQMAF introduces an automated profiling mechanism that measures multiple dimensions of data quality—completeness, consistency, accuracy, and structural conformity—and aggregates them into interpretable quality scores. Records are then categorized into high, medium, and low quality, enabling downstream systems to filter or adapt their behavior accordingly. A distinctive strength of DQMAF lies in integrating profiling with supervised machine learning models, producing scalable and reusable quality assessments applicable across domains such as social media, healthcare, IoT, and e-commerce. The framework incorporates modular preprocessing, feature engineering, and classification components using Decision Trees, Random Forest, XGBoost, AdaBoost, and CatBoost to balance performance and interpretability. We validate DQMAF on a publicly available Airbnb dataset, showing its effectiveness in detecting and classifying data issues with high accuracy. The results highlight its scalability and adaptability for real-world big data pipelines, supporting user protection, document and text-based classification, and proactive data governance while improving trust in analytics and AI-driven applications. Full article