Industry 5.0 reframes manufacturing around human-centric concerns: resilient operations, safe work, and decisions people can understand and contest. For PHM, that means elevating human-based features: competency, recency of practice, mentoring links, explainability, and fair exposure, rather than relying only on sensors or opaque models. Today those signals sit in ticket logs and massive databases, making them hard to audit, transfer, or reuse at scale.We suggest ONGOING, a representation layer framework that turns unstructured maintenance text into a human-auditable Knowledge Grid and a complex but modular feature vector, independent of any particular embedding model or projector. At its core, the grid tracks technician experiences by incrementing a part of the Knowledge Grid whenever tickets are resolved. Two mechanisms capture more advanced dynamics: knowledge transfer between people (e.g., mentorship) via a convex blend of Knowledge Grids, and neighborhood propagation that diffuses experience increases to semantically adjacent tasks through a Gaussian kernel. From each grid we derive interpretable features, such as hypervolume, sparsity, or maximum knowledge, that summarize knowledge distribution more accurately for better downstream use (e.g., dispatching optimizer models, LLMs, production forecast models).We implement the framework on a partner company's data, and deploy an instance at-scale (50000 tickets, 100 technicians) in real-time, using a multilingual sentence encoder and a toroidal SOM for ticket embedding.On our deployed instance, we designed a technician recommendation use-case. A maintenance expert study with human feedback over 55 real tickets found that grid-based recommendation were judged more pertinent than a scalar-based and a vector-based knowledge modeling approaches. Crucially, dispatchers could articulate rationales from visible grid neighborhoods and feature attributions, preserving interpretability.Beyond dispatch support, the Knowledge Grid enables training planning (identify coverage gaps), fairness monitoring (avoid single-point failure through over-reliance on “heroes”), and promotes workload balancing.

Nighttime cultural districts concentrate dense human activity, intense illumination, and continuous operations, which can generate pronounced nocturnal thermal anomalies that remain insufficiently examined in urban heat island (UHI) research. This study analyzes how nocturnal UHIs form and propagate across five representative music-cultural districts in Changsha, China, between 2013 and 2024. Using multi-source satellite observations in combination with XGBoost, principal component regression, structural equation modeling, and a Gaussian diffusion kernel, we reconstructed high-resolution nocturnal land surface temperature (LST) patterns and assessed the drivers of spatial variability. Results show a steady citywide warming trend, peaking in 2022 with a mean nocturnal LST of 17.03°C (maximum 21.32°C), accompanied by increasing spatial heterogeneity. Music-cultural districts showed heterogeneous thermal risk profiles, commercial corridors such as Jiefang West Road were consistently 1.0-2.1°C hotter than surrounding areas, while river- and park-adjacent districts such as Orange Island and Meixihu functioned as persistent cold islands (- 2.0 to - 2.8°C). Diffusion analysis revealed corridor-like propagation with gradients of 1.5-4.2°C km- 1 and effective stochastic diffusion coefficients of 1.30-1.49 m2 s- 1, indicating an influence radius of approximately 2-3km. Model-based uncertainty and attribution analyses highlighted the dominant role of building density and nighttime lighting (combined importance > 0.6) as risk-enhancing factors, whereas vegetation coverage and water proximity provided robust buffers. Mitigation scenario simulations suggest that isolated measures yield limited benefits (0.25-0.65°C cooling), while integrated strategies-expanding green-blue infrastructure, enhancing surface albedo, and moderating nighttime traffic-reduce diffusion intensity by 28-35% with cooling effects up to 1.8°C. These findings demonstrate that nocturnal UHI propagation not only shapes local microclimates but also generates uneven patterns of heat exposure risk within and around music-cultural districts. By treating music-cultural districts as a nightlife-related urban unit, this study offers an in-depth characterization of their nocturnal thermal dynamics, diffusion behavior, and policy-relevant mitigation levers. The proposed framework provides transferable evidence for climate-adaptive planning and public health risk mitigation in cities with vibrant nighttime economies.

Abstract Traditional extended Kalman filter is a popular method in vehicle state parameters estimation. However, its performance undergoes significant degradation when encountering non-Gaussian noise or noise of uncertain statistical properties. The maximum correntropy extended Kalman filter (MCEKF) based on Gaussian kernel can handle some non-Gaussian noise scenarios, but it is still less effective in dealing with noise of uncertain statistical characteristics and is typically sensitive to manually selected kernel bandwidths. To overcome these constraints, this paper proposes a new adaptive extended Kalman filtering algorithm based on the Cauchy-kernel maximum correntropy criterion (CKMCC), referred to as the Cauchy-kernel maximum correntropy adaptive extended Kalman filter (CKMCAEKF). By integrating the CKMCC and an adaptive method, the proposed algorithm eliminates sensitivity to selection of kernel bandwidth and enhances robustness against non-Gaussian noise and noise of uncertain statistical characteristics. Moreover, the robustness of the proposed method is theoretically proven. The effectiveness of the proposed method is validated through co-simulations on the MATLAB/Simulink and CarSim platforms. CKMCAEKF achieves superior estimation accuracy for vehicle state parameters to other existing methods under non-Gaussian measurement noise scenarios and scenarios of changes in statistical characteristics of measurement noise caused by sudden events.

The advancement of digital technology has made users increasingly reliant on online services, with user reviews serving as an essential resource for evaluating the quality of service provided by companies such as FirstMedia. However, these valuable data have not undergone comprehensive analysis to assess users’ emotional responses. This study aims to classify FirstMedia customers’ emotions into four categories (joy, sadness, anger, and neutral) and to evaluate the Support Vector Machine (SVM) method using four different kernel functions. Most existing studies primarily focus on polarity-based sentiment analysis and do not explicitly examine multi-emotion classification or kernel comparison in machine learning models. A total of 4,001 reviews were collected through web scraping from the Google Play Store and the X app and processed through several preprocessing steps. Emotion classification was conducted using the NRC Indonesian Emotion Lexicon, while word significance was determined using TF-IDF weighting. After preprocessing, 3,069 labeled reviews were retained and distributed as 1,065 neutral, 748 anger, 692 joy, and 564 sadness reviews, which were used for emotion classification. Model performance was evaluated using a hold-out validation scheme with an 80:20 train-test split and assessed through a confusion matrix. To address class imbalance, undersampling was applied, resulting in a balanced dataset for model training. The evaluation results show that the Linear kernel achieved the highest performance, with an accuracy of 82.63%, precision of 82.86%, recall of 82.63%, and an F1-score of 82.60%, outperforming the Gaussian, Polynomial, and Sigmoid kernels. This study demonstrates that multi-emotion sentiment analysis provides a more comprehensive understanding of user perceptions beyond conventional sentiment polarity, thereby supporting more informed evaluations of digital service quality.

Abstract Balancing exploration and exploitation is a fundamental challenge in informative path planning for environmental monitoring. Although numerous reward functions have been proposed in the literature, most have been evaluated under different datasets and experimental conditions, making direct comparison difficult. The novelty of this study lies in its development of four synthetic datasets, experimentally validated and designed with increasing spatial complexity (1 to 4 Regions of Interest, ROIs), to enable a fair and systematic comparison of three widely used Gaussian Process-based reward functions: Entropy, Upper Confidence Bound (UCB), and Level Set. The proposed framework integrates a greedy local path optimization algorithm that maximizes expected reward and incorporates a cross-validation strategy to reduce initial model variance and mitigate overfitting. Importantly, this study not only compares the individual performances of these reward functions but also analyzes how each one contributes to the trade-off between exploration and exploitation under varying environmental conditions. Experimental results show that Level Set performs best in high-variance environments (favoring exploration), UCB excels in low-variance settings with fast convergence (favoring exploitation), and Entropy provides stable long-term uncertainty reduction (balancing both aspects). With the inclusion of cross-validation, the model achieves up to 60% reduction in RMSE and 50% reduction in variance across all scenarios. These findings highlight the practical value of reward-aware path planning in robotic exploration tasks, particularly when aligned with the spatial complexity of the monitoring environment.

Universal machine learning interatomic potentials (uMLIPs) have recently been formulated and shown to generalize well. When applied out-of-sample, further data collection for improvement of the uMLIPs may, however, be required. In this work we demonstrate that, whenever the envisaged use of the MLIPs is global optimization, the data acquisition can follow an active learning scheme in which a gradually updated uMLIP directs the finding of new structures, which are subsequently evaluated at the density functional theory (DFT) level. In the scheme, we augment foundation models using a Δ-model based on this new data using local SOAP-descriptors, Gaussian kernels, and a sparse Gaussian process regression model. We compare the efficacy of the approach with different global optimization algorithms, random structure search, basin hopping, a Bayesian approach with competitive candidates (GOFEE), and a replica exchange formulation (REX). We further compare several foundation models, CHGNet, MACE-MP0, and MACE-MPA. The test systems are silver-sulfur clusters and sulfur-induced surface reconstructions on Ag(111) and Ag(100). Judged by the fidelity of identifying global minima, active learning with GPR-based Δ-models appears to be a robust approach. Judged by the total CPU time spent, the REX approach stands out as being the most efficient.

Radial basis function neural networks (RBFNNs) are widely applied due to their rapid modeling capabilities and efficient learning performance. However, when dealing with high-dimensional data, RBFNNs encounter two critical limitations: the hidden layer responses using Gaussian kernels suffer from ineffective activation and numeric underflow; and the estimation of output layer weights typically involves tedious parameter tuning and inefficient loading of high-dimensional feature matrices. To overcome these challenges, we first propose a dimensionality-adaptive Gaussian kernel function (DAGKF) equipped with a novel width adjustment mechanism that flexibly mitigates the numerical difficulties inherent in high-dimensional spaces. Moreover, to avoid processing entire feature matrices simultaneously, we introduce a multioutput coordinate descent (MOCD) algorithm that enables parallel computation across multioutput systems. Building upon MOCD, we further develop the joint residual MOCD (JRMOCD) algorithm, which incorporates a joint residual criterion for more effective weight estimation. The convergence of the JRMOCD algorithm is rigorously proven. Extensive experiments demonstrate the superior performance of the proposed methods, particularly in high-dimensional settings.

Enhancing patient engagement is essential for effective post-stroke robotic rehabilitation, yet limited effort has been made towards modulating and quantifying patient engagement during therapy. To Bridge this gap, we introduce a virtual reality (VR)-integrated robot-assisted system for upper limb rehabilitation, which innovatively enables simultaneous modulation and monitoring of user engagement in a line tracing task. Modulation is governed by two parameters: shape complexity and force noise disturbance level. Our system estimates engagement using physiological (GSR and pupil diameter) and behavioral (eye blink and gaze) indicators, benchmarked against the Game Engagement Questionnaire (GEQ). The study involved twenty healthy right-handed subjects. Results show that behavioral signals are more informative in predicting engagement than physiological signals, which were the focus of most prior efforts in estimating engagement. Our detailed analysis identifies an optimal 11-second window-initiated no earlier than 15 seconds into the trial-that yields the most accurate engagement metrics for estimation (MAE = 0.73, r = 0.42). Consistent with Mihaly Csikszentmihalyi's flow theory, the estimated engagement is maximized when task difficulty matches the user's skill level, with peak engagement modeled as a Gaussian function (R2 = 0.76, RMSE = 0.18). Taken together, this study confirms the potential of behavioral measurements for reliable, non-invasive engagement estimation during task performance, paving the way for adaptive systems that automatically adjust task difficulty to enhance patient engagement throughout the rehabilitation process.

<b>Aims</b> : High-resolution spectrographs are central to modern exoplanet research and are particularly effective for detecting Earth-like planets whose radial velocity (RV) signals can be only a few tens of centimeters per second. Achieving this level of precision requires highly accurate wavelength calibration. A key factor in this process is the modeling of the instrumental profile (IP), which describes the response of the spectrograph to incoming light. The true IP of a high-resolution instrument is often complex. It may show asymmetry or extended wings and change across the detector because of optical aberrations, variations in fiber illumination, and environmental effects. These features lead to systematic errors in the measured line centers when traditional parametric models such as Gaussian functions are used, and they limit the achievable RV precision.<br> <b>Methods:</b> This work introduces a non-parametric IP modeling method based on Gaussian Process Regression (GPR). The IP is treated as a smooth function with a flexible covariance structure instead of being constrained by a predefined analytic form. GPR learns both the global structure and small-scale features of the line shape directly from the data. Since the IP varies slowly across the detector, the method divides each spectral order into several consecutive spatial segments. Each segment is fitted independently, capturing local variations. The model includes measurement uncertainties and provides a probabilistic description of the IP. Adjacent segments are linked with smooth interpolation to ensure a continuous IP across the entire order. Model performance is evaluated using reduced chi-squared and root mean square error (RMSE), allowing quantitative assessment and comparison with traditional approaches.<br> <b>Results:</b> The method is tested with laser frequency comb (LFC) exposures from the fiber-fed High Resolution Spectrograph (HRS) on the 2.16 m telescope at Xinglong Observatory. The LFC produces a dense and highly stable set of emission lines and is well suited for validating IP reconstruction. Three experiments show clear and consistent improvements. Using odd-numbered lines to predict evennumbered ones within a single exposure reduces the RMSE by 35.6% compared with a Gaussian model, showing better determination of line centers. Applying an IP model trained on one exposure to a later exposure reduces the RMSE by 42.5%, demonstrating improved stability when the model is transferred between exposures. A comparison between two channels in the same exposure shows a 37.1% improvement in calibration consistency, indicating reduced channel-tochannel systematics.<br> <b>Conclusions:</b> The results show that GPR provides a more accurate description of the instrumental profile and its spatial variation than traditional parametric models. The improved reconstruction of the IP leads to more accurate line center measurements and a more stable and precise wavelength solution. This capability is important for pushing the RV precision of high-resolution spectrographs toward the centimeter-per-second level. GPR offers a promising approach for modeling instrumental profiles and supports the precision required for detecting Earth-like exoplanets.

Numerical predictions of cracking evolution on RAP-recycled asphalt mixtures using viscoelastic cohesive zone model with Gaussian damage function

Deep Fuzzy-NN modeling for the prediction of Zn(II) adsorption in columns using alkaline modified biochar: Integrated experimental and computational insights.

Gaussian fuzzy membership functions with fractal belief functions and an application

Homogeneous multi-antibiotics residual identification in various actual water via SERS spectra multilayer perceptron algorithm combined with Gaussian kernel density estimation data augmentation.

Abstract This study proposes a single-sensor method for impact force localization and time-history reconstruction in the frequency domain. The method constructs a resonance-focused modal inversion framework that extracts the modal vector at the impact location from a single-point acceleration response. An adaptive mode selection strategy filters out weakly responsive modes, thereby reducing errors caused by modal truncation, nodal placement, and measurement noise. The impact force is modeled using an asymmetric Gaussian function, and its parameters are identified via particle swarm optimization to achieve fast time-domain reconstruction.The performance of the method is systematically evaluated through numerical simulations on a cantilever beam and validated by experiments on a clamped steel plate. Results demonstrate that the proposed approach achieves high localization accuracy under low signal-to-noise ratios, reconstructs the force history with only a few resonance frequencies, and supports near real-time identification with minimal hardware. This approach provides a practical solution for structural health monitoring and vibrationbased event detection in space-constrained applications.

Kernel Gaussian processes based extended target tracking in polar coordinate

This study aims to analyze the factors influencing the Human Development Index (HDI) in West Java Province using the Geographically Weighted Regression (GWR) approach. The independent variables used in this study are the Open Unemployment Rate (TPT), School Participation Rate for ages 16–18 (APS_16_18), Population Density, and Gross Regional Domestic Product per Capita (PPK). The modeling was carried out by comparing various kernel functions, namely Gaussian, Bisquare, and Tricube, as well as two bandwidth approaches: fixed and adaptive. The results indicate that the GWR model with a Gaussian kernel and a fixed bandwidth approach provides the best performance based on the lowest AIC value. Compared to the classical Ordinary Least Squares (OLS) model, the GWR model offers a better explanation of spatial variation in HDI across the study area. Although the GWR model was not statistically significant overall based on the ANOVA test, local analysis showed that the variables TPT and PPK had significant effects in all districts and cities, while APS_16_18 and Population Density were not significant in any region. These findings demonstrate that the GWR model is capable of capturing spatial heterogeneity that is not detected by the global regression model.

Fusion dynamics of 40 Ca + 58,64 Ni and 48 Ti + 58,64 Ni reactions are theoretically investigated by using Wong formula with spherical & deformed choice of nuclear potential, symmetric-asymmetric Gaussian barrier distribution (SAGBD) model and coupled channel approach. The calculations with deformed nuclear potential are found enhanced with respect to the calculations of spherical choice of nuclear potential in Wong formula but such calculations do not retrieve the fusion data. With an additional radius parameter (ΔR) in deformed nuclear potential, the estimated cross-sections approximately recover the fusion data for chosen reactions. In SAGBD model, effects due to distinct nuclear structural properties are comprehensively assimilated by adopting a Gaussian function to Wong formula and the calculations fairly retrieve all the experimental data points. Furthermore, the effects of lowlying 2 + and 3 - quantum states of fusing pairs are also included in coupled channel calculations to reproduced the fusion data of 40 Ca + 58 Ni and 48 Ti + 58,64 Ni systems. However, for 40 Ca + 64 Ni reaction, 2 + … 3 - vibrational states of both projectile & target as well as neutron transfer channel having positive Q-value is primarily required for addressal of fusion data.

Water quality is a key indicator of a community’s health and welfare, yet it has deteriorated significantly due to pollution caused by human activities. This study aimed to evaluate Geographically Weighted Logistic Regression’s (GWLR) ability to handle spatial nonstationarity in the relationship between explanatory factors and water quality status in Pontianak City, and to compare its performance with logistic regression. Three modelling approaches were applied to classify water as polluted or non-polluted: (i) logistic regression with spatially invariant) parameters; (ii) GWLR with a fixed Gaussian kernel, producing spatially varying parameters using a fixed bandwidth; and (iii) GWLR with an adaptive Gaussian kernel, producing spatially varying parameters using an adaptive bandwidth. Model performance was compared using Akaike’s Information Criterion (AIC) and classification accuracy. The GWLR model with a fixed Gaussian kernel produced an AIC of 22.52, whereas the logistic regression model produced a slightly lower AIC of 22.39; both models achieved a classification accuracy of 92.86%, with the adaptive-kernel GWLR showing comparable classification performance. These results indicate that, for the parameter settings considered, GWLR offered performance comparable to, but not substantially better than logistic regression for modelling the factors affecting water quality, despite its capacity to address spatial nonstationarity.

This study investigates the prediction of Escherichia coli growth in environments where experimental data are limited, by integrating mathematical curve fitting with machine learning regression models. Two hybrid frameworks are developed: Fourier Series Curve Fitting combined with Gaussian Process Regression (FSCF-GPR), and Gaussian Curve Fitting integrated with Support Vector Machine Regression (GCF-SVMR). The raw dataset, initially composed of only 10 experimental measurements, was expanded to 114 data points through mathematical smoothing, providing a richer basis for model training. Model performance was assessed using Root Mean Square Error (RMSE), Mean Squared Error (MSE), Coefficient of Determination (R²), and Mean Absolute Error (MAE). Results demonstrate that the FSCF-GPR framework achieved outstanding predictive accuracy with an R² of 0.9999, while GCF-SVMR also showed strong performance with an R² of 0.9934. These findings highlight that data augmentation via curve fitting can substantially enhance the accuracy and robustness of machine learning approaches in microbiological growth prediction under data-scarce conditions.

Abstract Electronic interconnects benefit from copper due to its superior conductivity and low cost. Direct-write processes are desired for flexibility, ease and agility in mesoscale, hybrid and packaging electronics manufacturing. Vacuum cold spray (VCS) is an attractive process, but depends on optimization of many parameters to obtain efficient deposition and maximum fidelity. This study uses VCS with different powder feedstocks, nozzle diameters, nozzle standoffs and scan numbers to produce copper lines and pads on glass and silicon substrates. Electron microscopy reveals plasticity-based deposition, building films to thicknesses of several microns. Profilometry and image analysis portray the line profiles, with data fit to Gaussian curves to obtain accurate heights, widths and integrated cross-sectional areas. A figure of merit (FOM), combining height, rectangularity ratio and number of scans, is used to judge the deposition and geometric form of the lines. The FOM in this study has a wide range from 3 to 61 nm/scan. Both the line FOM and rectangularity are correlated with a drop in relative electrical resistivity. A 20-scan, 50-mm-long line is found to have a low electrical resistivity = 4.34 × 10 −8 Ωm, just 2.5 times that of pure bulk copper. The results suggest that VCS copper holds promise for direct writing of interconnects, and the FOM approach is proposed for comparative studies in process development.