Learning Modeling Research Articles

Multi-agent modelling and analysis of the knowledge learning of a human-machine hybrid intelligent organization with human-machine trust

Machine learning (ML) technologies have changed the paradigm of knowledge discovery in organizations and transformed traditional organizational learning to human-machine hybrid intelligent organizational learning. However, the general distrust among humans towards knowledge derived from machine learning has hindered effective knowledge exchange between humans and machines, thereby compromising the efficiency of human-machine hybrid intelligent organizational learning. To explore this issue, we used multi-agent simulation to construct a knowledge learning model of a human-machine hybrid intelligent organization with human-machine trust. The simulation showed that whether human-machine trust has a positive effect on knowledge level depends on the initial input and the magnitude of the effect depends on the human learning propensity (exploration and exploitation). When humans reconfigure machine learning excessively, whether human-machine trust has a positive effect on the knowledge level depends on human learning propensity (exploration and exploitation). Maintaining appropriate human-machine trust in turbulent environments assists humans in integrating diverse knowledge to meet changing knowledge needs. Our study extends the human-machine hybrid intelligence organizational learning model by modeling human-machine trust. It will assist managers in effectively designing the most economical level of human-machine trust, thereby enhancing the efficiency of human-machine collaboration in human-machine hybrid intelligent organization.

FOSTERING DIGITAL TRANSFORMATIONS IN MILITARY ENGINEERING EDUCATION: INTRODUCTION OF A TECHNOLOGY-ENHANCED LEARNING ENVIRONMENT

Digital tools have become integral to training military engineers, and the introduction of a technology-enhanced learning environment may improve the educational process at higher military educational institutions. This study explores the impact of digital transformations in military engineering education through a technology-enhanced learning environment. To answer the research questions, mixed methodology was used, which involved integration and data triangulation matrices. The survey included 17 experts who possess the competences to evaluate the integration of technology into the learning environment. The results showed that a positive technology-enhanced learning environment is represented through ICT-supported learning, e-learning, blended learning, and virtual learning environments. The researchers distinguished groups of digital tools used at higher military educational institutions (communication tools, content creation tools, Learning Management Systems, collaboration tools, assessment tools, simulation and modelling, and adaptive learning tools). The impact of digital tools applied for the formation of professional competence among future military engineers was evaluated. Then, the algorithm for creating and developing a positive technology-enhanced learning environment was explained as the set of systematic measures used to design and implement a learning environment that integrates technology. The measures to be taken to foster digital technologies while training future military engineers were described during the study. The outcomes can be used to improve the existing system of military engineering education at the Ukrainian military institutions and to maximize the professional training at the active military units. Keywords: digital tools, higher military educational institution, learning environment, professional competence, technology-enhanced learning

Machine learning assisted 5-part tooth segmentation method for CBCT-based dental age estimation in adults.

The utilization of segmentation method using volumetric data in adults dental age estimation (DAE) from cone-beam computed tomography (CBCT) was further expanded by using current 5-Part Tooth Segmentation (SG) method. Additionally, supervised machine learning modelling -namely support vector regression (SVR) with linear and polynomial kernel, and regression tree - was tested and compared with the multiple linear regression model. CBCT scans from 99 patients aged between 20 to 59.99 was collected. Eighty eligible teeth including maxillary canine, lateral incisor, and central incisor were used in this study. Enamel to dentine volume ratio, pulp to dentine volume ratio, lower tooth volume ratio, and sex was utilized as independent variable to predict chronological age. No multicollinearity was detected in the models. The best performing model comes from maxillary lateral incisor using SVR with polynomial kernel ( = 0.73). The lowest error rate achieved by the model was given also by maxillary lateral incisor, with 4.86 years of mean average error and 6.05 years of root means squared error. However, demands a complex approach to segment the enamel volume in the crown section and a lengthier labour time of 45 minutes per tooth.

Application of large language models in the field of education

This paper delves into the application of generative artificial intelligence and large language models in the field of education, with a particular focus on the rising trend of large language models. Large language models play a crucial role in intelligent tutoring systems by accurately understanding student queries through deep learning and providing personalized responses. Case studies showcase the exemplary utilization of natural language processing techniques and reasoning engines, albeit facing challenges related to real-time processing and privacy concerns. The latter part of the paper concentrates on the application of generative artificial intelligence and large language models in curiosity-driven learning and the integration of multimodal educational systems, emphasizing the technical frameworks and challenges associated with multimodal integration. Finally, the paper provides insights into future developments, highlighting research on the potential benefits in the field of education, while emphasizing concerns related to ethics and privacy.

Connected in Bad Times and in Good Times: Empathy Induces Stable Social Closeness.

Humans need social closeness to prosper. There is evidence that empathy can induce social closeness. However, it remains unclear how empathy-related social closeness is formed and how stable it is as time passes. We applied an acquisition-extinction paradigm combined with computational modeling and fMRI, to investigate the formation and stability of empathy-related social closeness. Female participants observed painful stimulation of another person with high probability (acquisition) and low probability (extinction) and rated their closeness to that person. The results of two independent studies showed increased social closeness in the acquisition block that resisted extinction in the extinction block. Providing insights into underlying mechanisms, reinforcement learning modeling revealed that the formation of social closeness is based on a learning signal (prediction error) generated from observing another's pain, whereas maintaining social closeness is based on a learning signal generated from observing another's pain relief. The results of a reciprocity control study indicate that this feedback recalibration is specific to learning of empathy-related social closeness. On the neural level, the recalibration of the feedback signal was associated with neural responses in anterior insula and adjacent inferior frontal gyrus and the bilateral superior temporal sulcus/temporoparietal junction. Together, these findings show that empathy-related social closeness generated in bad times, that is, empathy with the misfortune of another person, transfers to good times and thus may form one important basis for stable social relationships.

Junior High School Student Models in Online Learning as A Post-Covid Effect

Junior High School Student Models in Online Learning as A Post-Covid Effect

Interpretable machine learning of SPAC system via a mechanism-assisted gaussian process group: Representation of the system mechanism by data

Traditional models of the soil-plant-atmosphere continuum (SPAC) system are physics-based and their practical application has been hindered by issues of high parameterization, prior structural bias, and costly running. In this paper, we developed a machine learning modelling method based on the Gaussian process (GP) to avoid these difficulties of traditional modelling. At the same time, shortcomings of conventional machine learning in terms of interpretability and mechanism representation were addressed. The innovative method consisted of structuring a local GP group framework and improving kernel functions used in each local GP. The resultant model was endowed with advanced capacities, including representing interplays between subprocesses within the complex SPAC system, representing nonstationary subprocess dynamics caused by crop growth stage shifts, as well as automatically interpreting key low-order input interactions and dominant input variables for each subprocess. The performance of our model was examined on synthetic SPAC system datasets that covered three different soil conditions. Results demonstrated that its interpretations regarding subprocess mechanisms were robust across different soil conditions and consistent with domain knowledge. Compared to a conventional global GP model and two deep learning models (DNN and LSTM), our model performed significantly better not only in regular prediction experiments but also in generalization experiments that required models to be transferred across different conditions. Our study suggested that it is promising to enable a machine learning model interpretable by improving its feature representation function. For machine learning modelling of complex systems like the SPAC, it may be more important to enhance the model’s mechanism representation capability than simply using deeper networks.

Comparison and Analysis of Chinese and English Handwriting Recognition Algorithms in Deep Learning

Handwriting recognition is an important topic of current research on image recognition. This technology provides convenience for production and life, as well as guarantees for us in some aspects of safety. The author hopes to help people get a preliminary understanding of English handwriting recognition by summarizing and comparing two commonly used models in deep learning. This paper focuses on two models of widely used English handwriting recognition, which are the simple Recurrent Neural Network (RNN) model and the Long Short-Term Memory (LSTM) model. Firstly, the principles of the two models are introduced, and then gradient problems, fitting problems, training difficulty, and accuracy, as well as the reasons for these problems are all discussed and compared within the two models. Nowadays, the accuracy rate of English handwriting recognition is relatively high using deep learning methods and models. With the development of deep learning models, if more accurate string segmentation can be achieved, the accuracy rate of handwriting recognition can also be further improved.

Comparison and Analysis of the Effectiveness of Linear Regression, Decision Tree, and Random Forest Models for Health Insurance Premium Forecasting

Health insurance is a type of insurance that covers individual and family medical expenses and is important for the health and financial security of individuals and families. To better predict the demand for health insurance, three regression models in machine learning - random forest, linear regression, and decision tree - are widely used for health insurance prediction. Among these three regression models, random forest regression has the best prediction effect with a model score of 0.8564, which is the best prediction effect among the three models. Random forest regression is an integrated learning method that combines multiple decision tree models into a more powerful model that can effectively avoid overfitting problems and can handle large amounts of data. Therefore, random forest regression is a very effective method for health insurance prediction. The next model is the linear regression model with a model score of 0.7584. The linear regression model is a basic regression model that can be used to predict a linear relationship between two variables. In health insurance prediction, linear regression modeling can be used to predict the linear relationship between health insurance costs and related factors such as age, gender, and illness. The worst predictor was the decision tree, with a model score of 0.7097. The decision tree model can be used in Medicare forecasting to predict nonlinear relationships between Medicare costs and related factors such as age, gender, and illness.

Empowering Community Clinical Triage through Innovative Data-Driven Machine Learning

Efficient triaging and referral assessments are critical in ensuring prompt medical intervention in the community healthcare (CHC) system. However, the existing triaging systems in many community health services are an intensive, time-consuming process and often lack accuracy, particularly for various symptoms which might represent heart failure or other health-threatening conditions. There is a noticeable limit of research papers describing AI technologies for triaging patients. This paper proposes a novel quantitative data-driven approach using machine learning (ML) modelling to improve the community clinical triaging process. Furthermore, this study aims to employ the feature selection process and machine learning power to reduce the triaging process’s waiting time and increase accuracy in clinical decision making. The model was trained on medical records from a dataset of patients with “Heart Failure”, which included demographics, past medical history, vital signs, medications, and clinical symptoms. A comparative study was conducted using a variety of machine learning algorithms, where XGBoost demonstrated the best performance among the other ML models. The triage levels of 2,35,982 patients achieved an accuracy of 99.94%, a precision of 0.9986, a recall of 0.9958, and an F1-score of 0.9972. The proposed diagnostic model can be implemented for the CHC decision system and be developed further for other medical conditions.

Learning spatio-temporal patterns with Neural Cellular Automata.

Neural Cellular Automata (NCA) are a powerful combination of machine learning and mechanistic modelling. We train NCA to learn complex dynamics from time series of images and Partial Differential Equation (PDE) trajectories. Our method is designed to identify underlying local rules that govern large scale dynamic emergent behaviours. Previous work on NCA focuses on learning rules that give stationary emergent structures. We extend NCA to capture both transient and stable structures within the same system, as well as learning rules that capture the dynamics of Turing pattern formation in nonlinear PDEs. We demonstrate that NCA can generalise very well beyond their PDE training data, we show how to constrain NCA to respect given symmetries, and we explore the effects of associated hyperparameters on model performance and stability. Being able to learn arbitrary dynamics gives NCA great potential as a data driven modelling framework, especially for modelling biological pattern formation.

Predicting E-commerce customer satisfaction: Traditional machine learning vs. deep learning approaches

The rapid growth of e-commerce has increased the need for retailers to understand and predict customer satisfaction to support data-driven managerial decisions. This study analyzes online consumer behavior through a comparative machine learning modeling approach to forecast future customer satisfaction based on review ratings. Using a large dataset of over 100 k online orders from a major retailer, traditional machine learning models including random forest and support vector machines are benchmarked against deep learning techniques like multi-layer perceptrons. The predictive models are assessed for their ability to accurately predict customer satisfaction scores for the next orders based on key e-commerce features including delivery time, order value, and location. The findings demonstrate that the random forest model can predict future satisfaction with 92% accuracy, outperforming deep learning. The analysis further identifies core drivers of satisfaction such as delivery time and order accuracy. These insights enable retail managers to make targeted improvements, like optimizing logistics, to increase customer loyalty and revenue. This study provides a framework for leveraging predictive analytics and machine learning to unlock data-driven insights into online consumer behavior and satisfaction for superior retail decision-making. The focus on generalizable insights across a major retailer enhances the practical applicability of the machine learning approach for the retail sector.

Aboveground Biomass Inversion Based on Object-Oriented Classification and Pearson–mRMR–Machine Learning Model

Cities play a crucial role in the carbon cycle. Measuring urban aboveground biomass (AGB) is essential for evaluating carbon sequestration. Satellite remote sensing enables large-scale AGB inversion. However, the apparent differences between forest and grassland biomass pose a significant challenge to the accurate estimation of urban AGB using satellite-based data. To address this limitation, this study proposed a novel AGB estimation method using the eastern part of the Zhahe mining area in Huaibei City as the study area, which integrates land cover classification, feature selection, and machine learning modelling to generate high quality biomass maps of different vegetation types in an urban area with complex feature distribution. Utilizing the GEE platform and Sentinel-2 image, we developed an object-oriented machine learning classification algorithm, combining SNIC and GLCM to extract vegetation information. Optimal feature variables for forest and crop-grass AGB inversion were selected using the Pearson–mRMR algorithm. Finally, we constructed nine machine learning models for AGB inversion and selected the model with the highest accuracy to generate the AGB map of the study area. The results of the study are as follows: (1) Compared with the pixel-based classification method, the object-oriented classification method can extract the boundaries of different vegetation types more accurately. (2) Forest AGB is strongly correlated with vegetation indices and physiological parameters, while agri-grass AGB is primarily associated with vegetation indices and vegetation physiological parameters. (3) For forest AGB modelling, the RF-R model outperforms other machine learning models with an R2 of 0.77. For agri-grass AGB modelling, the XGBoost-R model is more accurate, with an R2 of 0.86. (4) The mean forest AGB in the study area was 4.60 kg/m2, while the mean agri-grass AGB was 0.71 kg/m2. High AGB values were predominantly observed in forested areas, which were mainly distributed along roads, waterways, and mountain ranges. Overall, this study contributes to a better understanding of the health of local urban ecosystems and provides valuable insights for ecosystem protection and the sustainable use of natural resources.

Seeing through their eyes: Revealing recreationists’ landscape preferences through viewshed analysis and machine learning

Planning for outdoor recreation requires knowledge about the needs and preferences of recreationists. While previous research has mainly relied on stated preferences, recent advances in spatial data collection and analysis have enabled the assessments of actual usage patterns. In this study, we explored how landscape characteristics interact with the attributes of recreationists to determine their area choice for recreation. Using a public participation GIS (PPGIS) approach we asked residents of a Swedish city in the boreal region to draw typical recreational routes and identify favourite places for recreation on a digital online map (1389 routes, 385 individuals). We employed a novel methodology, where LiDAR data was used to calculate what was visible along all routes and at favourite places (viewsheds) in order to more realistically capture the landscape that each recreationist had experienced. Using machine learning modelling, we compared landscape characteristics of experienced areas with areas available to each recreationist. Our novel approach yielded accurate models that revealed that water environments, recreational infrastructure and deciduous forests increased the probability of choosing an area for recreation, while urban environments, noise, forest clearcuts and young forests had the opposite effect. Characteristics of the recreationists such as age, gender, level of education, or of the activity, such as type of activity performed, did not meaningfully influence area choice. Our findings suggest that it is possible to improve the conditions for recreation by developing recreational infrastructure, maintaining recreation opportunities close to waters, and adapting forest management in areas important for recreation.

Satellite Image Fusion Airborne LiDAR Point-Clouds-Driven Machine Learning Modeling to Predict the Carbon Stock of Typical Subtropical Plantation in China

In the current context of carbon neutrality, afforestation is an effective means of absorbing carbon dioxide. Stock can be used not only as an economic value index of forest wood resources but also as an important index of biomass and carbon storage estimation in forest emission reduction project evaluation. In this paper, we propose a data-driven machine learning framework and method for predicting plantation stock based on airborne LiDAR + satellite remote sensing, and carried out experimental verification at the site of the National Forest emission reduction project in Southern China. We used step-up regression and random forest (RF) to screen LiDAR and Landsat 8 OLI multispectral indicators suitable for the prediction of plantation stock, and constructed a plantation stock model based on machine learning (support vector machine regression, RF regression). Our method is compared with traditional statistical methods (stepwise regression and partial least squares regression). Through the verification of 57 plantation field survey data, the accuracy of the stand estimation model constructed using the RF method is generally better (ΔR2 = 0.01~0.27, ΔRMSE = 1.88~13.77 m3·hm−2, ΔMAE = 1.17~13.57 m3·hm−2). The model evaluation accuracy based on machine learning is higher than that of the traditional statistical method, and the fitting R2 is greater than 0.91, while the fitting R2 of the traditional statistical method is 0.85. The best fitting models were all support vector regression models. The combination of UAV point clouds and satellite multi-spectral images has the best modeling effect, followed by LiDAR point clouds and Landsat 8. At present, this method is only applicable to artificial forests; further verification is needed for natural forests. In the future, the density and quality of higher clouds could be increased. The validity and accuracy of the method were further verified. This paper provides a method for predicting the accumulation of typical Chinese plantations at the forest farm scale based on the “airborne LiDAR + satellite remote sensing” data-driven machine learning modeling, which has potential application value for the current carbon neutrality goal of the southern plantation forest emission reduction project.

Risk Management in Agile Al/Ml Projects: Identifying and Mitigating Data and Model Risks

Purpose: This study addresses the crucial challenge of managing risks associated with data and models in Agile Artificial Intelligence (AI) and Machine Learning (ML) projects. It aims to develop a systematic framework for effective risk control utilizing agile methodologies. Methodology: The research is grounded in an interpretivist approach and utilizes a deductive method. It constructs a comprehensive framework for identifying and mitigating risks, integrating risk management seamlessly into Agile processes for AI and ML development. Findings: The study introduces four technological themes critical for risk mitigation: dynamic distribution of resources, model robustness, risk integration, and quality assessment of information. These themes provide actionable strategies for reducing risks throughout the Agile AI/ML development lifecycle, ensuring that risk assessment and mitigation are integral to project planning and execution. Unique contribution to theory, practice, and policy: The study contributes to both theory and practice by offering a detailed, actionable framework for risk management in Agile AI/ML projects. It advocates for the adoption of adaptive technologies and tools, continuous stakeholder engagement, and adherence to ethical standards. Recommendations include validation of the framework through empirical research and ongoing longitudinal evaluations to adapt and refine risk management practices. This approach aims to enhance the reliability and efficiency of project outputs in dynamic environments, providing a significant foundation for policy development in technology project management.

Comparing machine learning and inverse modeling approaches for the source term estimation

Comparing machine learning and inverse modeling approaches for the source term estimation

Multimorbidity in middle-aged women and COVID-19: binary data clustering for unsupervised binning of rare multimorbidity features and predictive modeling

BackgroundMultimorbidity is typically associated with deficient health-related quality of life in mid-life, and the likelihood of developing multimorbidity in women is elevated. We address the issue of data sparsity in non-prevalent features by clustering the binary data of various rare medical conditions in a cohort of middle-aged women. This study aims to enhance understanding of how multimorbidity affects COVID-19 severity by clustering rare medical conditions and combining them with prevalent features for predictive modeling. The insights gained can guide the development of targeted interventions and improved management strategies for individuals with multiple health conditions.MethodsThe study focuses on a cohort of 4477 female patients, (aged 45-60) in Piedmont, Italy, and utilizes their multimorbidity data prior to the COVID-19 pandemic from their medical history from 2015 to 2019. The COVID-19 severity is determined by the hospitalization status of the patients from February to May 2020. Each patient profile in the dataset is depicted as a binary vector, where each feature denotes the presence or absence of a specific multimorbidity condition. By clustering the sparse medical data, newly engineered features are generated as a bin of features, and they are combined with the prevalent features for COVID-19 severity predictive modeling.ResultsFrom sparse data consisting of 174 input features, we have created a low-dimensional feature matrix of 17 features. Machine Learning algorithms are applied to the reduced sparsity-free data to predict the Covid-19 hospital admission outcome. The performance obtained for the corresponding models are as follows: Logistic Regression (accuracy 0.72, AUC 0.77, F1-score 0.69), Linear Discriminant Analysis (accuracy 0.7, AUC 0.77, F1-score 0.67), and Ada Boost (accuracy 0.7, AUC 0.77, F1-score 0.68).ConclusionMapping higher-dimensional data to a low-dimensional space can result in information loss, but reducing sparsity can be beneficial for Machine Learning modeling due to improved predictive ability. In this study, we addressed the issue of data sparsity in electronic health records and created a model that incorporates both prevalent and rare medical conditions, leading to more accurate and effective predictive modeling. The identification of complex associations between multimorbidity and the severity of COVID-19 highlights potential areas of focus for future research, including long COVID and intervention efforts.

Actual evapotranspiration and energy balance estimation from vineyards using micro-meteorological data and machine learning modeling

Actual evapotranspiration (ETa) can be commonly estimated using numerical models based on i) weather and plant-based parameters, ii) from remotely sensed data and energy balance algorithms, and lately, iii) through the development and implementation of machine learning (ML) modeling techniques. In this work, supervised ML models were developed from a vineyard located in Talca, Chile, (i) to estimate actual evapotranspiration (ETa) (Model 1; M1) using the micrometeorological approach [Eddy Covariance; EC; sensible (H), latent (LE), soil heat fluxes (G) and net radiation (Rn)] and data from an automatic meteorological station (AMS) in reference conditions as ground-truth (inputs); (ii) to estimate energy balance components (Model 2; M2) from AMS data (inputs) and EC energy balance data as targets; (iii) to estimate ETa from the EC’s measured ETa data as target and thermal time data (degree hours; DH) calculated from air temperature with a base of 5 °C increments from 5 – 45 °C as inputs (Model 3; M3) and iv) to estimate energy balance components (targets from EC) from the same inputs of Model 3 (Model 4; M4). Results showed that the developed ML models had high accuracy and performance with no signs of over or under-fitting with a high correlation (R) and slope (b) close to unity (M1; R=0.94; b=0.89; M2; R=0.97; b=0.93; M3; R=0.97; b=0.89–0.95; M4; R=0.98; b=0.97). Furthermore, models were directly deployed over another vineyard located 22 km West of the modeled vineyard at 60 m lower over the sea level with significant performances and R values (R = 0.64–0.87; b = 0.66–1.00 for M1 to M4, respectively). These models could be used for precision irrigation to increase water use efficiency and better control canopy vigor, balance fruit and vegetative components, and ultimately improve berry and wine quality traits.

Coupling of microbial-explicit model and machine learning improves the prediction and turnover process simulation of soil organic carbon

Modeling soil organic carbon (SOC) is helpful for understanding its distribution and turnover processes, which can guide the implementation of effective measures for carbon (C) sequestration and enhance land productivity. Process-based simulation with high interpretability and extrapolation, and machine learning modeling with high flexibility are two common methods for investigating SOC distribution and turnover. To take advantage of both methods, we developed a hybrid model by coupling of a two-carbon pool microbial model and machine learning for SOC modeling. Here, we assessed the SOC model's predictive, mapping, and interpretability capabilities for the SOC turnover process on Ningbo region. The results indicate that the microbial model with density-dependence (β ​= ​2) and microbial biomass carbon simulation performed better in modeling the parameters of the microbial-based C cycle, such as microbial carbon use efficiency (CUE), microbial mortality rate, and assimilation rate. By integrating this optimal microbial model and random forest (RF) model, the hybrid model improved the prediction accuracy of SOC, with an increased R2 from 0.74 to 0.84, residual prediction deviation increased from 1.97 to 2.50, and reduced the root-mean-square error from 4.65 to 3.67 ​g ​kg−1 compared to the conventional RF model. As a result, the predicted SOC distribution exhibited high spatial variation and provided abundant details. Microbial CUE and potential C input, represented by net primary productivity, emerged as the primary factors driving SOC distribution in Ningbo region. Projections of SOC under the CMIP6 SSP2-4.5 scenario revealed that regional C loss in high SOC areas was mainly caused by decreased microbial CUE and C input, induced by climate change. Our findings highlight the potential of combining the microbial-explicit model and machine learning to improve SOC prediction accuracy and understand SOC feedback in a changing climate.