Artificial Model Research Articles

Artificial Intelligence Recognition Model Using Liquid-Based Cytology Images to Discriminate Malignancy and Histological Types of Non-Small-Cell Lung Cancer

Introduction: Artificial intelligence image recognition has applications in clinical practice. The purpose of this study was to develop an automated image classification model for lung cancer cytology using a deep learning convolutional neural network (DCNN). Methods: Liquid-based cytology samples from 8 normal parenchymal (N), 22 adenocarcinoma (ADC), and 15 squamous cell carcinoma (SQCC) surgical specimens were prepared, and 45 Papanicolaou-stained slides were scanned using whole-slide imaging. The final dataset of 9,141 patches consisted of 2,737 N, 4,756 ADC, and 1,648 SQCC samples. Densenet-121 was used as the DCNN to classify N versus malignant (ADC+SQCC) and ADC versus SQCC images. AdamW optimizer and 5-fold cross-validation were used in the training. Results: For malignancy prediction, the sensitivity, specificity, and accuracy were 0.97, 0.85, and 0.94, respectively, in the patch-level classification, and 0.92, 0.88, and 0.91, respectively, in the case-level classification. For SQCC prediction, the sensitivity, specificity, and accuracy were 0.86, 0.91, and 0.90, respectively, in the patch-level classification and 0.73, 0.82, and 0.78, respectively, in the case-level classification. Conclusion: The DCNN model performed excellently in predicting malignancy and histological types of lung cancer. This model may be useful for predicting cytopathological diagnosis in clinical situations by reinforcing training.

Experimental Exploration of Cellulose Material for Battery Separators and Artificial Neural Network-Driven Predictive Modeling for Enhanced Thermal Safety in Electric Vehicles

Abstract Lithium-ion batteries (LIBs) are the most reliable energy storage devices nowadays because of their high energy density, long life cycle, and low self-discharge rate. But still, the safety concern is a significant problem in the area. When talking about LIB safety, thermal effects come first; this leads to thermal runaway, fires, and explosions. The critical component of LIB that has a great role in safety is the separator, which serves the purpose of preventing direct contact between the positive and negative electrodes while enabling the movement of lithium ions. This work aimed to find naturally available cellulose material for the LIB separator and to predict the performance of the material by artificial neural network (ANN) for better control of thermal problems that happen with traditional polymer separator materials. The cellulose derived from banana peels is isolated and characterized for its potential use as a separator material. The study conducts the four selected characterization approaches, scanning electronics microscopy (SEM) with three different resolutions to assess the morphology of the extracted cellulose, differential scanning calorimetry (DSC) to measure the heat flow with temperature change on the cellulose and the value obtained 231.22 J/g at a maximum temperature of 323.18 °C, thermogravimetric analysis (TGA) was used to examine the weight loss of the cellulose with respect to temperature variation, which results in a weight loss of 59.37% when the temperature reaches 235 °C, which is considered favorable, and a differential thermal analysis (DTA) was used to know the temperature difference in the banana peel cellulose (BPC), which results in a temperature of 330.23 °C. This morphological and thermal analysis technique for the BPC is used to determine the heat-related properties of the BPC, including phase transitions, thermal stability, and reaction. In addition, these results show BPC as an alternative material for separators in comparison to the existing polymer-based materials. Furthermore, these experimental results are used to train an ANN to predict the performance of BPC material using a binary classification. Because of the training process, 97.58% accuracy was achieved.

Prediction of Mechanical Properties of Lattice Structures: An Application of Artificial Neural Networks Algorithms.

The yield strength and Young's modulus of lattice structures are essential mechanical parameters that influence the utilization of materials in the aerospace and medical fields. Currently, accurately determining the Young's modulus and yield strength of lattice structures often requires conduction of a large number of experiments for prediction and validation purposes. To save time and effort to accurately predict the material yield strength and Young's modulus, based on the existing experimental data, finite element analysis is employed to expand the dataset. An artificial neural network algorithm is then used to establish a relationship model between the topology of the lattice structure and Young's modulus (the yield strength), which is analyzed and verified. The Gibson-Ashby model analysis indicates that different lattice structures can be classified into two main deformation forms. To obtain an artificial neural network model that can accurately predict different lattice structures and be deployed in the prediction of BCC-FCC lattice structures, the artificial network model is further optimized and validated. Concurrently, the topology of disparate lattice structures gives rise to a certain discrete form of their dominant deformation, which consequently affects the neural network prediction. In conclusion, the prediction of Young's modulus and yield strength of lattice structures using artificial neural networks is a feasible approach that can contribute to the development of lattice structures in the aerospace and medical fields.

Comparative computer simulation and empirical analysis of MIDAS and artificial neural network-UMIDAS models for short- and long-term US GDP forecasting

Comparative computer simulation and empirical analysis of MIDAS and artificial neural network-UMIDAS models for short- and long-term US GDP forecasting

Research status of the tumor stroma ratio in prognosis and treatment of pancreatic cancer

An increasing number of studies suggested that the tumor microenvironment exerts a substantial influence on the pathophysiology of pancreatic cancer. As a crucial component of the tumor microenvironment,the tumor stroma plays a pivotal role in the occurrence,development,and chemotherapy resistance of pancreatic cancer. By serving as a proxy for the interaction between tumor cells and the microenvironment,the tumor stroma ratio(TSR) has emerged as a focal point of investigation in recent years. At present,numerous studies show that a low TSR is a protective factor for the prognosis of resectable pancreatic cancer. Additionally, patients with a low TSR are more suitable for the gemcitabine and albumin-bound paclitaxel chemotherapy regimen. But these researches are not conclusive, and there is still a gap between guiding precision treatment. Further research and exploration are required. Integration of artificial intelligence deep learning models into traditional pathological and imaging assessments facilitates precise evaluation of the TSR. It can also enable stratification and precision treatment of pancreatic cancer patients based on this index.

Performance degradation evaluation of square steel tubes with local penetration damage

Steel tubular structures are susceptible to welding defects, cracking defects and environmental corrosion, which can negatively affect the stability and reliability of structure. This study tested the axial compression performance of twenty-six square steel tubular stubs with localized penetration damage simulated by artificial notch, considering the influence of notch length, orientation, and position. The test results showed that the failure mode of the specimens was dominated by the local buckling along the direction of the notch axis. The failure mode of the specimen with vertical notch showed notch opening, while those with horizontal notch showed notch closing, resulting in the secondary peak value in the load-displacement curve. The existence of local penetration damage reduced the ultimate strength and ductility of the tubular stubs. Compared to the ultimate strength of the specimens with middle notch, those with corner notch were weakened more, up to 27 %. Furthermore, the effects of notch dimension, position, orientation, and tube wall thickness on the axial bearing capacity of square steel tubular stubs are discussed by using a validated finite element model. Owing to the analysis, the ultimate strength of the stubs is reduced by increasing notch length, width and depth, whereas the reduction effect tends to be weakened as the notch orientation angle increases. The notch located in the corner is the most unfavorable for the axial compression performance of the square steel tubular stubs. A semi-empirical formula for evaluating the ultimate strength of square steel tubular stubs with local penetration damage is proposed by introducing the strength reduction factor. An artificial neural network-based ultimate strength prediction model is developed that can consider the effects of wide-range parameters, achieving high accuracy.

Changes in donor vessel physiology following coronary computed tomography angiography guided chronic total occlusion percutaneous coronary intervention: insights from computed tomography fractional flow reserve and artificial intelligence-guided ischemia model.

Donor vessel fractional flow reserve (FFR) usually increases following successful chronic total occlusion (CTO) percutaneous coronary intervention, as documented by pressure wires. In this case, donor vessel physiology changes were assessed using FFR derived from coronary computed tomography angiography (CCTA) and an artificial Intelligence-guided quantitative CCTA ischemia model in combination with pressure wire-based FFR.

Shortwave Infrared Imaging of a Quantum Dot-Based Magnetic Guidewire Toward Non-Fluoroscopic Peripheral Vascular Interventions.

Peripheral vascular interventions (PVIs) offer several benefits to patients with lower extremity arterial diseases, including reduced pain, simpler anesthesia, and shorter recovery time, compared to open surgery. However, to monitor the endovascular tools inside the body, PVIs are conducted under X-ray fluoroscopy, which poses serious long-term health risks to physicians and patients. Shortwave infrared (SWIR) imaging of quantum dots (QDs) has shown great potential in bioimaging due to the non-ionizing penetration of SWIR light through tissues. In this paper, a QD-based magnetic guidewire and its system is introduced that allows X-ray-free detection under SWIR imaging and precise steering via magnetic manipulation. The QD magnetic guidewire contains a flexible silicone tube encapsulating a QD polydimethylsiloxane (PDMS) composite, where HgCdSe/HgS/CdS/CdZnS/ZnS/SiO2 core/multi-shell QDs are dispersed in the PDMS matrix for SWIR imaging upon near-infrared excitation, as well as a permanent magnet for magnetic steering. The SWIR penetration of the QD magnetic guidewire is investigated within an artificial tissue model (1% Intralipid) and explore the potential for non-fluoroscopic PVIs within a vascular phantom model. The QD magnetic guidewire is biocompatible in its entirety, with excellent resistance to photobleaching and chemical alteration, which is a promising sign for its future clinical implementation.

Predictive modelling of cooling consumption in nursing homes using artificial neural networks: Implications for energy efficiency and thermal comfort

The growing need for cooling within the built environment, propelled by climate change and the expansion of nursing homes due to the increase in life expectancy, highlights the urgency of implementing energy-efficient strategies in buildings occupied by older populations. As of today, there remains a need for comprehensive research into the influence of indoor and outdoor conditions, building, operational, and occupant characteristics, on energy consumption specifically for nursing homes. This study develops a systemic artificial neural network-based model with a multi-layer perceptron architecture to assess HVAC energy implications during the cooling season for older populations. Using monitored data from eight nursing homes, the model includes cooling area, construction age, outdoor and indoor temperatures, and outdoor relative humidity as inputs, and cooling consumption as the output. Results show excellent predictive capability (R2=0.95), with mean error of −0.5 kWh, root mean squared error of 13.7 kWh, mean absolute error of 10.2 kWh, and relative error of 0.051. These outcomes are better compared to linear models (R2≈0.65) under the same data set. Adjusting operative temperatures adaptively can significantly enhance resident comfort and achieve up to 23.4 % energy savings, particularly in hotter, drier climates. These findings are of paramount importance for effective energy management in buildings.

GAIT-CKD (Gait Analysis using Artificial Intelligence for digital Therapeutics of patients with Chronic Kidney Disease): design and methods.

Digital therapeutics are emerging as treatments for diseases and disabilities. In chronic kidney disease (CKD), gait is a potential biomarker for health status and intervention effectiveness. This study aims to analyze gait characteristics in CKD patients, providing baseline data for digital therapeutics development. At baseline and after an 8-week intervention, we performed bioimpedance analysis measurements, the Timed Up and Go, Tinetti, and grip strength tests, and gait analysis in 217 healthy individuals and 276 patients with CKD. Demographic and clinical information was collected, including underlying diseases and medications, laboratory tests, and quality of life satisfaction surveys. Gait analysis was performed using skeleton data, which involved acquiring three-dimensional skeleton data of a walker using a single Kinect sensor. The performance of an artificial intelligence-based classification model in distinguishing between healthy individuals and those with CKD was then investigated. Simultaneously, inertia measurement unit analysis was conducted using measurements taken from the wrist and waist. Most subjects received a health intervention via an app, and their gait was assessed for improvements after an 8-week period. Incidents such as falls, fractures, hospitalizations, and deaths will be investigated in years 1 and 3. This study confirmed that the gaits of healthy individuals and CKD patients were different, and the effect of the 8-week app-based health intervention will be analyzed. The study will yield important baseline data for creating digital therapeutics for CKD patients' diet/exercise in the future.

Machine learning chain models for multi-response prediction of electrical equipment in substation subjected to earthquakes

Engineering structures often exhibit multiple potential vulnerable positions during strong earthquakes, such as porcelain insulators and connection flanges of electrical equipment in substations. This paper proposes two types of machine learning (ML) chain models, where multiple individual models are linked end to end, to predict multiple peak seismic responses of substation equipment at the same time using intensity measures (IMs). One is a simple chain where the next input is the previous output, while in another one, the next input is combined by IMs and the previous output. The training mechanisms of ML chain models are presented by means of bio-inspired multi-objective optimization techniques for the hyperparameters of multiple individual ML models. A case study on a 1100 kV transformer bushing is then conducted. Artificial neural network-gradient boosting regression and artificial neural network-kernel ridge regression chain models are respectively established for predicting the peak stresses at two most vulnerable positions. Both the evaluation indicators and shaking table tests show enough prediction accuracy of the two ML chain models. Such chain models are qualified for identifying the initial damage to equipment, which can be employed for assisting post-earthquake rapid judgement and relief.

An artificial intelligence ensemble model for paddy leaf disease diagnosis utilizing deep transfer learning

An artificial intelligence ensemble model for paddy leaf disease diagnosis utilizing deep transfer learning

Development, Calibration and Validation of Time Series Analysis and Artificial Neural Network Joint Model for Urban Noise Prediction

Noise in large urban areas, which is mainly generated by road traffic and by the human activities carried out nearby and inside the area under study, is a relevant problem. The continuous exposure to high noise levels, in fact, can lead to several problems, largely documented in the scientific literature. The analysis and forecasting of the noise level in a given area are, then, fundamental for control and prevention, especially when field measurements present peculiar trends and slopes, which can be modeled with a Time Series Analysis approach. In this paper, a hybrid model is presented for the analysis and the forecasting of noise time series in urban areas: this technique is based on the application of a deterministic decomposition model followed in cascade by a predictor of the forecasting errors based on an artificial neural network. Two variants of the hybrid model have been implemented and presented. The time series used to calibrate and validate the model is composed of sound pressure level measurements detected on a busy road near the commercial port of an Italian city. The proposed hybrid model has been calibrated on a part of the entire time series and validated on the remaining part. Residuals and error analysis, together with a detailed statistical description of the simulated noise levels and error metrics describe in detail the method’s performances and its limitations.

Artificial intelligence-based pulmonary embolism classification: Development and validation using real-world data.

This paper presents an artificial intelligence-based classification model for the detection of pulmonary embolism in computed tomography angiography. The proposed model, developed from public data and validated on a large dataset from a tertiary hospital, uses a two-dimensional approach that integrates temporal series to classify each slice of the examination and make predictions at both slice and examination levels. The training process consists of two stages: first using a convolutional neural network InceptionResNet V2 and then a recurrent neural network long short-term memory model. This approach achieved an accuracy of 93% at the slice level and 77% at the examination level. External validation using a hospital dataset resulted in a precision of 86% for positive pulmonary embolism cases and 69% for negative pulmonary embolism cases. Notably, the model excels in excluding pulmonary embolism, achieving a precision of 73% and a recall of 82%, emphasizing its clinical value in reducing unnecessary interventions. In addition, the diverse demographic distribution in the validation dataset strengthens the model's generalizability. Overall, this model offers promising potential for accurate detection and exclusion of pulmonary embolism, potentially streamlining diagnosis and improving patient outcomes.

Development and internal validation of an artificial intelligence-assisted bowel sounds auscultation system to predict early enteral nutrition-associated diarrhoea in acute pancreatitis: a prospective observational study.

Aims/Background An artificial intelligence-assisted prediction model for enteral nutrition-associated diarrhoea (ENAD) in acute pancreatitis (AP) was developed utilising data obtained from bowel sounds auscultation. This model underwent validation through a single-centre, prospective observational study. The primary objective of the model was to enhance clinical decision-making by providing a more precise assessment of ENAD risk. Methods The study enrolled patients with AP who underwent early enteral nutrition (EN). Real-time collection and analysis of bowel sounds were conducted using an artificial intelligence bowel sounds auscultation system. Univariate analysis, multicollinearity analysis, and logistic regression analysis were employed to identify risk factors associated with ENAD. The random forest algorithm was utilised to establish the prediction model, and partial dependence plots were generated to analyse the impact of risk factors on ENAD risk. Validation of the model was performed using the optimal model Bootstrap resampling method. Predictive performance was assessed using accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and an area under the receiver operating characteristic (ROC) curve. Results Among the 133 patients included in the study, the incidence of ENAD was 44.4%. Six risk factors were identified, and the model's accuracy was validated through Bootstrap iterations. The prediction accuracy of the model was 81.10%, with a sensitivity of 84.30% and a specificity of 77.80%. The positive predictive value was 82.60%, and the negative predictive value was 80.10%. The area under the ROC curve was 0.904 (95% confidence interval: 0.817-0.997). Conclusion The artificial intelligence bowel sounds auscultation system enhances the assessment of gastrointestinal function in AP patients undergoing EN and facilitates the construction of an ENAD predictive model. The model demonstrates good predictive efficacy, offering an objective basis for precise intervention timing in ENAD management.

The nonlinear rainfall–quick flow relationships in a humid mountainous area: Roles of soil thickness and forest type

The nonlinear behavior of hillslope quick flow (QF) in response to variations in soil thickness and forest types remains unclear due to uncertainties or unknown features in field studies, hindering accurate interpretation of their impact on QF. To this end, three artificial physical models with different mineral soil thickness (0 cm, shallow 50 cm, and thick 150 cm) beneath a 50 cm root zone for each forest (i.e., the broadleaf and coniferous forests) were constructed and intensively monitored. Results showed that the threshold-based QF was a function of the sum of rainfall amount and antecedent soil wetness, which was isolated by two breakpoints—the generation and rising thresholds—across all artificial physical models. These breakpoints indicated the storage thresholds for QF generation and the transition from slow to fast QF in humid forests. The broadleaf forest required more rainfall to reach the generation threshold due to higher litter interception capacity and soil water deficit by plant water uptake compared to the coniferous forest. Above the rising threshold phase, QF increased rapidly and linearly due to the establishment of flow path connectivity driven by soil water content, with the broadleaf forest showing a faster increase in QF than the coniferous forest. The dominant preferential/unsaturated flow in humid mountainous areas was also confirmed. Preferential flow paths were well-connected in the root zone and extended to the shallow mineral soil, resulting in an insignificant difference in mean QF between forest types. The thick mineral soil disrupted preferential flow path connectivity, leading to a greater amount of water being laterally diverted and retained within the soil. Consequently, a sharp decrease in QF was observed in hillslope with the root zone capping thick mineral soil for both forests. Our results highlighted the application of threshold-based theory and controlled experiments in enhancing the comprehension of the fundamental physical mechanism governing runoff processes in humid forests.

Machine learning based methods for ratemaking health care insurance

In insurance, proposing an accurate premium that is adjusted to the insured risk profile allows companies to better manage their portfolios and to be more competitive. Machine learning methods have recently been adopted for various improvements in insurance ratemaking, especially in the automobile industry. These models are specifically used to mine potential data information and to build a predictive model for a variable of interest using explanatory variables. In this paper, we aim to provide a pricing method for ratemaking individual healthcare insurance contracts using machine learning algorithms that are applied to a Tunisian healthcare insurance portfolio. We start with a simple Classification and Regression Tree, and we work toward more advanced methods that are Random Forest, Extreme gradient boosting, Support Vector Regression, and Artificial Neural network regression model. The predictive performance of these non-parametric methods is compared with the standard generalized linear model. Our results showed the applicability of machine learning in the healthcare insurance market and that the XGBoost algorithm outperforms the predictive capacity of the classical generalized linear model.

Artificial Intelligence-Based Models for Prediction of Mortality in ICU Patients: A Scoping Review.

Healthcare professionals may be able to anticipate more accurately a patient's timing of death and assess their possibility of recovery by implementing a real-time clinical decision support system. Using such a tool, the healthcare system can better understand a patient's condition and make more informed judgements about distributing limited resources. This scoping review aimed to analyze various death prediction AI (Artificial Intelligence) algorithms that have been used in ICU (Intensive Care Unit) patient populations. The search strategy of this study involved keyword combinations of outcome and patient setting such as mortality, survival, ICU, terminal care. These terms were used to perform database searches in MEDLINE, Embase, and PubMed up to July 2022. The variables, characteristics, and performance of the identified predictive models were summarized. The accuracy of the models was compared using their Area Under the Curve (AUC) values. Databases search yielded an initial pool of 8271 articles. A two-step screening process was then applied: first, titles and abstracts were reviewed for relevance, reducing the pool to 429 articles. Next, a full-text review was conducted, further narrowing down the selection to 400 key studies. Out of 400 studies on different tools or models for prediction of mortality in ICUs, 16 papers focused on AI-based models which were ultimately included in this study that have deployed different AI-based and machine learning models to make a prediction about negative patient outcome. The accuracy and performance of the different models varied depending on the patient populations and medical conditions. It was found that AI models compared with traditional tools like SAP3 or APACHE IV score were more accurate in death prediction, with some models achieving an AUC of up to 92.9%. The overall mortality rate ranged from 5% to more than 60% in different studies. We found that AI-based models exhibit varying performance across different patient populations. To enhance the accuracy of mortality prediction, we recommend customizing models for specific patient groups and medical contexts. By doing so, healthcare professionals may more effectively assess mortality risk and tailor treatments accordingly. Additionally, incorporating additional variables-such as genetic information-into new models can further improve their accuracy.

How reliable is the artificial intelligence product large language model ChatGPT in orthodontics?

To evaluate the reliability of information produced by the artificial intelligence-based program ChatGPT in terms of accuracy and relevance, as assessed by orthodontists, dental students, and individuals seeking orthodontic treatment. Frequently asked and curious questions in four basic areas related to orthodontics were prepared and asked in ChatGPT (Version 4.0), and answers were evaluated by three different groups (senior dental students, individuals seeking orthodontic treatment, orthodontists). Questions asked in these basic areas of orthodontics were about: clear aligners (CA), lingual orthodontics (LO), esthetic braces (EB), and temporomandibular disorders (TMD). The answers were evaluated by the Global Quality Scale (GQS) and Quality Criteria for Consumer Health Information (DISCERN) scale. The total mean DISCERN score for answers on CA for students was 51.7 ± 9.38, for patients was 57.2 ± 10.73 and, for orthodontists was 47.4 ± 4.78 (P = .001). Comparison of GQS scores for LO among groups: students (3.53 ± 0.78), patients (4.40 ± 0.72), and orthodontists (3.63 ± 0.72) (P < .001). Intergroup comparison of ChatGPT evaluations about TMD was examined in terms of the DISCERN scale, with the highest value given in the patients group (57.83 ± 11.47) and lowest value in the orthodontist group (45.90 ± 11.84). When information quality evaluation about EB was examined, it GQS scores were >3 in all three groups (students: 3.50 ± 0.78; patients: 4.17 ± 0.87; orthodontists: 3.50 ± 0.82). ChatGPT has significant potential in terms of usability for patient information and education in the field of orthodontics if it is developed and necessary updates are made.

Views of health sciences undergraduates on ChatGPT, an artificial intelligence-powered language model: A qualitative study

ABSTRACT Artificial Intelligence (AI) has significantly impacted the field of education, particularly in health sciences, where tools such as ChatGPT are increasingly utilised. ChatGPT, powered by AI, presents both opportunities and challenges that warrant investigation. This qualitative study explored the perceptions, experiences, and expectations of undergraduate students in health sciences regarding ChatGPT, an AI-powered language model. Through in-depth interviews, data were collected and analysed thematically, focusing on three main themes: perceptions of ChatGPT, student experiences with the tool, and their future expectations. The findings reveal that while students see ChatGPT as a revolutionary tool and a valuable source of information, they also acknowledge concerns about its potential for providing incorrect information. The study emphasises the importance of critical thinking when using AI models in education and provides guidance for responsible usage.