An integrated framework combining Finite Element Analysis (FEA) and Artificial Neural Networks (ANN) is presented to enhance the prediction and design of bioprinted scaffolds. By leveraging the strengths of data-driven learning and physics-based simulations, the hybrid approach (ANN + FEA) achieves superior predictive accuracy and generalization compared to standalone approaches. Validation against experimental results demonstrates that a single ANN model yields a relative error of 5.17% when predicting scaffold Young’s modulus. Incorporating FEA simulation based on ANN-predicted geometry and material properties reduces the relative error to 4.72%, representing an 8.6% improvement. The framework also enables the accurate simulation for unseen combinations of printing parameters located far from the experimental data manifold, reducing prediction errors from 14.2% (ANN-only) to 5.7% (hybrid). By integrating predictive modeling, simulation, and data augmentation, this approach offers an efficient pathway for optimizing scaffold designs and accelerating the development of biomaterials with tailored mechanical performance.

ABSTRACTA propensity weighted (PSW) methodology was recently proposed for assessing the treatment effect conditional to the probability of non‐specific response to a treatment (prob‐NSRT). Prob‐NSRT was estimated using an artificial neural network (ANN) model applied to pre‐randomization and study endpoint observations in a placebo arm of a placebo‐controlled clinical trial. Placebo data were initially used to estimate prob‐NSRT, then the ANN model was applied to the data of each individual in each treatment arm (placebo + active) for estimating the individual prob‐NSRT, and finally all data in the trial enriched by the prob‐NSRT values were used to assess the treatment effect. One of the major limitations of this methodology was that the ANN model was developed and applied to analyze data in the same dataset. To overcome this limitation, a new artificial intelligence driven nonlinear mixed effect modeling approach (AI‐NLME) is proposed. This approach involves the development of the ANN model using a dataset that is independent from the dataset used to estimate the treatment effect. A case study is presented using data from a randomized, placebo‐controlled trial in major depressive disorders. The AI‐NLME approach provided an effective tool for controlling the confounding effect of treatment non‐specific response, for increasing signal detection, for decreasing heterogeneity in the response, for increasing the effect size, for better assessing the responder rate, and for providing a reliable estimate of the “true” treatment effect. These findings provide convergent evidence on the potential role of AI‐NLME to become the reference approach for analyzing placebo‐controlled clinical trials.

Predicting diameter and tensile strength of electrospun fibers for biomedicine: A comparison of Box-Behnken design, traditional machine learning and deep learning.

ABSTRACTThis research paper involved modeling the water absorption behavior of polystyrene (PS) composites with sawdust and kenaf fiber (KF) reinforcement using the Artificial Neural Network (ANN) method. The composites were made by manual mixing combined with the hand lay‐up process at room temperature (25°C ± 2°C) and cured in an open mold over 7 days at ambient temperature. The water absorption measurements were done according to the ASTM D1037‐99. The findings were that the water uptake was enhanced by filler content as well as immersion duration in the sawdust composite and KF. The ANN model also had good accuracy; the coefficients of determination (R2) in all the ANN models were more than 0.98 in all the training, validating, and test sets of both types of materials. Also, values of root mean square error (RMSE) were low (less than 1 wt%), indicating that this model was very accurate in forecasting the behavior of water absorption. Parity plots indicated that there was a good balance of the performance of the predictions, which captured the low and high values of absorption. Moreover, the p value was lower than 0.05, which showed ANOVA results are statistically significant.

A promising material used in radiation synovectomy of small joints is hydroxyapatite, labeled with 177Lu. During the design and production of radiopharmaceuticals, the condition of the radiolabeling process directly influences the radiochemical yield and consequently the quality of the final product so this process necessitates precise optimization. In this investigation, a central composite design based on response surface methodology and artificial neural networks modeling coupled with genetic algorithm technique is applied to build predictive models and explore key parameters' effect in hydroxyapatite's radiolabeling process with 177Lu radionuclide. The variables that directly affected the labeling reaction were the initial 177Lu radioactivity, pH, radiolabeling reaction time, and temperature. Based on the validation data set, the statistical values demonstrate that the artificial neural networks model performs better than the response surface methodology model. The artificial neural networks model has a small mean squared error (9.08 artificial neural networks < 12.36 response surface methodology) and a high coefficient of determination (R2: 0.99 artificial neural networks > 0.93 response surface methodology). The optimum conditions to achieve maximum radiochemical yield based on response surface methodology using artificial neural networks modeling coupled with genetic algorithm were at the initial radioactivity of 177Lu radionuclide = 0.082 Gigabecquerel (GBq), pH = 6.75, time= 22 (min), and temperature = 37.8 (oC). The ability to generate more data with fewer experiments for optimization and improved production is a pertinent advantage of multivariate optimization methods over traditional methods in radiation-related activities. The central composite design and artificial neural network- genetic algorithm optimization approaches are successfully utilized to create prediction models and investigate the impact of critical variables in the radiolabeling of hydroxyapatite with 177Lu radionuclide.

Optimization of ultrasound-assisted extraction of polysaccharides from Akebia Fruit using an artificial neural network model: Characteristics and antioxidant activity.

Abstract Adhesive joints are widely preferred as fastening elements due to their versatility and load distribution capabilities. However, their performance can be significantly affected under varying environmental conditions. Elevated service temperatures, in particular, tend to adversely impact the mechanical strength of bonded joints. In this study, short Kevlar fibers (SKF) were incorporated alongside carbon fibers (CF) to enhance the performance of single-lap joints (SLJs) exposed to high temperatures. Additionally, the influence of vibration applied to the adhesive joints was investigated. The shear strength of the joints was modeled using Response Surface Methodology (RSM) and Artificial Neural Networks (ANN). The results demonstrated that the addition of SKF improved the strength of SLJs under high-temperature conditions, and the application of ultrasonic vibration further amplified this enhancement. The R2 values of the RSM and ANN models are 0.994 and 0.97, respectively. The highest shear and bonding strength, along with a cohesive failure mode, were observed in joints containing 1.5% CF and 1% SKF. Dynamic Mechanical Analysis (DMA) confirmed that the inclusion of SKF positively influenced the storage modulus, loss modulus, and glass transition temperature of the adhesive.

In order to enhance the efficiency of water usage intended for drinking or utility purposes, the determination of water quality has assumed greater significance. In this study, water samples were taken from a region of the Kızılırmak River over the course of 1 year and analyzed. A probabilistic water quality assessment was then made, taking uncertainties into account, using artificial neural networks (ANN) and Monte Carlo simulation (MCS) methods. The Weighted Arithmetic Water Quality Index method was employed. Analyses of dissolved oxygen (DO), pH, temperature, turbidity, chloride, and sulfate were conducted on samples collected between the years 2023 and 2024. The independent variables (temperature, sulfate, and chloride) were generated, and the dependent variables (pH, turbidity, and DO) were estimated with ANN. The R value and the root mean square error (RMSE) of ANN models are used to evaluate the effectiveness of the model by assessing both the accuracy and the margin of error. The findings of the sensitivity analysis demonstrated that the parameters of DO, pH, and turbidity exerted a significant influence on the quality of water in the Kızılırmak River. A methodology was implemented in which three distinct ANN prediction models function collectively. The study yielded reliability levels corresponding to different WQI categories, namely 10% for "excellent" quality, 29% for "good" quality, 59% for "poor" quality, and 86% for "very poor" quality. This indicates that the ANN and MCS models are effective estimation tools for determining the water quality of the Kızılırmak River.

ABSTRACTThe crashworthiness optimization of closed‐cell aluminum foam‐filled sandwiched crash boxes is a critical aspect of vehicle occupant safety, aimed at enhancing the energy absorption capability of these structures during collisions. This study is focused on enhancing the crash box energy absorption capacity by using closed‐cell‐sandwiched aluminum foam characterized by lightweight and high‐energy absorption properties. The design of the experiment (DOE) is used to determine the minimum number of runs by considering cell size, void fraction, and density as input parameters and energy absorption as output parameters. The finite element analysis (FEA) is conducted using ABAQUS with tetrahedral element type under impact loading conditions by considering good mesh quality, well‐defined boundary conditions, and material models. An artificial neural network (ANN) integrated with a genetic algorithm (GA) is used to predict and optimize the maximum possible energy absorption capacity. After analysis, the maximum energy absorption of 255 J is identified from 27 runs, achieved with a combination of cell size, porosity, and density of (10, 15, and 2.6). To optimize energy absorption and determine optimal parameters, results from Abaqus are input into the ANN model. The ANN generates a fitting function with a high R value (0.989) and minimum error (1.34). The fitness function is then exported to the GA optimization tool, refining it to achieve an optimized energy absorption of 256.69 J. The optimal parameters identified through this process are cell size 10, porosity 0.162, and density 2.6. From the results obtained, we can conclude that the use of integrated computational methodologies can enhance crashworthiness optimization of complex foam geometries to provide a high‐performance energy‐absorbing crash box.

Time-of-flight based one-dimensional position estimation of radioactive sources using artificial neural network model

Fabrication and characterization of biological biosensors in sports injury treatment: High sensitivity of silver oxide using artificial neural network modeling

Optimization of ultrasonic extraction processes, structural characteristics and potential antipyretic mechanism of a glucan from Tetrastigma hemsleyanum Diels.

Soil contamination remains a critical environmental concern, necessitating efficient techniques for site characterization and remediation. The radon-deficit technique (RDT) offers a non-invasive approach to identifying organic contamination, relying on the behavior of radon-222 (222Rn) as a tracer. However, RDT results are influenced by environmental variables such as soil moisture, temperature, and atmospheric pressure, potentially leading to uncertainties. This study evaluates the application of machine learning (ML) models-including linear regression (LR), random forest (RF), artificial neural network (ANN), and gradient boosting machine (GBM)-to predict 222Rn activity in soil gas based on environmental parameters. A year-long dataset of continuous measurements was collected from an uncontaminated granite-based site in Madrid, encompassing variables such as soil moisture, ambient and soil temperatures, and atmospheric conditions. ANN and RF models exhibited superior performance in predicting 222Rn variability, identifying soil moisture and ambient temperature as the most influential predictors. The findings demonstrate that ML can significantly enhance the reliability of RDT by accounting for environmental variability, enabling more accurate identification of contamination hotspots. While the application of these models requires substantial datasets, they offer a promising tool for improving the efficacy of contamination screening and long-term remediation monitoring. Further studies are recommended to explore ML's predictive capacity in contaminated sites and expand the approach to diverse geological contexts.

Assessing the carbon footprint and environmental impact of bone hydrochar for arsenic remediation: A life cycle approach to hydrothermal waste recycling.

Predicting disinfection by-products (DBPs) in supply water within a real water distribution network using an artificial neural network.

ABSTRACTThe maritime industry faces increasing pressure to enhance vessel efficiency and reduce environmental impact, largely through structural weight reduction. This study presents the design and optimization of a lightweight Enclosed Mast for naval ships, using both steel and composite materials. The aim was to improve aerodynamic efficiency and reduce stress concentrations. Finite element analysis (FEA) and computational fluid dynamics (CFD) were employed to evaluate structural and aerodynamic performance. An artificial neural network (ANN), coupled with a genetic algorithm, identified optimal designs, while Monte Carlo simulations addressed uncertainties, and sensitivity analysis determined the most influential input parameters. The optimized mast achieved a 50% weight reduction and a 36% decrease in body thickness, along with improved stress distribution. Aerodynamically, it reduced low‐velocity zones and drag by stabilizing the wake and smoothing flow. The ANN model achieved over 97% accuracy and a low error of 0.05, validating the approach for future marine structure optimization.

The thermochemical conversion of biomass into biochar is a key process in sustainable resource management. However, accurately predicting biochar yield remains challenging due to the diverse nature of biomass and the complexities of pyrolysis. This review examines the application of artificial neural networks (ANNs) in biochar yield prediction, highlighting their potential to improve accuracy. However, the “black-box” nature of ANNs, along with the high dimensionality of biomass data and the complexity of industrial systems, limits their interpretability and broader applicability. Emerging hybrid ANN models, combining data-driven and mechanistic approaches, offer a solution by enhancing predictive performance and model transparency. Future research should focus on developing integrated datasets covering diverse biomass types and pyrolysis conditions, as well as incorporating real-time data and feedback mechanisms to improve scalability and effectiveness in industrial biochar production, leading to enhanced economic and environmental outcomes.

Retained surgical items (RSIs) remain a persistent challenge in patient safety, with retained surgical sponges (RSS) being the most common. Traditional RSI prevention methods, including manual counting, radiofrequency identification (RFID), and radiography, have demonstrated limitations, leading to persistent surgical errors. Artificial intelligence (AI), particularly deep learning models, has emerged as a promising solution for improving RSS detection and reducing human error in the operating room. This review examines the application of AI in RSI prevention, focusing on deep learning techniques such as Convolutional Neural Networks (CNNs) and Artificial Neural Networks (ANNs). CNN models analyze visual data such as images and videos, while ANN models recognize complex data patterns. Studies have demonstrated that CNN-based models significantly enhance RSS detection in x-rays and laparoscopic video feeds, often outperforming human observers. Object detection models, such as YOLO (You Only Look Once), have shown promise in real-time RSS tracking, making them particularly valuable in complex surgical environments. In addition, ANN-based computer-aided detection (CAD) systems, when combined with radiopaque markers, have improved accuracy in identifying retained sponges. Despite these advancements, several challenges remain, including data set limitations, false positives, and difficulties distinguishing gauze from surrounding tissue. Further research is needed to refine these models, expand their applications beyond RSS, and integrate them effectively into surgical workflows. The adoption of AI-based detection systems has the potential to enhance patient safety, reduce health care costs, and prevent surgical never events, marking a crucial step toward reducing RSIs in modern surgical practice.

Biomacromolecular engineering of redox-active chitosan-polypyrrole-clay hybrid materials for machine learning-assisted desulfurization and supercapacitor application.

This study indicates the utilization of two Machine learning models that is Artificial Neural Network (ANN) and Random Forest(RF), for the prediction of equivalent noise levels (Leq), across four locations covering commercial and silent zones in Yanam, Puducherry. The total Passenger Car Unit (PCU) data, observed Leq of every location was used effectively for the training and testing of the ANN model, resulting in aCoefficient of Determination (R²) value of 0.912.Additionally, a Random Forest regression model was applied for performance comparison, with a result of an R² of 0.925.The noise maps on weekdays and Sunday revealed the highest and the lowest noise levels at the locations during the 12 hour time interval. The study indicates that the noise levels are proportional to traffic volume, highlighting the need to implement effective strategies for reducing noise pollution in the area.