Neural Network Based Machine Learning Research Articles

RoseNNa: A performant, portable library for neural network inference with application to computational fluid dynamics

The rise of neural network-based machine learning ushered in high-level libraries, including TensorFlow and PyTorch, to support their functionality. Computational fluid dynamics (CFD) researchers have benefited from this trend and produced powerful neural networks that promise shorter simulation times. For example, multilayer perceptrons (MLPs) and Long Short Term Memory (LSTM) recurrent-based (RNN) architectures can represent sub-grid physical effects, like turbulence. Implementing neural networks in CFD solvers is challenging because the programming languages used for machine learning and CFD are mostly non-overlapping, We present the roseNNa library, which bridges the gap between neural network inference and CFD. RoseNNa is a non-invasive, lightweight (1000 lines), and performant tool for neural network inference, with focus on the smaller networks used to augment PDE solvers, like those of CFD, which are typically written in C/C++ or Fortran. RoseNNa accomplishes this by automatically converting trained models from typical neural network training packages into a high-performance Fortran library with C and Fortran APIs. This reduces the effort needed to access trained neural networks and maintains performance in the PDE solvers that CFD researchers build and rely upon. Results show that RoseNNa reliably outperforms PyTorch (Python) and libtorch (C++) on MLPs and LSTM RNNs with less than 100 hidden layers and 100 neurons per layer, even after removing the overhead cost of API calls. Speedups range from a factor of about 10 and 2 faster than these established libraries for the smaller and larger ends of the neural network size ranges tested. Program summaryProgram Title: RoseNNaCPC Library link to program files:https://doi.org/10.17632/srbrfx8k74.1Developer's repository link:https://github.com/comp-physics/roseNNaLicensing provisions: MITProgramming language: Fortran90 and PythonNature of problem: Neural network applications in computational fluid dynamics exhibit high promise. However, they are often not deployed on HPC systems due to limited deep-learning support in Fortran or C/C++. Previous attempts to solve this issue rely on significant manual intervention. This leaves researchers with no practical solution for conveniently porting their trained models for use in performant HPC codebases.Solution method: RoseNNa converts trained neural networks from popular machine learning frameworks to Fortran/C code for inference, requiring minimal manual intervention. Metaprogramming enables automatic code creation, and ONNX ensures flexible machine learning library support.

Predicting Major Bleeding in Patients with Venous Thromboembolism on Extended Anticoagulation Therapy Using Follow-up Data and Long Short-Term Memory Based Recurrent Neural Network

Background: It is often challenging for physicians to decide on the duration of anticoagulation treatment for patients with venous thromboembolism (VTE), as they need to weigh the risk of recurrent thrombosis and bleeding at the same time. Several clinical models have been developed to help physicians identify patients with high risk of bleeding. However, these tools use only the baseline clinical information and are not able to incorporate the clinical conditions and events that occur over time that may influence the risk of bleeding. Therefore, capturing the patterns and relationships in the continuously changing time series of clinical data could be a more powerful approach to developing predictor models than just relying on the baseline clinical information. Nonetheless, creating predictive models from follow-up clinical information is challenging given that time series clinical data are non-uniform, high-dimensional, multivariate observations. Here, we present the first attempt at creating a model that uses the time series follow-up information to predict major bleeding over time using a Recurrent Neural Network (RNN) with Long Short-Term Memory (LSTM) cells. Methods: 2542 patients diagnosed with VTE were enrolled in a prospective cohort study over 8 years. In addition to recording their clinical information at baseline, 6-month follow-up interviews were conducted using a standard script to monitor bleeding status and record clinical information. Major bleeding was defined by the International Society on Thrombosis and Haemostasis, with suspected bleeding events classified by an independent adjudication committee. Overall, 118 patients had major bleeding - a 4.6% incidence rate. The median and mode of the clinical variables were used to impute missing numerical and categorical values, respectively, and for patients who had no follow-up information, for whom bleeding occurred before the first follow-up, an artificial follow-up data point was generated from their corresponding baseline data. Thereafter, the data was divided into two stratified sets: 70% for training, and 30% for testing. Five supervised neural network-based machine learning models with different architectures were trained on the baseline dataset, or the follow-up dataset, or both to predict major bleeding. After training, these machine learning models were tested on the testing set and compared to the conventional clinical models, modified to make them compatible with the available predictor variables in our dataset, including the CHAP, the HAS-BLED, the VTE-BLEED, the RIETE, the ACCP, and the OBRI, which only use the baseline information. Results: Overall, the models that used the follow-up information had a higher area under the Receiver Operating Curve (AUROC) or c-statistic compared to the other models that only relied on the baseline dataset. In particular, the LSTM RNN model was able to achieve AUROC of 81.3% that is more than 10% higher compared to the best performing clinical model. We discovered that the LSTM RNN model mostly relied on features such as number of concomitant medications, years since baseline visit, use of specific antibiotics or antiplatelet agents, and presence of new hypertension to predict bleeding from the follow-up dataset. Furthermore, half of the bleeding events occurred within the first year after patients' baseline visits - a trend reflected in the predictions made by LSTM RNN model. Finally, the models that used both the baseline and the follow-up datasets showed different results depending on their architectures; that is, the simpler ensemble model achieved AUROC of 82.5% while the more complex model had AUROC of 70.8% due to overfitting. Conclusion: We have shown that using time series follow-up data can improve bleeding risk prediction in patients with VTE who are on extended anticoagulant therapy compared to just using the baseline data, and clinicians might benefit from using such an approach. Furthermore, our results indicate that LSTM RNN is a suitable architecture to model routine clinical follow-up data. Finally, we believe using time series data could improve the performance of the other clinical models that are currently based on one-time baseline measurements.

A graph neural network-based machine learning model for sentiment polarity and behavior identification of COVID patients

A graph neural network-based machine learning model for sentiment polarity and behavior identification of COVID patients

An artificial neural network-based machine learning approach to correct coarse-mesh-induced error in computational fluid dynamics modeling of cell culture bioreactor

Computational Fluid Dynamics (CFD) is a valuable tool for studying fluid environments within cell culture bioreactors and optimizing processing parameters, but it can be computationally expensive. This study developed an artificial neural network (ANN)-based machine learning model to predict and correct the coarse-mesh-induced errors in CFD modeling of a spinner flask bioreactor. A baseline ANN model was trained to predict the velocity error function between the coarse and optimized reference mesh results at one rotational speed (90 rpm), demonstrating that the ANN-based approach could correct the coarse-mesh velocity with RMSE values of nodal velocities improved by an average of ∼20% at different rotational speeds. The effect of ANN structure, input data normalization, and training dataset combinations on prediction performance was evaluated. More neurons and hidden layers generated better results but required more computational time for training. The model’s generalization capabilities were further evaluated in case studies of correcting velocity and Kolmogorov length at different fluid viscosity and bioreactor impeller geometry conditions. Results suggested that the ANN model had better generalization in correcting Kolmogorov length than velocity. This research provides insights into using a machine learning approach to enhance CFD modeling in bioreactor applications, contributing to advancing tissue engineering processes.

Multimodal Three-Dimensional Printing for Micro-Modulation of Scaffold Stiffness Through Machine Learning.

The ability to precisely control a scaffold's microstructure and geometry with light-based three-dimensional (3D) printing has been widely demonstrated. However, the modulation of scaffold's mechanical properties through prescribed printing parameters is still underexplored. This study demonstrates a novel 3D-printing workflow to create a complex, elastomeric scaffold with precision-engineered stiffness control by utilizing machine learning. Various printing parameters, including the exposure time, light intensity, printing infill, laser pump current, and printing speed were modulated to print poly (glycerol sebacate) acrylate (PGSA) scaffolds with mechanical properties ranging from 49.3 ± 3.3 kPa to 2.8 ± 0.3 MPa. This enables flexibility in spatial stiffness modulation in addition to high-resolution scaffold fabrication. Then, a neural network-based machine learning model was developed and validated to optimize printing parameters to yield scaffolds with user-defined stiffness modulation for two different vat photopolymerization methods: a digital light processing (DLP)-based 3D printer was utilized to rapidly fabricate stiffness-modulated scaffolds with features on the hundreds of micron scale and a two-photon polymerization (2PP) 3D printer was utilized to print fine structures on the submicron scale. A novel 3D-printing workflow was designed to utilize both DLP-based and 2PP 3D printers to create multiscale scaffolds with precision-tuned stiffness control over both gross and fine geometric features. The described workflow can be used to fabricate scaffolds for a variety of tissue engineering applications, specifically for interfacial tissue engineering for which adjacent tissues possess heterogeneous mechanical properties (e.g., muscle-tendon).

Starting at Go: Protein structure prediction succumbs to machine learning

This year’s Albert Lasker Basic Medical Research Award recognizes the invention of AlphaFold, a revolutionary advance in the history of protein research which for the first time offers the practical ability to accurately predict the three-dimensional arrangement of amino acids in the vast majority of proteins on a genomic scale on the basis of sequence alone [J. Jumper et al., Nature 596, 583–589 (2021) and K. Tunyasuvunakool et al., Nature 596, 590–596 (2021)]. This extraordinary achievement by Demis Hassabis and John Jumper and their coworkers at Google’s DeepMind and other collaborators was built on decades of experimental protein structure determination (structural biology) as well as the gradual development of multiple strategies incorporating biologically inspired statistical approaches. But when Jumper and Hassabis added a brew of innovative neural network-based machine learning approaches to the mix, the results were explosive. Realizing the half-century-old dream of predicting protein structure has already accelerated the pace and creativity of many areas of Chemistry, Biology, and Medicine.

Nanoscale soil-water retention mechanism of unsaturated clay via MD and machine learning

In this article, we investigate the nanoscale soil-water retention mechanism of unsaturated clay through molecular dynamics and machine learning. Pyrophyllite was chosen due to its stable structure and as the precursor of other 2:1 clay minerals. A series of molecular dynamics simulations of clay at low degrees of saturation were conducted. Soil water was represented by a point cloud through the center-of-mass method. Water-air interface area was measured numerically by the alpha-shape method. The soil-water retention mechanism at the nanoscale was analyzed by distinguishing adsorptive pressure and capillary pressure at different mass water contents and considering the apparent capillary interface area (i.e., water-air interface area per unit water volume). The water number density profile was used to quantify the adsorption effect. A neural-network based machine learning technique was utilized to construct functional relationships among matric suction, the mass water content, and the apparent water-air interface area. Our numerical results have demonstrated from a nanoscale perspective that the adsorption effect is dominated by the van der Waals force and hydroxyl hydration between the clay surface and water. As the mass water content increases, the adsorption pressure decreases, and capillarity plays a prominent role in the soil-water retention mechanism at the nanoscale.

Diagnostic Accuracy of Convolutional Neural Network Based Machine Learning Algorithms in Endoscopic Severity Prediction of Ulcerative Colitis: A Systematic Review & Meta-Analysis.

Endoscopic assessment of ulcerative colitis (UC) can be performed by Mayo Endoscopic Score (MES), or the Ulcerative Colitis Endoscopic Index of Severity (UCEIS). In this meta-analysis, we assessed the pooled diagnostic accuracy parameters of deep machine learning, by means of convolutional neural network (CNN) algorithms, in predicting UC severity on endoscopic images. Databases including Medline, Scopus, and Embase, were searched in June 2022. Outcomes of interest were the pooled accuracy, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV). Standard meta-analysis methods were employed using the random-effects model, and heterogeneity was assessed using the I2statistics. Twelve studies were included in the final analysis. The pooled diagnostic parameters of CNN-based machine learning algorithms in endoscopic severity assessment of UC were: Accuracy 91.5% (95% confidence interval (CI) [88.3-93.8],I2=84%), sensitivity 82.8% ([78.3-86.5],I2=89%), specificity 92.4% ([89.4-94.6],I2=84%), PPV 86.6% ([82.3-90],I2=89%) and NPV 88.6% ([85.7-91],I2=78%). Subgroup analysis revealed significantly better sensitivity and PPV with UCEIS scoring system as compared to MES (93.6% [87.5-96.8],I2=77% vs 82% [75.6-87],I2=89%, p=0.003; and 93.6% [88.7-96.4],I2=68% vs 83.6% [76.8-88.8],I2=77%; p=0.007, respectively). CNN-based machine learning algorithms demonstrated excellent pooled diagnostic accuracy parameters in the endoscopic severity assessment of UC. Utilizing UCEIS scores in CNN training might offer better results than MES. Further studies are warranted to establish these findings in real-life.

Parametric analysis and prediction of energy consumption of electric vehicles using machine learning

Emission regulations for all automobiles have been introduced to reduce global warming caused by vehicles. Hybrid electric vehicles (HEVs) are being developed to address consumer demand for environmentally friendly automobiles with more power and better fuel efficiency. HEVs are propelled by a combination of an internal combustion engine (ICE) and one or more electric motors that draw power from a secondary battery, which is commonly a lithium-based battery. The fuel economy of such a hybrid drivetrain system can be enhanced above that of traditional ICE automobiles. Although the global world is now focussing on electric vehicles (EVs) over HEVs due to environmental pollution. In this study, a 1-dimensional model was developed for an electric vehicle (EV), and a parametric analysis was made for the eight different cycles using GT-Suite software. The parameters included motor power, state of charge of the battery, vehicle speed, distance travelled, and energy consumption. In light of the parametric analysis obtained using GT-Suite software, this paper also predicts the energy consumption of EV batteries using a neural network-based machine learning (ML) method. After selecting input parameters through a correlation coefficient index (CI) process, the proposed neural network-based prediction model has achieved 89% accuracy in predicting battery energy consumption which will help EV drivers to plan. It will also help automobile engineers to design more efficient and scalable EVs.

A study of the effect of printing process parameters on the delamination and surface finish properties of aluminium-infused PLA printed via material extrusion

ABSTRACT Demand for the newer manufacturing techniques to produce materials with difference and specific properties are kept on increasing. The needed specific properties are obtained through composite materials by reinforcing components onto a base material to improve some critical properties. In that way, the polylactic acid (PLA) incorporated with aluminium was used to print the samples through fused filament fabrication (FFF) process. During printing, the printing process parameters have considerable effect on the material properties. At first, specimens are printed as per the Taguchi’s experimental design with L9 Orthogonal array design and the outputs are obtained experimentally. As an extension, this study proposed two neural network-based machine learning models to predict the influence of process parameters like printing temperature (PT), layer height (LH), raster angle (RA) and fill pattern (FP) on the surface finish (SF), surface finish near to the drilled holes (SFNDH), and delamination factor (DF). Both the proposed models recorded the overall prediction accuracy about 99.99%. To ensure the results predicted by the NN models, validation experiments are conducted and compared with the model predicted results. The experimental results are in good agreement with the predicted values which are obtained through numerical modelling.

Unconventional growth of methane hydrates: A molecular dynamics and machine learning study

Natural gas hydrate reservoirs in natural settings subjected to external driving forces are deformed and locally dissociated, while they reform due to endothermic dissociation reaction. Here we report a molecular dynamics (MD) and machine learning (ML) study of unconventional growth characteristics of methane hydrates (MH). MD results show that depending on the strain of MH substrate, methane clathrate hydrates grow via four distinct growth modes, in which diverse unconventional cages are critically involved. Interestingly, a novel new MH structure named as MH-VII composed of 51263, 51472 and 425861 polyhedral cages is discovered. Moreover, a long short-term memory (LSTM) neural network-based ML model using short-time MD information is developed to effectively predict the complex dynamic growth of MH in terms of clathrate cages and F4 order parameter. This work provides new insights and perspectives into the growth of clathrate hydrates, and as-developed LSTM-based ML model opens a critical pathway for exploring time-dependent behaviors of clathrate hydrates under complex conditions.

Encoding integers and rationals on neuromorphic computers using virtual neuron

Neuromorphic computers emulate the human brain while being extremely power efficient for computing tasks. In fact, they are poised to be critical for energy-efficient computing in the future. Neuromorphic computers are primarily used in spiking neural network–based machine learning applications. However, they are known to be Turing-complete, and in theory can perform all general-purpose computation. One of the biggest bottlenecks in realizing general-purpose computations on neuromorphic computers today is the inability to efficiently encode data on the neuromorphic computers. To fully realize the potential of neuromorphic computers for energy-efficient general-purpose computing, efficient mechanisms must be devised for encoding numbers. Current encoding mechanisms (e.g., binning, rate-based encoding, and time-based encoding) have limited applicability and are not suited for general-purpose computation. In this paper, we present the virtual neuron abstraction as a mechanism for encoding and adding integers and rational numbers by using spiking neural network primitives. We evaluate the performance of the virtual neuron on physical and simulated neuromorphic hardware. We estimate that the virtual neuron could perform an addition operation using just 23 nJ of energy on average with a mixed-signal, memristor-based neuromorphic processor. We also demonstrate the utility of the virtual neuron by using it in some of the μ-recursive functions, which are the building blocks of general-purpose computation.

Machine learning-assisted multi-scale modeling

Neural network-based machine learning is capable of approximating functions in very high dimension with unprecedented efficiency and accuracy. This has opened up many exciting new possibilities, one of which is to use machine learning algorithms to assist multi-scale modeling. In this review, we use three examples to illustrate the process involved in using machine learning in multi-scale modeling: ab initio molecular dynamics, ab initio meso-scale models, such as Landau models and generalized Langevin equation, and hydrodynamic models for non-Newtonian flows.

An Automated Cell Detection Method for TH-positive Dopaminergic Neurons in a Mouse Model of Parkinson's Disease Using Convolutional Neural Networks.

Quantification of tyrosine hydroxylase (TH)-positive neurons is essential for the preclinical study of Parkinson's disease (PD). However, manual analysis of immunohistochemical (IHC) images is labor-intensive and has less reproducibility due to the lack of objectivity. Therefore, several automated methods of IHC image analysis have been proposed, although they have limitations of low accuracy and difficulties in practical use. Here, we developed a convolutional neural network-based machine learning algorithm for TH+ cell counting. The developed analytical tool showed higher accuracy than the conventional methods and could be used under diverse experimental conditions of image staining intensity, brightness, and contrast. Our automated cell detection algorithm is available for free and has an intelligible graphical user interface for cell counting to assist practical applications. Overall, we expect that the proposed TH+ cell counting tool will promote preclinical PD research by saving time and enabling objective analysis of IHC images.

Battery-free and AI-enabled multiplexed sensor patches for wound monitoring.

Wound healing is a dynamic process with multiple phases. Rapid profiling and quantitative characterization of inflammation and infection remain challenging. We report a paper-like battery-free in situ AI-enabled multiplexed (PETAL) sensor for holistic wound assessment by leveraging deep learning algorithms. This sensor consists of a wax-printed paper panel with five colorimetric sensors for temperature, pH, trimethylamine, uric acid, and moisture. Sensor images captured by a mobile phone were analyzed by neural network-based machine learning algorithms to determine healing status. For ex situ detection via exudates collected from rat perturbed wounds and burn wounds, the PETAL sensor can classify healing versus nonhealing status with an accuracy as high as 97%. With the sensor patches attached on rat burn wound models, in situ monitoring of wound progression or severity is demonstrated. This PETAL sensor allows early warning of adverse events, which could trigger immediate clinical intervention to facilitate wound care management.

A machine learning potential for simulating infrared spectra of nanosilicate clusters.

The use of machine learning (ML) in chemical physics has enabled the construction of interatomic potentials having the accuracy of abinitio methods and a computational cost comparable to that of classical force fields. Training an ML model requires an efficient method for the generation of training data. Here, we apply an accurate and efficient protocol to collect training data for constructing a neural network-based ML interatomic potential for nanosilicate clusters. Initial training data are taken from normal modes and farthest point sampling. Later on, the set of training data is extended via an active learning strategy in which new data are identified by the disagreement between an ensemble of ML models. The whole process is further accelerated by parallel sampling over structures. We use the ML model to run molecular dynamics simulations of nanosilicate clusters with various sizes, from which infrared spectra with anharmonicity included can be extracted. Such spectroscopic data are needed for understanding the properties of silicate dust grains in the interstellar medium and in circumstellar environments.

Detection for COVID-19 Chest X-ray Based on Convolutional Neural Network

COVID-19 has been the most serious public health problem of the past decade. To date, the pandemic has taken a huge toll on the globe in terms of human lives lost, economic impact and increased poverty. However, due to its viral characteristics, determining whether a patient carries COVID-19 is not easy. RT-PCR methods are the gold standard for detecting COVID-19, but their time cost, as well as the need for specific equipment and instrumentation, limit their ubiquity in some medically underdeveloped areas. Chest X-rays, a test with high ubiquity and rapid results, require a certain number of professionals to read and determine whether a patient is likely to have COVID-19. Therefore, it is important to have a system to assist in the determination in areas where professionals are lacking. In this experiment, a convolutional neural network-based machine learning technique was used to create a model for recognizing COVID-19. Although there is no clinical evidence to prove its effectiveness, the model can assist professionals in judgment to a certain extent.

ænet-PyTorch: A GPU-supported implementation for machine learning atomic potentials training.

In this work, we present ænet-PyTorch, a PyTorch-based implementation for training artificial neural network-based machine learning interatomic potentials. Developed as an extension of the atomic energy network (ænet), ænet-PyTorch provides access to all the tools included in ænet for the application and usage of the potentials. The package has been designed as an alternative to the internal training capabilities of ænet, leveraging the power of graphic processing units to facilitate direct training on forces in addition to energies. This leads to a substantial reduction of the training time by one to two orders of magnitude compared to the central processing unit implementation, enabling direct training on forces for systems beyond small molecules. Here, we demonstrate the main features of ænet-PyTorch and show its performance on open databases. Our results show that training on all the force information within a dataset is not necessary, and including between 10% and 20% of the force information is sufficient to achieve optimally accurate interatomic potentials with the least computational resources.

Imaging Sensor-Based High-Throughput Measurement of Biomass Using Machine Learning Models in Rice

Phenomics technologies have advanced rapidly in the recent past for precision phenotyping of diverse crop plants. High-throughput phenotyping using imaging sensors has been proven to fetch more informative data from a large population of genotypes than the traditional destructive phenotyping methodologies. It provides accurate, high-dimensional phenome-wide big data at an ultra-super spatial and temporal resolution. Biomass is an important plant phenotypic trait that can reflect the agronomic performance of crop plants in terms of growth and yield. Several image-derived features such as area, projected shoot area, projected shoot area with height constant, estimated bio-volume, etc., and machine learning models (single or multivariate analysis) are reported in the literature for use in the non-invasive prediction of biomass in diverse crop plants. However, no studies have reported the best suitable image-derived features for accurate biomass prediction, particularly for fully grown rice plants (70DAS). In this present study, we analyzed a subset of rice recombinant inbred lines (RILs) which were developed from a cross between rice varieties BVD109 × IR20 and grown in sufficient (control) and deficient soil nitrogen (N stress) conditions. Images of plants were acquired using three different sensors (RGB, IR, and NIR) just before destructive plant sampling for the quantitative estimation of fresh (FW) and dry weight (DW). A total of 67 image-derived traits were extracted and classified into four groups, viz., geometric-, color-, IR- and NIR-related traits. We identified a multimodal trait feature, the ratio of PSA and NIR grey intensity as estimated from RGB and NIR sensors, as a novel trait for predicting biomass in rice. Among the 16 machine learning models tested for predicting biomass, the Bayesian regularized neural network (BRNN) model showed the maximum predictive power (R2 = 0.96 and 0.95 for FW and DW of biomass, respectively) with the lowest prediction error (RMSE and bias value) in both control and N stress environments. Thus, biomass can be accurately predicted by measuring novel image-based parameters and neural network-based machine learning models in rice.

Bayesian Regularization Neural Network-Based Machine Learning Approach on Optimization of CRDI-Split Injection with Waste Cooking Oil Biodiesel to Improve Diesel Engine Performance

The present study utilized response surface methodology (RSM) and Bayesian neural network (BNN) to predict the characteristics of a diesel engine powered by a blend of biodiesel and diesel fuel. The biodiesel was produced from waste cooking oil using a biocatalyst synthesized from vegetable waste through the wet impregnation technique. A multilevel central composite design was utilized to predict engine characteristics, including brake thermal efficiency (BTE), nitric oxide (NO), unburned hydrocarbons (UBHC), smoke emissions, heat release rate (HRR), and cylinder peak pressure (CGPP). BNN and the logistic–sigmoid activation function were used to train the experimental data in the artificial neural network (ANN) model, and the errors and correlations of the predicted models were calculated. The study revealed that the biocatalyst was capable of producing a maximum yield of 93% at 55 °C under specific reaction conditions, namely a reaction time of 120 min, a stirrer speed of 900 rpm, a catalyst loading of 7 wt.%, and a molar ratio of 1:9. Further, the ANN model was found to exhibit comparably lower prediction errors (0.001–0.0024), lower MAPE errors (3.14–4.6%), and a strong correlation (0.984–0.998) compared to the RSM model. B100-80%-20% was discovered to be the best formulation for emission property, while B100-90%-10% was the best mix for engine performance and combustion at 100% load. In conclusion, this study found that utilizing the synthesized biocatalyst led to attaining a maximum biodiesel yield. Furthermore, the study recommends using ANN and RSM techniques for accurately predicting the characteristics of a diesel engine.