• Journal of The Korean Society of Agricultural Engineers
• /
• v.63 no.1
• /
• pp.103-116
• /
• 2021

Analysis of runoff is substantial for effective water management in the watershed. Runoff occurs by reaction of a watershed to the rainfall and has non-linearity and uncertainty due to the complex relation of weather and watershed factors. ANN (Artificial Neural Network), which learns from the data, is one of the machine learning technique known as a proper model to interpret non-linear data. The performance of ANN is affected by the ANN's structure, the number of hidden layer nodes, learning rate, and activation function. Especially, the activation function has a role to deliver the information entered and decides the way of making output. Therefore, It is important to apply appropriate activation functions according to the problem to solve. In this paper, ANN models were constructed to estimate runoff with different activation functions and each model was compared and evaluated. Sigmoid, Hyperbolic tangent, ReLU (Rectified Linear Unit), ELU (Exponential Linear Unit) functions were applied to the hidden layer, and Identity, ReLU, Softplus functions applied to the output layer. The statistical parameters including coefficient of determination, NSE (Nash and Sutcliffe Efficiency), NSEln (modified NSE), and PBIAS (Percent BIAS) were utilized to evaluate the ANN models. From the result, applications of Hyperbolic tangent function and ELU function to the hidden layer and Identity function to the output layer show competent performance rather than other functions which demonstrated the function selection in the ANN structure can affect the performance of ANN.

• Journal of Digital Convergence
• /
• v.20 no.4
• /
• pp.389-396
• /
• 2022

Deep Learning is based on a perceptron, and is currently being used in various fields such as image recognition, voice recognition, object detection, and drug development. Accordingly, a variety of learning algorithms have been proposed, and the number of neurons constituting a neural network varies greatly among researchers. This study analyzed the learning characteristics according to the number of neurons of the currently used SGD, momentum methods, AdaGrad, RMSProp, and Adam methods. To this end, a neural network was constructed with one input layer, three hidden layers, and one output layer. ReLU was applied to the activation function, cross entropy error (CEE) was applied to the loss function, and MNIST was used for the experimental dataset. As a result, it was concluded that the number of neurons 100-300, the algorithm Adam, and the number of learning (iteraction) 200 would be the most efficient in deep learning learning. This study will provide implications for the algorithm to be developed and the reference value of the number of neurons given new learning data in the future.

• Nuclear Engineering and Technology
• /
• v.54 no.8
• /
• pp.3059-3072
• /
• 2022

A Jacobian-Free Newton Krylov Two-Nodal Coarse Mesh Finite Difference algorithm based on Nodal Expansion Method (NEM_TNCMFD_JFNK) is successfully developed and proposed to solve the three-dimensional (3D) and multi-group reactor physics models. In the NEM_TNCMFD_JFNK method, the efficient JFNK method with the Modified Incomplete LU (MILU) preconditioner is integrated and applied into the discrete systems of the NEM-based two-node CMFD method by constructing the residual functions of only the nodal average fluxes and the eigenvalue. All the nonlinear corrective nodal coupling coefficients are updated on the basis of two-nodal NEM formulation including the discontinuity factor in every few newton steps. All the expansion coefficients and interface currents of the two-node NEM need not be chosen as the solution variables to evaluate the residual functions of the NEM_TNCMFD_JFNK method, therefore, the NEM_TNCMFD_JFNK method can greatly reduce the number of solution variables and the computational cost compared with the JFNK based on the conventional NEM. Finally the NEM_TNCMFD_JFNK code is developed and then analyzed by simulating the representative PWR MOX/UO₂ core benchmark, the popular NEACRP 3D core benchmark and the complicated full-core pin-by-pin homogenous core model. Numerical solutions show that the proposed NEM_TNCMFD_JFNK method with the MILU preconditioner has the good numerical accuracy and can obtain higher computational efficiency than the NEM-based two-node CMFD algorithm with the power method in the outer iteration and the Krylov method using the MILU preconditioner in the inner iteration, which indicates the NEM_TNCMFD_JFNK method can serve as a potential and efficient numerical tool for reactor neutron diffusion analysis module in the JFNK-based multiphysics coupling application.

• Journal of Microbiology and Biotechnology
• /
• v.31 no.4
• /
• pp.550-558
• /
• 2021

Lactobacillus kefiranofaciens contains two types of β-galactosidase, LacLM and LacZ, belonging to different glycoside hydrolase families. The difference in function between them has been unclear so far for practical application. In this study, LacLM and LacZ from L. kefiranofaciens ATCC51647 were cloned into constitutive lactobacillal expression vector pMG36e, respectively. Furtherly, pMG36n-lacs was constructed from pMG36e-lacs by replacing erythromycin with nisin as selective marker for food-grade expressing systems in Lactobacillus plantarum WCFS1, designated recombinant LacLM and LacZ respectively. The results from hydrolysis of o-nitrophenyl-β-galactopyranoside (ONPG) showed that the β-galactosidases activity of the recombinant LacLM and LacZ was 1460% and 670% higher than that of the original L. kefiranofaciens. Moreover, the lactose hydrolytic activity of recombinant LacLM was higher than that of LacZ in milk. Nevertheless, compare to LacZ, in 25% lactose solution the galacto-oligosaccharides (GOS) production of recombinant LacLM was lower. Therefore, two β-galactopyranosides could play different roles in carbohydrate metabolism of L. kefiranofaciens. In addition, the maximal growth rate of two recombinant strains were evaluated with different temperature level and nisin concentration in fermentation assay for practical purpose. The results displayed that 37℃ and 20-40 U/ml nisin were the optimal fermentation conditions for the growth of recombinant β-galactosidase strains. Altogether the food-grade Expression system of recombinant β-galactosidase was feasible for applications in the food and dairy industry.

• Journal of the Korean Society of Propulsion Engineers
• /
• v.25 no.6
• /
• pp.1-11
• /
• 2021

To develop an efficient procedure related to the flamelet library, the machine learning process based on artificial neural network(ANN) is applied for the gaseous hydrogen/liquid oxygen combustor under a supercritical pressure condition. For hidden layers, 25 combinations based on Rectified Linear Unit(ReLU) and hyperbolic tangent are adopted to find an optimum architecture in terms of the computational efficiency and the training performance. For activation functions, the hyperbolic tangent is proper to get the high learning performance for accurate properties. A transformation learning data is proposed to improve the training performance. When the optimal node is arranged for the 4 hidden layers, it is found to be the most efficient in terms of training performance and computational cost. Compared to the interpolation procedure, the ANN procedure reduces computational time and system memory by 37% and 99.98%, respectively.

• Journal of Ginseng Research
• /
• v.47 no.1
• /
• pp.9-22
• /
• 2023

As a famous herbal medicine in China and Asia, ginseng (Panax ginseng C. A. Meyer) is also known as the "King of All Herbs" and has long been used in medicine and healthcare. In addition to the obvious biological activities of ginsenosides, ginseng polysaccharides (GPs) exhibit excellent antitumor, antioxidant stress, and immunomodulatory effects. In particular, GPs can exert an antitumor effect and is a potential immunomodulator. However, due to the complexity and diversity in the structures and components of GPs, their specific physicochemical properties, and underlying mechanisms remain unclear. In this article, we have summarized the factors influencing the antitumor activity of GPs and their mechanism of action, including the stimulation of the immune system, regulation of the gut microbiota, and direct action on tumor cells

• Journal of Radiopharmaceuticals and Molecular Probes
• /
• v.9 no.1
• /
• pp.23-33
• /
• 2023

Prostate cancer ranks as the world's second most frequently diagnosed cancer among men, and is responsible for the fifth highest number of cancer-related deaths in this population. The development of effective diagnostic and therapeutic approaches for prostate cancer remains a major challenge in the field of oncology. Over the past few years, the prostate-specific membrane antigen (PSMA) has raised as a hopeful tracer for the diagnosis and treatment of prostate cancer.Various radioisotopes, such as ¹³¹I, ^99mTc, ⁶⁸Ga, and ¹⁷⁷Lu, have been used to label PSMA analogues, with varying degrees of success. Among these, ⁶⁸Ga-PSMA-11 and ¹⁷⁷Lu-PSMA-617 have emerged as the most promising radioligands for clinical use. Recently, researchers have been exploring the use of other radioisotopes, such as ²¹¹At, ⁸⁹Zr, ^64/67Cu, and ^203/212Pb, for the labeling of PSMA-targeted radioligands. These radioisotopes have unique properties that may offer advantages over existing radioligands, such as longer half-lives, higher specific activities, and different emission profiles. Efforts are currently underway to develop these radiopharmaceuticals and make them more widely available for clinical use. These exciting developments highlight the potential of PSMA-targeted radioligands for the diagnosis and treatment of prostate cancer, and provided significant implications for the management of this disease in the future. The current study aims to provide a comprehensive summary of the latest research and clinical applications of radiolabeled PSMA inhibitors for diagnoses and therapy of prostate cancer, emphasizing the exciting developments in the field and their potential impact on clinical practice.

• Journal of Microbiology and Biotechnology
• /
• v.33 no.2
• /
• pp.235-241
• /
• 2023

Hyaluronidase (HAase) can enhance drug diffusion and dissipate edema by degrading hyaluronic acid (HA) in the extracellular matrix into unsaturated HA oligosaccharides in mammalian tissues. Microorganisms are recognized as valuable sources of HAase. In this study, a new hyaluronate lyase (HAaseD) from Bacillus sp. CQMU-D was expressed in Escherichia coli BL21, purified, and characterized. The results showed that HAaseD belonged to the polysaccharide lyase (PL) 8 family and had a molecular weight of 123 kDa. HAaseD could degrade chondroitin sulfate (CS) -A, CS-B, CS-C, and HA, with the highest activity toward HA. The optimum temperature and pH value of HAaseD were 40℃ and 7.0, respectively. In addition, HAaseD retained stability in an alkaline environment and displayed higher activity with appropriate concentrations of metal ions. Moreover, HAaseD was an endolytic hyaluronate lyase that could degrade HA to produce unsaturated HA oligosaccharides. Together, our findings indicate that HAaseD from Bacillus sp. CQMU-D is a new hyaluronate lyase and with excellent potential for application in industrial production.

• The Journal of the Convergence on Culture Technology
• /
• v.9 no.4
• /
• pp.611-616
• /
• 2023

ECG(electrocardiogram) is a test used to measure the rate and regularity of heartbeats, as well as the size and position of the chambers, the presence of any damage to the heart, and the cause of all heart diseases can be found. Because the ECG signal obtained using the ECG-KIT includes noise in the ECG signal, noise must be removed from the ECG signal to apply to the deep learning. In this paper, the noise of the ECG signal was removed using the digital IIR Butterworth low-pass filter. When the performance evaluation of the three activation functions, sigmoid(), ReLU(), and tanh() functions, was compared using the deep learning model of LSTM, it was confirmed that the activation function with the smallest error was the tanh() function. Also, When the performance evaluation and elapsed time were compared for LSTM and GRU models, it was confirmed that the GRU model was superior to the LSTM model.