Deep learning, which has recently shown excellent performance, has a problem that the amount of computation and required memory are large. Model compression is very useful because it saves memory and reduces storage size while maintaining model performance. Model compression methods reduce the number of edges by pruning weights that are deemed unnecessary in the calculation. Existing weight pruning methods using ADMM construct an optimization problem by a layer-by-layer addition of pre-defined removal-ratio constraints. Decomposing into two subproblems through the ADMM process, one can solve them through gradient descent and projection. However, the layer-by-layer removal ratios must be structurally specified, causing a sharp increase in training time due to a large number of parameters, and hardly feasible to use for large models that actually require weight pruning. Our proposed method performs weight pruning, producing similar performance, by setting a global removal ratio for the entire model without prior knowledge of structural characteristics in order to solve the shortcomings of the existing ADMM weight-pruning methods. To effectively avoid performance degradation, the method removes a relatively small number of previous layers in charge of feature extraction. Experiments show high-quality performance, not necessarily setting layer-by-layer removal ratios. Additionally, experiments increasing layers yield an insight for feature extraction in pruned layers. The experiment of the proposed method to the LeNet-5 model using MNIST data results in a higher compression ratio of 99.3% outperforming those of other existing algorithms. We also demonstrate the effectiveness of the proposed method in YOLOv4, an object detection model requiring substantial computation.

Numerical analysis has been performed to investigate the stress distribution characteristics of auxiliary axle system in a compact special vehicle. Structural and running stability of compact special vehicle is largely affected by auxiliary axle system installation. Structural characteristics of stress and deformation distributions with safe factor in the auxiliary axle system are analyzed. It can be seen that maximum stress happens near the local axle position, and deformation is also predicted. Simulation results are also compared with experimental test data. These results from this study could be applicable for optimal design of diverse axle systems in the compact special vehicle.

Most of the commercial vehicle steering system is hydraulic. In this system, Breakage and leakage of oil tank is the cause of serious problems to inhibiting the vehicle direction control of the driver. In this paper, FSI(Fluid-Structure Interaction) analysis was performed considering the sloshing of oil in the tank for structural safety evaluation of the power steering oil tank. Additionally, vibration analysis of simple and fast evaluation method was performed by assuming the fluid to mass. As a result, sloshing analysis of oil in the tank was confirmed from the FSI vibration analysis, and it could get the distribution of the flow pressure 8.2kPa～-5.5kPa. Second, stress change of the FSI vibration analysis results was greater in the 33.6MPa to 0.25MPa during the transitional period 2cycle, then it showed a stable result. Third, maximum stress of FSI vibration analysis considering sloshing of the oil was 2.22MPa. maximum stress of vibration analysis assuming the fluid to mass was 4.50MPa and 103% higher than the FSI vibration analysis, but the oil tank was safely evaluate structurally safety factor 14.1. Finally, without the FSI vibration analysis by applying a weight of 0.5 to the result of vibration analysis assumes fluid to mass, it could be obtained results similar to FSI vibration analysis.

In recent years, there was many conflagration about special structure such as wooden cultural assets, warehouses and factories. The common causes of increase in the fire damage were difficulty of the initial suppression and absence of equipment for appropriate disaster prevention. A prediction of the air injection diameter of the destruction-spray nozzle, a core technology of destruction-spray fire vehicle which is possible for fire suppression of special structure were studied. As a result, changes in water flow according to the air injection diameter is not large, but air flow rate showed a difference more than up to four times. And then, through the result data of the flow analysis, the air injection diameter was obtained in the target mix ratio of the air according to the water injection pressure. Finally, by the formula derivation for the air injection diameter of target mix ratio of the air, the air injection diameter according to the water injection pressure change could be predicted within an error of 10%.

Background : Panos extract is a mixture of four Panax plant extracts namely Dendropanax morbifera, Panax ginseng, Acanthopanax senticosus and Kalopanax septemlobus. We intended to use Panos extract for ZnO nanoparticles(NPs) synthesis and application for waste water treatment. Methods and Results : In the present study, we have synthesized Panos ZnO nanoparticles via co precipitation method. Characterization of the NPs has been done using X-ray diffraction (XRD), Fourier transformed infrared spectroscopy (FTIR) and UV-Visible spectroscopy. An average of 75% efficacy in degrading the methylene blue dye has been observed. The nanoparticles showed antibacterial activity against E. coli and S. aureus. Conclusion : The results shows that Panos ZnO NPs can be a potential eco-friendly and economical tool for waste water management in the current scenario where there an intense urge to remediate the polluted environment through novel approaches such as Nanobiotechnology.

Background : Panax ginseng the king of Oriental medicines, has enormous potential against many of the disease symptoms and ailments. Panax quinquefolius is an another important member of Panax family which has similar benefits and medicinal properties. Both the species contains ginsenosides a pharmacologically active component abundant in their roots. In an effort to establish the new ginseng cultivars with improved agronomical characters such as root shape, stress tolerant and higher ginsenoside contents, the reciprocal interspecific hybrids were generated Methods and Results : Four elite cultivars and two new lines of ginseng were chosen to produce interspecies reciprocal hybrids with P. quinquefolius, by pollen dusting on the emasculated flowers. Among the F1 populations, the reciprocal hybrids generated by using new line 0837 both as maternal and paternal parent showed better properties and hybrid vigor. They showed strong root phenotypes with many lateral and fine roots thus having 10% to 20% higher ginsenoside contents compared to the parental populations. Among the major ginsenosides, the pharmacologically active ginsenosides such as Re, Rb1, Rb2 and F1 were enriched and accounted for the 70% of the PPD ginsenosides in 0837/Pq and Pq/0837 crosses, displaying strong dominance. Conclusion : Thus, based on our result we could conclude that P. ginseng line 0837 is the superior variety compared to the already existing lines for performing the interspecific hybrids with promising outcome in their root quality and ginsenoside content.

Since 1997, Korean Ministry of Knowledge Economy and Ministry of Environment have established data on some 400 basic raw and subsidiary materials and process like energy, petro-chemical, steel, cement, glass, paper, construction materials, transportation, recycling and disposal etc by initiating establishment of LCI database. Regarding agriculture, Rural Development Administration has conducted establishment of LCI database for major farm products like rice, barley, beans, cabbage and radish etc from 2009, and released that they would establish LCI database for 50 items until 2020 later on. The domestic LCI database for seeds, seedling, agrochemical, inorganic, fertilizer and organic fertilizer etc is only at initial stage of establishment, so overseas LCI databases are brought and being used. However, since the domestic and overseas natural environments differ, they fall behind in reliability. Therefore, this study has the purpose to select organic farming materials, survey the production process for various types of organic farming materials and establish LCI database for the effects of greenhouse gas emitted by each crop during the process. As for selecting methods, in this study organic farming materials were selected in the method of direct observation of material and bottom-up method a survey method with focus on the organic farming materials admitted into rice production. For the basic unit of carbon emission amount by the production of 1kg of organic farming material, the software PASS 4.1.1 developed by Korea Accreditation Board under Ministry of Knowledge Economy was used. The study had the goal to ultimately provide basic unit to calculate carbon emission amount in executing many institutions like goal management system and carbon performance display system etc in agricultural sector to be conducted later on. As a result, emission basic units per 1kg of production were calculated to be 0.04968kg- CO2 for oystershells, 0.004692kg-CO2 for expanded rice hull, and 1.029kg-CO2 for bordeaux mixture liquid.