• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.12
• /
• pp.519-524
• /
• 2023

In hashing-based image retrieval, the hash code of a manipulated image is different from the original image, making it difficult to search for the same image. This paper proposes and evaluates a self-supervised deephashing model that generates perceptual hash codes from feature information such as texture, shape, and color of images. The comparison models are autoencoder-based variational inference models, but the encoder is designed with a fully connected layer, convolutional neural network, and transformer modules. The proposed model is a variational inference model that includes a SimAM module of extracting geometric patterns and positional relationships within images. The SimAM module can learn latent vectors highlighting objects or local regions through an energy function using the activation values of neurons and surrounding neurons. The proposed method is a representation learning model that can generate low-dimensional latent vectors from high-dimensional input images, and the latent vectors are binarized into distinguishable hash code. From the experimental results on public datasets such as CIFAR-10, ImageNet, and NUS-WIDE, the proposed model is superior to the comparative model and analyzed to have equivalent performance to the supervised learning-based deephashing model. The proposed model can be used in application systems that require low-dimensional representation of images, such as image search or copyright image determination.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.25 no.6
• /
• pp.541-554
• /
• 2023

The utilization of underground spaces in relation to tunnels and energy/waste storage is on the rise. To ensure the stability of underground spaces, it is crucial to reinforce rock fractures and discontinuities. Discontinuities, such as joints, can weaken the strength of the rock and lead to groundwater inflow into underground spaces. In order to enhance the strength and stability of the area around these discontinuities, rock grouting techniques are employed. However, during rock grouting, it is impossible to visually confirm whether the grouting material is being smoothly injected as intended. Without proper injection, the expected increases in strength, durability, and degree of consolidation may not be achieved. Therefore, it is necessary to predict in advance whether the grouting material is being injected as designed. In this study, we aimed to assess the injection performance based on injection variables such as the water/cement mixture ratio, injection pressure, and injection flow using UDEC (Universal Distinct Element Code) numerical program. Additionally, numerical results were validated by the lab experiment. The results of this study are expected to help optimize variables such as injection material properties, injection time, and pump pressure in the grouting design in the field.

• Structural Engineering and Mechanics
• /
• v.90 no.6
• /
• pp.611-633
• /
• 2024

This study intends to investigate the response of multi-cell (MC) beams to flexural loads in which the primary reinforcement is composed of both metallic and non-metallic materials. "Multi-cell" describes beam sections with multiple longitudinal voids separated by thin webs. Seven reinforced concrete MC beams measuring 300×200×1800 mm were tested under flexural loadings until failure. Two series of beams are formed, depending on the type of main reinforcement that is being used. A control RC beam with no openings and six MC beams are found in these two series. Series one and two are reinforced with metallic and non-metallic main reinforcement, respectively, in order to maintain a constant reinforcement ratio. The first crack, ultimate load, deflection, ductility index, energy absorption, strain characteristics, crack pattern, and failure mode were among the structural parameters of the beams under investigation that were documented. The primary variables that vary are the kind of reinforcing materials that are utilized, as well as the kind and quantity of mesh layers. The outcomes of this study that looked at the experimental and numerical performance of ferrocement reinforced concrete MC beams are presented in this article. Nonlinear finite element analysis (NLFEA) was performed with ANSYS-16.0 software to demonstrate the behavior of composite MC beams with holes. A parametric study is also carried out to investigate the factors, such as opening size, that can most strongly affect the mechanical behavior of the suggested model. The experimental and numerical results obtained demonstrate that the FE simulations generated an acceptable degree of experimental value estimation. It's also important to demonstrate that, when compared to the control beam, the MC beam reinforced with geogrid mesh (MCGB) decreases its strength capacity by a maximum of 73.33%. In contrast, the minimum strength reduction value of 16.71% is observed in the MC beams reinforced with carbon reinforcing bars (MCCR). The findings of the experiments on MC beams with openings demonstrate that the presence of openings has a significant impact on the behavior of the beams, as there is a decrease in both the ultimate load and maximum deflection.

• Journal of Intelligence and Information Systems
• /
• v.26 no.1
• /
• pp.135-149
• /
• 2020

Artificial intelligences are changing world. Financial market is also not an exception. Robo-Advisor is actively being developed, making up the weakness of traditional asset allocation methods and replacing the parts that are difficult for the traditional methods. It makes automated investment decisions with artificial intelligence algorithms and is used with various asset allocation models such as mean-variance model, Black-Litterman model and risk parity model. Risk parity model is a typical risk-based asset allocation model which is focused on the volatility of assets. It avoids investment risk structurally. So it has stability in the management of large size fund and it has been widely used in financial field. XGBoost model is a parallel tree-boosting method. It is an optimized gradient boosting model designed to be highly efficient and flexible. It not only makes billions of examples in limited memory environments but is also very fast to learn compared to traditional boosting methods. It is frequently used in various fields of data analysis and has a lot of advantages. So in this study, we propose a new asset allocation model that combines risk parity model and XGBoost machine learning model. This model uses XGBoost to predict the risk of assets and applies the predictive risk to the process of covariance estimation. There are estimated errors between the estimation period and the actual investment period because the optimized asset allocation model estimates the proportion of investments based on historical data. these estimated errors adversely affect the optimized portfolio performance. This study aims to improve the stability and portfolio performance of the model by predicting the volatility of the next investment period and reducing estimated errors of optimized asset allocation model. As a result, it narrows the gap between theory and practice and proposes a more advanced asset allocation model. In this study, we used the Korean stock market price data for a total of 17 years from 2003 to 2019 for the empirical test of the suggested model. The data sets are specifically composed of energy, finance, IT, industrial, material, telecommunication, utility, consumer, health care and staple sectors. We accumulated the value of prediction using moving-window method by 1,000 in-sample and 20 out-of-sample, so we produced a total of 154 rebalancing back-testing results. We analyzed portfolio performance in terms of cumulative rate of return and got a lot of sample data because of long period results. Comparing with traditional risk parity model, this experiment recorded improvements in both cumulative yield and reduction of estimated errors. The total cumulative return is 45.748%, about 5% higher than that of risk parity model and also the estimated errors are reduced in 9 out of 10 industry sectors. The reduction of estimated errors increases stability of the model and makes it easy to apply in practical investment. The results of the experiment showed improvement of portfolio performance by reducing the estimated errors of the optimized asset allocation model. Many financial models and asset allocation models are limited in practical investment because of the most fundamental question of whether the past characteristics of assets will continue into the future in the changing financial market. However, this study not only takes advantage of traditional asset allocation models, but also supplements the limitations of traditional methods and increases stability by predicting the risks of assets with the latest algorithm. There are various studies on parametric estimation methods to reduce the estimated errors in the portfolio optimization. We also suggested a new method to reduce estimated errors in optimized asset allocation model using machine learning. So this study is meaningful in that it proposes an advanced artificial intelligence asset allocation model for the fast-developing financial markets.

• Journal of Animal Science and Technology
• /
• v.47 no.5
• /
• pp.731-744
• /
• 2005

Ninety Hanwoo steers(initial BW=167.2$\pm$13.4kg) were used to determine the effect of energy concentrations in concentrates and slaughter age on performance, carcass characteristics. Steers were allotted by BW to one of eighteen pens. Eighteen pens were randomly allotted to a low(70-70-71-72% for the growing, the early, the middle or the late fattening periods), medium(70-71-72-73%) or high(70-72- 73-74%) TDN level of concentrates. Five steers for each treatment of energy level were slaughtered every one month from 26 month of age to 31 month of age. Concentrates was fed restrictedly to achieve a predicted gain of 0.7-0.9kg from growing stage to middle fattening stage. All steers were fed orchard grass(Dactylis glomerata L.) hay as roughage during the growing period, fed rice straw gradually substituted for orchard grass hay during the early fattening period, and fed rice straw only thereafter. Overall body weight and feed intake were not affected by TDN levels of concentrates. Average daily gain for all treatments was higher than 0.9kg/d during the 19-21 month of age and decrease thereafter, but sustained above 0.7kg/d. Mean concentrates intake for all treatments was 1.0-1.3% of live BW during the growing period and 1.5% during the early fattening. Thereafter, it decreased up to 1.4% during the middle fattening and 1.0% during late fattening period. Delay of slaughter end point resulted in a gradual increase of rib-eye area, back fat thickness and marbling score, especially after slaughter age of 29 month there was significant increases(P<0.05). The appearance rate of 1+ and 1 grade related to the slaughter ages was 100% at 29, 30 and 31 months of age, whereas those at 26, 27 and 28 months were 93, 86 and 80%, respectively. Dressing rate was significantly(P<0.05) increased and rate of retailed cut weight significantly(P<0.05) decreased when slaughter age increased. In economic analysis, there was pronounced increase in net income up to 32-46% after slaughter age of 29 months. Under the conditions of this study, high TDN intake is not necessarily required for high quality Hanwoo meat production and slaughter age of 29 month might be the optimum for Hanwoo steers.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.48 no.6
• /
• pp.1-6
• /
• 2011

A transparent oxide thin film transistors (Transparent Oxide-TFT) have been fabricated by RF magnetron sputtering at room temperature using amorphous indium zinc oxide (a-IZO) as both of active channel and source/drain, gate electrodes and co-sputtered $HfO_2-Al_2O_3$ (HfAIO) as gate dielectric. In spite of its high dielectric constant > 20), $HfO_2$ has some drawbacks including high leakage current and rough surface morphologies originated from small energy band gap (5.31eV) and microcrystalline structure. In this work, the incorporation of $Al_2O_3$ into $HfO_2$ was obtained by co-sputtering of $HfO_2$ and $Al_2O_3$ without any intentional substrate heating and its structural and electrical properties were investigated by x-ray diffraction (XRD), atomic force microscopy (AFM) and spectroscopic ellipsometer (SE) analyses. The XRD studies confirmed that the microcrystalline structures of $HfO_2$ were transformed to amorphous structures of HfAIO. By AFM analysis, HfAIO films (0.490nm) were considerably smoother than $HfO_2$ films (2.979nm) due to their amorphous structure. The energy band gap ($E_g$) deduced by spectroscopic ellipsometer was increased from 5.17eV ($HfO_2$) to 5.42eV (HfAIO). The electrical performances of TFTs which are made of well-controlled active/electrode IZO materials and co-sputtered HfAIO dielectric material, exhibited a field effect mobility of more than $10cm^2/V{\cdot}s$, a threshold voltage of ~2 V, an $I_{on/off}$ ratio of > $10^5$, and a max on-current of > 2 mA.

• Korean Journal of Remote Sensing
• /
• v.35 no.3
• /
• pp.457-470
• /
• 2019

Automatic Target Recognition (ATR) using Synthetic Aperture Radar (SAR) has been attracted attention in the fields of surveillance, reconnaissance, and national security due to its advantage of all-weather and day-and-night imaging capabilities. However, there have been some difficulties in automatically identifying targets in real situation due to various observational and environmental conditions. In this paper, ATR problems in Extended Operating Conditions (EOC) were investigated. In particular, we considered partial occlusions of the target (10% to 50%) and differences in the depression angle between training ($17^{\circ}$) and test data ($30^{\circ}$ and $45^{\circ}$). To simulate various occlusion conditions, SARBake algorithm was applied to Moving and Stationary Target Acquisition and Recognition (MSTAR) images. The ATR accuracies were evaluated by using the template matching and Adaboost algorithms. Experimental results on the depression angle showed that the target identification rate of the two algorithms decreased by more than 30% from the depression angle of $45^{\circ}$ to $30^{\circ}$. The accuracy of template matching was about 75.88% while Adaboost showed better results with an accuracy of about 86.80%. In the case of partial occlusion, the accuracy of template matching decreased significantly even in the slight occlusion (from 95.77% under no occlusion to 52.69% under 10% occlusion). The Adaboost algorithm showed better performance with an accuracy of 85.16% in no occlusion condition and 68.48% in 10% occlusion condition. Even in the 50% occlusion condition, the Adaboost provided an accuracy of 52.48%, which was much higher than the template matching (less than 30% under 50% occlusion).

• Journal of Nutrition and Health
• /
• v.51 no.5
• /
• pp.465-479
• /
• 2018

Purpose: This study examined the frequency of sugar-sweetened beverage consumption in Korean adolescents and the major factors influencing the excessive consumption of sugar-sweetened beverages, such as sociodemographic characteristics, health-related behaviors, psychological characteristics, and dietary habits. Methods: The analysis was performed using the data from the 2017 Korean Youth Risk Behavior Web-based Survey. The subjects included in the analysis were 30,885 middle school students and 31,391 high school students, totaling 62,276. The frequency of sugar-sweetened beverage consumption was calculated by summing the consumption frequencies of soda, high caffeine or energy drinks, and sweet beverages over the last 7 days. The Rao-Scott chi-square test, reflecting information on the sampling design, was used to test the differences in the rate of sugar-sweetened beverage consumption according to each factor. Logistic regression analysis was performed to examine the factors influencing the excessive consumption of sugar-sweetened beverages. Results: The rate of sweetened beverage consumption was higher in boys than in girls, in high school students than in middle school students, in students whose father's education level was lower, in those whose subjective academic performance was lower, and in those who smoked or consumed alcohol. In addition, the rate of sugar-sweetened beverage consumption was higher in those who experienced severe stress, suicidal ideation, sadness, or a sense of despair. The rate of sugar-sweetened beverage consumption was also high in those who skipped breakfast; who frequently consumed fast foods, ramen, or snacks; and who frequently ate meals at convenience stores, supermarkets, or school stores. Conclusion: The rate of sugar-sweetened beverage consumption in Korean adolescents is related to various factors, such as sociodemographic characteristics, health-related behaviors, psychological characteristics, and dietary habits.

• Journal of Life Science
• /
• v.24 no.2
• /
• pp.105-111
• /
• 2014

The purpose of this study was to characterize equine heat-shock protein (Hsp) genes and analyze their expression pattern in various horse tissues and blood leukocytes after exercise. In a previous study, RNA sequencing of blood and skeletal muscles of thoroughbreds before and after exercise was performed using differently expressed gene (DEG) analysis. Three Hsp genes (HspH1, Hsp90${\alpha}$ and Hsp70) were selected by DEG analysis and were found to be differentially expressed in either blood or muscle. To validate and extend previous observations on these genes, we performed RT-PCR analyses of horse tissue as well as real-time qPCR analyses of blood leukocytes after exercise. mRNA expression of these Hsp genes was found to be ubiquitous in the analyzed tissues (including thyroid, colon, skeletal muscle, cecum, kidney, spinal cord, heart, and lung). In addition, Hsp mRNA expression of these genes in extracted whole blood increased after 120 minutes of exercise compared to the baseline condition. These results are in agreement with the results of human and other experimental animals, suggesting that regulatory mechanisms that are responsible for upregulation of Hsp gene transcription may be conserved among species. Further investigations to correlate Hsp gene expression patterns with athletic performance or recovery processes after exercise are warranted.

• Journal of the Korean Electrochemical Society
• /
• v.23 no.2
• /
• pp.25-38
• /
• 2020

Photoelectrochemical water splitting has been considered as the most promising technology for generating hydrogen energy. Transition metal dichalcogenide (TMD) compounds have currently attracted tremendous attention due to their outstanding ability towards the catalytic water-splitting hydrogen evolution reaction (HER). Herein, we report the synthesis method of various transition metal dichalcogenide including MoS₂, MoSe₂, WS₂, and WSe₂ nanosheets as excellent catalysts for solar-driven photoelectrochemical (PEC) hydrogen evolution. Photocathodes were fabricated by growing the nanosheets directly onto Si nanowire (NW) arrays, with a thickness of 20 nm. The metal ion layers were formed by soaking the metal chloride ethanol solution and subsequent sulfurization or selenization produced the transition metal chalcogenide. They all exhibit excellent PEC performance in 0.5 M H₂SO₄; the photocurrent reaches to 20 mA cm^-2 (at 0 V vs. RHE) and the onset potential is 0.2 V under AM1.5 condition. The quantum efficiency of hydrogen generation is avg. 90%. The stability of MoS₂ and MoSe₂ is 90% for 3h, which is higher than that (80%) of WS₂ and WSe₂. Detailed structure analysis using X-ray photoelectron spectroscopy for before/after HER reveals that the Si-WS₂ and Si-WSe₂ experience more oxidation of Si NWs than Si-MoS₂ and Si-MoSe₂. This can be explained by the less protection of Si NW surface by their flake shape morphology. The high catalytic activity of TMDs should be the main cause of this enhanced PEC performance, promising efficient water-splitting Si-based PEC cells.