• Journal of the Korean Ceramic Society
• /
• v.33 no.6
• /
• pp.665-671
• /
• 1996

Ferroelectric Pb(Fe1/2Nb1/2)O3 thin films were successfully fabricated on ITO/Glass substrate by sol-gel proces-sing and characterized to determine the dielectric and electric properties. Viscosity of PEN sol measured to investigate rheological properties was 3.25 cP which was proper for coating. The sol also showed Newtonian behavior. RTA(Rapid Thermal Annealing) was used for the annealing of the thin film and 1200~1700$\AA$ thick PEN thin films were fabricated by repeating the intermediate and the final annealing. After the deposition of Pt as top electrode by vacuum evaporation dielectric and electric properties were measured. Dielectric properties of FFN thin film were enhanced by increasing the perovskite phase fraction with increasing the annealing temperature. Measured dielectric constant of 1700$\AA$ PFN thin film annealed at $650^{\circ}C$ was 890 at 1kHz Capacitatnce density and dielectric loss were 47 fF/${\mu}{\textrm}{m}$2 and 0.47 respectively. As a result of measuring Curie temperature PFN thin films had Curie point with a rang of 110~12$0^{\circ}C$ and showed broad dielectric peak at that point. Leakage current of the PFN thin films were increased with increasing the annealing tempera-ture.

• Structural Engineering and Mechanics
• /
• v.63 no.4
• /
• pp.429-438
• /
• 2017

Structural design has an imperative role in deciding the failure possibility of a Reinforced Concrete (RC) structure. Recent research works achieved the goal of predicting the structural failure of the RC structure with the assistance of machine learning techniques. Previously, the Artificial Neural Network (ANN) has been trained supported by Particle Swarm Optimization (PSO) to classify RC structures with reasonable accuracy. Though, keeping in mind the sensitivity in predicting the structural failure, more accurate models are still absent in the context of Machine Learning. Since the efficiency of multi-objective optimization over single objective optimization techniques is well established. Thus, the motivation of the current work is to employ a Multi-objective Genetic Algorithm (MOGA) to train the Neural Network (NN) based model. In the present work, the NN has been trained with MOGA to minimize the Root Mean Squared Error (RMSE) and Maximum Error (ME) toward optimizing the weight vector of the NN. The model has been tested by using a dataset consisting of 150 RC structure buildings. The proposed NN-MOGA based model has been compared with Multi-layer perceptron-feed-forward network (MLP-FFN) and NN-PSO based models in terms of several performance metrics. Experimental results suggested that the NN-MOGA has outperformed other existing well known classifiers with a reasonable improvement over them. Meanwhile, the proposed NN-MOGA achieved the superior accuracy of 93.33% and F-measure of 94.44%, which is superior to the other classifiers in the present study.

• Journal of Nutrition and Health
• /
• v.26 no.1
• /
• pp.21-33
• /
• 1993

This balance study was performed to investigate nutritional effects of guar gum and calcium supplement for 8 weeks. 36 male rate of Sprague-Dawley strain weighing 50g were blocked into six groups : 5CN, 5GN, 5CS, 5GS, FFN, 10GN. Food intake, body weight gains, FER, PER, bioavailabity of energy, protein, fat, Ca and P, and Ca & P content of femur and kidney was measured. The results obtained in this study were summarized as follows: (1) Feed intake, weight gains, FER, PER and bioavailabilities of protein and P and the contents of Ca & P in femur of 5GB were significantly higher than 5CN and 5CS (p<0.05) but was not significant different with those 5GN. Fat bioavailabilities of 5GS was significantly lower than 5CN and 5GN (p<0.05) but was not significant different with those of 5CS. Ca contents of kidney of 5GS was significantly higher than 5CN and 5GN(p<0.05) but was not significant different with those of 5SC. (2) According to guar gum levels feed intake. FER and PER were not sifgnificantly different among groups but body weight gains, bioavailabilities of energy, protein, fat and Ca, Ca & P content of femur was significantly decreased (p<0.05). Bioavailabilities of P was significantly increased according to the guar gum levels (p<0.05). Ca contents in Kidney of 10GN was significantly higher those of 5GN(p<0.05) but P contents in kidney was not significant difference by guar gum levels. Therefore 5% supplemental guar gum in diet of obese man and diebetics could be recommended. But it should be careful in using 10% guar gum in diet.

• Korean Journal of Community Nutrition
• /
• v.5 no.2
• /
• pp.273-279
• /
• 2000

This study was carried out to evaluate the calcium and phosphorus intakes and correlation with growth performance of 25male breast and formula-fed infants from 1 to 3 months postpartum. There were four groups breast-fed(BF) and three formula-fed groups(FFM, FFN and FFP). There was no significant difference in the height and weight of infants according to feeding methods and formula brands. Milk intake and the concentration of calcium and phosphorus from human milk and the formulas were measured. The average calcium content of human milk was 26.6$\pm$4.1mg/dl and 13.5$\pm$2.0mg/dl, respectively. The average calcium intakes of the brest-fed and formula-fed infants were 205.5$\pm$29.3mg/day and the 460.5$\pm$70.6mg/day, respectively the average calcium intake of breast-fed infants was significantly lower than that of formula fed infant. The percent of RDA(41%) of breast-fed infants was lower than that(75.4%) of formula fed infant. The average phosphorus intake of the brest-fed and formula-fed infants were 105.1$\pm$20.1mg/day and 288.3$\pm$37.3mg/day, respectively. The precent of RDA(27%) for brest fed was significantly lower than that(75.4%) of formula fed. The Ca/P ratios were 2.1 in brest fed and 1.6 in formula fed infant. The average calcium and phosphorus intakes of the formula fed infants was higher than that of the brest-fed infants. This data suggests the calcium and phosphorus intakes form human milk sufficiently support the growth of infants. Therefore, the level of calcium and phosphorus recommended dietary allowances for infants should be reduced.

• Korean Journal of Community Nutrition
• /
• v.4 no.1
• /
• pp.30-36
• /
• 1999

This study was carried out to longitudinally investigate the iron and zinc intakes and correlation with growth performance of 25 male breast-and formula-fed infants from 1 to 3 months postpartum, longitudinally. There were four groups breast fed(BF) and three formula fed groups((FFM, FFN and FFP). Milk intakes and the concentration of iron and zinc from human milk and the formulas were measured. The iron content of human milk was $2.07\pm{1.05(0.63~5.65)}\mu{g}$/ml. The zinc content was $2.43\pm{1.14(0.70~5.30)}\mu{/ml.}$ Both were not significantly different among postpartum months. The average iron intake of the breast-fed and formula-fed infants was $1.6\pm{0.7mg/}$day and $8.4\pm{2.3mg}$/day, respectively. The iron intake of breast-fed infants was higher than previous reports. And that of formula-fed infants was higher than the RDA. The average zinc intake of the breast-fed group was $1.9\pm{0.9mg/}$day and formula-fed infants' was $2.7\pm{0.7mg/}$day, which was higher than the RDA. There was no correlation between these mineral intakes and the growth performance during 1 to 3 postpartum months. So, extensive studies of the iron content of human milk and wide cross-sectional studies for establishing iron and zinc recommended dietary allowances for infants are needed.

• Journal of IKEEE
• /
• v.28 no.1
• /
• pp.53-64
• /
• 2024

With the advancement of NLP(Natural Language Processing) models, conversational AI such as ChatGPT is becoming increasingly popular. To enhance processing speed and reduce power consumption, it is important to implement the Transformer algorithm, which forms the basis of the latest natural language processing models, in hardware. In particular, the multi-head attention and feed-forward network, which analyze the relationships between different words in a sentence through matrix multiplication, are the most computationally intensive core algorithms in the Transformer. In this paper, we propose a new variable systolic array based on the number of input words to enhance matrix multiplication speed. Quantization maintains Transformer accuracy, boosting memory efficiency and speed. For evaluation purposes, this paper verifies the clock cycles required in multi-head attention and feed-forward network and compares the performance with other multipliers.