• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.11
• /
• pp.471-480
• /
• 2023

The Korean construction order volume in South Korea grew significantly from 91.3 trillion won in public orders in 2013 to a total of 212 trillion won in 2021, particularly in the private sector. As the size of the domestic and overseas markets grew, the scale and complexity of EPC (Engineering, Procurement, Construction) projects increased, and risk management of project management and ITB (Invitation to Bid) documents became a critical issue. The time granted to actual construction companies in the bidding process following the EPC project award is not only limited, but also extremely challenging to review all the risk terms in the ITB document due to manpower and cost issues. Previous research attempted to categorize the risk terms in EPC contract documents and detect them based on AI, but there were limitations to practical use due to problems related to data, such as the limit of labeled data utilization and class imbalance. Therefore, this study aims to develop an AI model that can categorize the contract terms based on the FIDIC Yellow 2017(Federation Internationale Des Ingenieurs-Conseils Contract terms) standard in detail, rather than defining and classifying risk terms like previous research. A multi-text classification function is necessary because the contract terms that need to be reviewed in detail may vary depending on the scale and type of the project. To enhance the performance of the multi-text classification model, we developed the ELECTRA PLM (Pre-trained Language Model) capable of efficiently learning the context of text data from the pre-training stage, and conducted a four-step experiment to validate the performance of the model. As a result, the ensemble version of the self-developed ITB-ELECTRA model and Legal-BERT achieved the best performance with a weighted average F1-Score of 76% in the classification of 57 contract terms.

• The Journal of the Society of Korean Medicine Diagnostics
• /
• v.12 no.2
• /
• pp.1-7
• /
• 2008

Objectives: The basis research of aged people's healthcare technology is very important, because our nation goes into an aged society rapidly. Especially in traditional medicine, pulse shapes of aged people were supposed to weak and empty pulse until now, but any objective measurement and compare research do not taken. Methods and Results: Therefore in this study we design observation study to find characteristic of pulse between younger group (age $20{\sim}30$) and elder group (age $43{\sim}80$) in Busan area. We used 3-dimensional pulse analyzer '3-D MAC' (Daeyomedi Co., Ltd.). And major parameters for comparison are 1) pulse height (h1), 2) applied pressure, 3) pulse power (energy), 4) high pressure duration (w) and 5) R-AI. Conclusions: According to analysis result, elder people's pulse has more large value in pulse height, applied pressure and pulse power than younger's that. It reveals that elder people in modernized lifestyle has Yang symptoms as well Yin symptoms. Consequently we need new method to care elder people's health, and this means more important direction than old text book. From now on verification research with objective diagnosis device like 3-D MAC, must be activated to build of evidence in traditional medicine. Cause traditional medicine's spirit of care and prevention is suitable to care elder people's cardiac and cerebral vascular disease those are have close relation with elder people.

• Asian Journal of Turfgrass Science
• /
• v.9 no.1
• /
• pp.81-89
• /
• 1995

The variations of the soil texture, $SiO_2$ /$Al_2$$O_3$ ratio, buffer action, exchangeable base, ex-changeable hydrogen, and mineral nutrients were investigated to estimate the grade of the soil fertility of the soil derived from the different parent rocks such as the granite in Kwangnung and the basalt in Chejudo. The results investigated were showed as follows : Basalt soils in Chejudo belong to sandy clay, light clay and sandy clay loam, while gramite soils in Kwangnung sandy loam. The $SiO_2$ /$AI_2$$O_3$ ratio of the grassland in Chejudo was 1.11 and that of the oak forest soils was 1.24, while granite soils in Kwangnung 1.54 and 1.46, respectively. The buffer actions of ba-salt soils against the N /10 HCI and $Ca(OH)_2$ were stronger than those of granite soils. The $SiO_2$/$Al_2$$O_3$ + $Fe_2$$O_3$ ratios of grassland and oak forest soils of basalt in Chejudo showed 1.10 and 1.24 respectively, while those of the grassland and oak forest of Kwangnung 1.44 and 1.33. The base exchange capacity of basalt soils which has higher value of exchangeable hydrogen was stronger than that of granite soils. But the base saturation of granite soils showed higher value than that of basalt soils. Water contents of basalt soils in Chejudo was lower than that of granite soils fo Kwangnung Basalt soils in Chejudo contain still more humus and total nitrogen than gran-ite soils in kwangnung, The amount of available nitrogen, available phosphorus and exchangeable calcium of granite soils were more than that of basalt soils, Therefore, estimating the soil fertility, granite soils in Kwangnung is higher than that of basalt soils in Chejudo.

• Electronics and Telecommunications Trends
• /
• v.38 no.1
• /
• pp.55-64
• /
• 2023

Recently, preparations for 6G have led to the increasing interest in integrated or hybrid communication networks considering low-orbit satellite communication networks with terrestrial mobile communication networks. In addition, the demand for frequency allocation for new mobile services from low-orbit small satellites to provide global internet of things (IoT) services is increasing. The operation of such satellites and terrestrial mobile communication networks may inevitably cause interference in adjacent bands and the same band frequency between satellites and terrestrial systems. Focusing on the results of the recent ITU-R WP4C meeting, this study introduces the current status of frequency sharing and interference issues between satellites and terrestrial systems, and frequency allocation issues for new mobile satellite operations. Coexistence and compatibility studies with terrestrial IMT in L band and 2.6 GHz band, operated by Inmassat and India, respectively, and a new frequency allocation study (WRC-23 AI 1.18) are carried out to reflect satellite IoT demand. For the L band, technical requirements have been developed for emission from IMT devices at 1,492 MHz to 1,518 MHz to bands above 1,518 MHz. Related studies in the 2 GHz and 2.6 GHz bands are not discussed due to lack of contributions at the recent meeting. In particular, concerning the WRC-23 agenda 1.18 study on the new frequency allocation method of narrowband mobile satellite work in the Region 1 candidate band 2,010 MHz to 2,025 MHz, Region 2 candidate bands 1,695 MHz to 1,710 MHz, 3,300 MHz to 3,315 MHz, and 3,385 MHz to 3,400 MHz, ITU-R results show no new frequency allocation to narrow mobile satellite services. Given the expected various collaborations between satellites and the terrestrial component are in the future, interference issues between terrestrial IMT and mobile satellite services are similarly expected to continuously increase. Therefore, participation in related studies at ITU-R WP4C and active response to protect terrestrial IMT are necessary to protect domestic radio resources and secure additional frequencies reflecting satellite service use plans.

• Phonetics and Speech Sciences
• /
• v.15 no.3
• /
• pp.75-82
• /
• 2023

Automatic speech recognition (ASR) has been revolutionized with deep learning-based approaches, among which self-supervised learning methods have proven to be particularly effective. In this study, we aim to enhance the performance of OpenAI's Whisper model, a multilingual ASR system on the Korean language. Whisper was pretrained on a large corpus (around 680,000 hours) of web speech data and has demonstrated strong recognition performance for major languages. However, it faces challenges in recognizing languages such as Korean, which is not major language while training. We address this issue by fine-tuning the Whisper model with an additional dataset comprising about 1,000 hours of Korean speech. We also compare its performance against a Transformer model that was trained from scratch using the same dataset. Our results indicate that fine-tuning the Whisper model significantly improved its Korean speech recognition capabilities in terms of character error rate (CER). Specifically, the performance improved with increasing model size. However, the Whisper model's performance on English deteriorated post fine-tuning, emphasizing the need for further research to develop robust multilingual models. Our study demonstrates the potential of utilizing a fine-tuned Whisper model for Korean ASR applications. Future work will focus on multilingual recognition and optimization for real-time inference.

• Korean Journal of Materials Research
• /
• v.14 no.3
• /
• pp.191-195
• /
• 2004

Tunnel junctions with AI-N barriers fabricated by microwave-excited plasma were studied. When the Al thickness, nitridation time, and annealing temperature were 1 nm (0.8 nm), 50 s (35 s), and $280^{\circ}C$ ($300^{\circ}C$), TMR ratio and resistance-area product (RA) were 49% (34%) and $3 ${\times}$ 10^4$ $\Omega$$\mu\m^2$ ($1.5 ${\times}$ 10^4$ $\Omega$$\mu\m^2$), respectively. In order to clarify the annealing temperature dependence of TMR ratio, the local transport properties were measured for Ta 5 nm/Cu 20 nm/Ta 5 nm$29_{76}$ $Fe_{24}$ 2 nm/Cu 5 nm/M $n_{75}$$Ir_{25}$ 10 nm/ $Co_{71}$ $Co_{29}$ 4nm/Al-N junction with Al thickness of 0.8 nm and nitridation time of 35s at various temperatures. The increase of TMR ratio after annealing at $300^{\circ}C$, where the TMR ratio of the corresponding MTJ had the maximum value of 34%, can be well explained by the enhancement of the average barrier height ($\Phi_{ave}$) and the reduction of its fluctuation. After further annealing at $340^{\circ}C$, the leakage current was observed and the TMR ratio decreaseded

• Reproductive and Developmental Biology
• /
• v.34 no.3
• /
• pp.235-240
• /
• 2010

This study was performed to comprehend the plasma proteins expressed specifically during early pregnancy in pregnant or non-pregnant Hanwoo using proteomic analysis technique. Plasma samples (0, 2, 3, 4, 7, and 11 weeks after AI) were obtained from pregnant (P, n=3) or non-pregnant (NP, n=4) Hanwoo, respectively. To evaluate proteins differentially expressed, 2-dimensional electrophoresis (2DE) was conducted. Normalized protein spots were selected for the significant expression variation deviated over two fold in its expression level between two groups; Molecular functions of the proteins were DNA binding, protein binding, hemoglobin binding, ferrochelatase and transporter activity and arylestera, respectively. According to western blotting, haptoglobin was specifically expressed only in NP group during early pregnancy; however, paraoxonase 1 was highly expressed in pregnant group. Based on these results, pregnancy was maintained successfully by the activation of specific plasma proteins associated with immune system and antioxidant regulation during early pregnancy in Hanwoo.

• Korean journal of food and cookery science
• /
• v.21 no.5
• /
• pp.616-628
• /
• 2005

The purpose of this study was to monitor the current nutrition labeling practice on the packaging of the processed foods providing consumers with reliable nutrition information, which has been considered as a useful aid for food selection and a potent educational tool for nutrition in a daily life. The 2,214 processed foods purchased at the a wholesale market in July, 2003, were divided by food categories issued from the 2003 food codes and assessed in the terms of the nutrition composition labeling and nutrition claims. Nutrition composition labeling was found on 533 out of the 2,214 processed foods items. The findings of this study were as follows : Milk and dairy products had 66.4% of nutrition composition labeling, which was the largest percentage among the food categories. Tables format (71.3%) is the most widely used type of nutrition labeling. The title of nutrition composition labeling used the most widely was nutrition composition (71.3%). Nutrition composition included many different ways of expression, such as a table of nutrition composition, and indication of nutrition composition, etc. The expression unit of the nutrition composition labeling was use per 100g or 100mL (40.5%). The processed foods with nutrition claims were 18.0% (399 items). Nutrition claims were divided into two categories: nutrient content claims and nutrient comparative claims. The most frequently used nutrient content claims were contained (43.9%) and more or plus in the nutrient comparative claims (16.0%). Ca was the most popular item as a nutrition claim nutrient (33.6%).

• The Journal of Korean Academic Society of Nursing Education
• /
• v.23 no.4
• /
• pp.419-429
• /
• 2017

Purpose: The purpose of this study was to investigate the effect of the development and application of the new nurses' core skill program on the clinical skill, confidence, knowledge and performance of new nurses. Methods: This study was a nonequivalent control group post-test design on 46 new nurses with 3 to 6 months work experience in a general hospital. Data were collected from August 2015 to August 2016. A training program of core fundamental nursing skills was applied on 23 new nurses from 2016 in the experiment group, and an existing training method was applied on 23 new nurses from 2015 in the comparison group. The collected data were analyzed using SPSS/WIN 23.0. Results: The experimental group ($89{\pm}4.29$) showed higher score in practical skill assessment compared to the control group ($85.09{\pm}3.99$) and showed a statistically significant difference (p<001). Conclusion: A training program of core fundamental nursing skills for new nurses proved effective in improving practical skills. Further research on the development of a systemic training program that can improve clinical nursing knowledge and work ability for new nurses according to the extended enforcement of a Nursing Care Integrated Service ward is needed.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2007.06a
• /
• pp.51-52
• /
• 2007

Built-in voltage in organic light-emitting diodes was studied using modulated photocurrent technique ambient conditions. From the bias voltage-dependent photocurrent, built-in voltage of the device is determined. The applied bias voltage when the magnitude of modulated photo current is zero corresponds to a built-in voltage. Built-in voltage in the device is generated due to a difference of work function of the anode and cathode. A device was made with a structure of anode/$Alq_3$/cathode to study a built-in voltage. ITO and ITO/PEDOT:PSS were used as an anode, and Al and LiF/AI were used as a cathode. It was found that an incorporation of PEDOT:PSS layer between the ITO and $Alq_3$ increases a built-in voltage by about 0.4V. This is consistent to a difference of a highest occupied energy states of ITO and PEDOT:PSS. This implies that a use of PEDOT:PSS layer in anode improves the efficiency of the device because of a lowering of anode barrier height. With a use bilayer cathode system LiF/Al, it was found that the built-in voltage increases as the LiF layer thickness increases in the thickness range of 0~1nm. For 1nm thick LiF layer, there is a lowering of electron barrier by about 0.2eV with respect to an Al-only device. It indicates that a very thin alkaline metal compound LiF lowers an electron barrier height.