• Annual Conference on Human and Language Technology
• /
• 2017.10a
• /
• pp.76-81
• /
• 2017

외국어로 구성된 용어를 발음에 기반하여 자국의 언어로 표기하는 것을 음차 표기라 한다. 국가 간의 경계가 허물어짐에 따라, 외국어에 기원을 두는 용어를 설명하기 위해 뉴스 등 다양한 웹 문서에서는 동일한 발음을 가지는 외국어 표기와 한국어 표기를 혼용하여 사용하고 있다. 이에 좋은 검색 결과를 가져오기 위해서는 외국어 표기와 더불어 사람들이 많이 사용하는 다양한 음차 표기를 함께 검색에 활용하는 것이 중요하다. 음차 표기 모델과 음차 표기 대역 쌍 추출을 통해 음차 표현을 생성하는 기존 방법 대신, 본 논문에서는 신뢰할 수 있는 다양한 음차 표현을 찾기 위해 문서에서 음차 표기 후보를 찾고, 이 음차 표기 후보가 정확한 표기인지 판별하는 방식을 제안한다. 다양한 딥러닝 모델을 비교, 검토하여 최종적으로 음차 표기 대역 쌍 판별에 특화된 모델인 Distance LSTM-CNN 모델을 제안하며, 제안하는 모델의 Batch Size 영향을 줄이고 학습 시 수렴 속도 개선을 위해 Layer Normalization을 적용하는 방법을 보인다.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.14 no.6
• /
• pp.95-103
• /
• 2006

Wheel treads of E.M.U. are usually under a heavy thermal load by brake frictional heat between wheel and brake shoe and damaged by repeated thermal and mechanical loads. To examine the cause of wheel tread damage of E.M.U.'s in service running, a systematic approach has been used. This study is composed of three parts. Frictional heat analysis was conducted in the first part by finite element method. Two kinds of brake shoes in service were considered. In the second part, experimental study was carried out on a brake dynamometer. Temperatures were measured for the two brake shoes. And experimental study in service running E.M.U.'s was performed. Wheel and brake shoe temperatures were measured by using thermocouples and temperature indicating strips. Finally metallurgical characteristics were examined by a SEM/EDS and the cause of the wheel damage was analyzed. It seems that aggregated ferrous component is a main cause of the wheel tread damage.

• Journal of the Korean Association of Oral and Maxillofacial Surgeons
• /
• v.27 no.2
• /
• pp.150-156
• /
• 2001

This study was performed to investigate the changes of nerve after the injection of alcohol and glycerol at the infraorbital nerve in rats. Using the eighteen Sprague-Dawley rats, weighing $200{\sim}250g$, 99% alcohol, pure glycerol, and sterile saline was injected to the epineurium of the infraorbital nerve. Glycerol injected rats were devided into 0.01ml, 0.03ml and 0.05ml groups. The alcohol and control group were injected 0.03ml at the left infraorbital nerve. The following results were obtained by histopathological examination after 1 week, 1 month, and 3 months. A few inflammatory cell infiltration and no signs of nerve degeneration were noted in control group. Total nerve degeneration was noted in the alcohol group and no regeneration was noted in 1month, and partial regeneration was noted at 3month. The nerve degeneration was noted at the periphery of nerve bundle in 0.01ml glycerol injection group. Total degeneration was noted in the 0.03ml and 0.05ml glycerol injection group and the degree was propotional to dose. These results suggest that injection of alcohol and glycerol are effective to nerve blockage by nerve degeneration, and nerve degeneration by glycerol injection is propotional to dose and nerve regeneration by glycerol injection is inversely propotional to dose.

• International Journal of Highway Engineering
• /
• v.18 no.3
• /
• pp.115-125
• /
• 2016

PURPOSES : This study proposes a methodology to collect data necessary for microlevel emission estimation, such as second-by-second speeds and road grades, and to accordingly estimate emissions. METHODS : To ease data collection for microlevel emission estimation, a vehicle equipped with speed- and location-recording instruments as well as equipment for measuring road geometry was used. As a case study, this vehicle and the proposed methodology were used on a 10-km-long highway in Yongin City, Korea. Emissions from the vehicle during driving were estimated in various microscale driving conditions. RESULTS : Differences in the estimated emission under different microscale driving conditions cannot be ignored. Compared with the estimations obtained when second-by-second data were not considered, CO and NOx emissions were more than threefold higher when considering second-by-second speed; similarly, CO and NOx emission estimations were higher by approximately 10% and 3%, respectively, when considering second-by-second road grade. CONCLUSIONS : The proposed method can estimate vehicle emissions under real-world driving conditions in such applications as road design and traffic policy assessments.

• The Journal of the Korea Contents Association
• /
• v.4 no.4
• /
• pp.79-87
• /
• 2004

In this paper, we propose a probability map based genetic algorithm to detect lips from portrait image. The existing genetic algorithm used to get an optimal solution is modified in order to get multiple optimal solutions for lips detection. Each individual consists of two chromosomes to represent coordinates x, y in space. Also the algorithm introduce a preserving zone in the population, a modified uniform crossover, a selection without individual duplication. Using probability map of H, 5 components, the proposed algorithm has adaptability in the segmentation of objects with similar colors. In experiments, we analyzed relationships of primary parameters and found that the algorithm can apply to the detection of other ROIs easily

• International Journal of Highway Engineering
• /
• v.17 no.6
• /
• pp.65-73
• /
• 2015

PURPOSES : This study tries to develop the accident models of 4-legged signalized intersections in Busan Metropolitan city with random parameter in count model to understanding the factors mainly influencing on accident frequencies. METHODS : To develop the traffic accidents modeling, this study uses RP(random parameter) negative binomial model which enables to take account of heterogeneity in data. By using RP model, each intersection's specific geometry characteristics were considered. RESULTS : By comparing the both FP(fixed parameter) and RP modeling, it was confirmed the RP model has a little higher explanation power than the FP model. Out of 17 statistically significant variables, 4 variables including traffic volumes on minor roads, pedestrian crossing on major roads, and distance of pedestrian crossing on major/minor roads are derived as having random parameters. In addition, the marginal effect and elasticity of variables are analyzed to understand the variables'impact on the likelihood of accident occurrences. CONCLUSIONS : This study shows that the uses of RP is better fitted to the accident data since each observations'specific characteristics could be considered. Thus, the methods which could consider the heterogeneity of data is recommended to analyze the relationship between accidents and affecting factors(for example, traffic safety facilities or geometrics in signalized 4-legged intersections).

• Annual Conference on Human and Language Technology
• /
• 2014.10a
• /
• pp.127-132
• /
• 2014

본 논문은 위키피디아로부터 한국어-영어 간 병렬 문장을 추출하기 위해 이질적 언어 자원의 순차적 매칭을 적용한 유사도 계산 방법을 제안한다. 선행 연구에서는 병렬 문장 추출을 위해 언어 자원별로 유사도를 계산하여 선형 결합하였고, 토픽모델을 이용해 추정한 단어의 토픽 분포를 유사도 계산에 추가로 이용함으로써 병렬 문장 추출 성능을 향상시켰다. 하지만, 이는 언어 자원들이 독립적으로 사용되어 각 언어자원이 가지는 오류가 문장 간 유사도 계산에 반영되는 문제와 관련이 적은 단어 간의 분포가 유사도 계산에 반영되는 문제가 있다. 본 논문에서는 이질적인 언어 자원들을 이용해 순차적으로 단어를 매칭함으로써 언어 자원들의 독립적인 사용으로 각 자원의 오류가 유사도에 반영되는 문제를 해결하였고, 관련이 높은 단어의 분포만을 유사도 계산에 이용함으로써 관련이 적은 단어의 분포가 반영되는 문제를 해결하였다. 실험을 통해, 언어 자원들을 이용해 순차적으로 매칭한 유사도 계산 방법은 선행 연구에 비해 F1-score 48.4%에서 51.3%로 향상된 성능을 보였고, 관련이 높은 단어의 분포만을 유사도 계산에 이용한 방법은 약 10%에서 34.1%로 향상된 성능을 얻었다. 마지막으로, 제안한 유사도 방법들을 결합함으로써 선행연구의 51.6%에서 2.7%가 향상된 54.3%의 성능을 얻었다.

• Journal of the Korean Society for Precision Engineering
• /
• v.27 no.3
• /
• pp.73-79
• /
• 2010

A preliminary study on a length-variable rotor blade for a small unmanned helicopter has been conducted. After surveys on previous researches, and examining requirements for application to a small unmanned helicopter, a length-variable rotor blade was designed and manufactured to be driven by centrifugal force from rotor revolution with no mechanical actuator. The rotor blade was divided into a fixed inboard section and an outboard section sliding in span-wise direction. In order to determine the operating conditions of the length-variable rotor during revolution, and to derive the design variables of extension spring and rotor weight, a series of analyses from multi-body dynamics solution were conducted. The manufactured prototype was verified of its length-varying mechanism from a rotor stand, the results and required future improvements are discussed.

• The KSFM Journal of Fluid Machinery
• /
• v.15 no.6
• /
• pp.5-10
• /
• 2012

The effect of pressure-side opening length on three-dimensional flow fields and aerodynamic losses downstream of a cavity squealer tip has been investigated in a turbine rotor cascade for the squealer rim height-to-chord ratio and tip gap height-tochord ratio of $h_{st}/c$ = 5.05% and h/c = 2.0% respectively. The opening length-to-camber ratio is changed to be $OL/c_c$ = 0.0, 0.1, 0.2, 0.3, 0.4, 0.5, and 0.7 The results show that longer OL leads not only to weaker secondary flow but also to lower aerodynamic loss in the tip leakage vortex region, while it significantly widens the area with high aerodynamic loss in the passage vortex region. The aerodynamic loss coefficient mass-averaged all over the measurement plane is kept almost constant for $0.0{\leq}OL/c_c{\leq}0.3$, whereas it increases rapidly for $OL/c_c$ > 0.3 in proportion to $OL/c_c$. There is little deterioration in flow turning with increasing $OL/c_c$.

• International Journal of Highway Engineering
• /
• v.19 no.2
• /
• pp.35-44
• /
• 2017

PURPOSES: This study suggests a specific methodology for the prediction of road surface temperature using vehicular ambient temperature sensors. In addition, four kind of models is developed based on machine learning algorithms. METHODS : Thermal Mapping System is employed to collect road surface and vehicular ambient temperature data on the defined survey route in 2015 and 2016 year, respectively. For modelling, all types of collected temperature data should be classified into response and predictor before applying a machine learning tool such as MATLAB. In this study, collected road surface temperature are considered as response while vehicular ambient temperatures defied as predictor. Through data learning using machine learning tool, models were developed and finally compared predicted and actual temperature based on average absolute error. RESULTS : According to comparison results, model enables to estimate actual road surface temperature variation pattern along the roads very well. Model III is slightly better than the rest of models in terms of estimation performance. CONCLUSIONS : When correlation between response and predictor is high, when plenty of historical data exists, and when a lot of predictors are available, estimation performance of would be much better.