http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Large-Signal Output Equivalent Circuit Modeling for RF MOSFET IC Simulation
Seoyoung Hong,Seonghearn Lee 대한전자공학회 2015 Journal of semiconductor technology and science Vol.15 No.5
An accurate large-signal BSIM4 macro model including new empirical bias-dependent equations of the drain-source capacitance and channel resistance constructed from bias-dependent data extracted from S-parameters of RF MOSFETs is developed to reduce S22-parameter error of a conventional BSIM4 model. Its accuracy is validated by finding the much better agreement up to 40 ㎓ between the measured and modeled S22-parameter than the conventional one in the wide bias range.
일회용 플라스틱 커피컵 사용을 줄이기 위한 텀블러 홀더 및 캠페인 제안 : 텀블러 사용 활성화 방안 서비스 제안을 중심으로
SeoYoung Hong,Jongho Lee 한국서비스디자인학회 2018 서비스디자인융합연구 Vol.2 No.2
일회용 폐기물 문제의 해결안으로 다회용 용기사용의 필요성은 항상 언급되어왔다. 본 연구에서는 일회용 커피컵 사용자의 사용 경험 및 여정을 관찰 분석하여 총 7가지 페인포인트를 도출하였고, 이 중 분리배출의 문제가 아닌 텀블러 사용 활성화 방안의 제안을 통하여 일회용 컵의 무분별한 배출 문제의 해결을 도모하였다. 아이디어 단계에서 도출한 2가지의 컨셉을 결합하여, 일상생활에서 밀레니얼 세대들이 조금 더 자연스럽게 텀블러를 활용할 수 있도록 돕는 텀블러 홀더와 이의 활성화를 위한 프로모션 캠페인의 서비스 시나리오를 제시하였다. 이를 토대로 텀블러 사용자들이 겪고 있는 기본적인 문제를 해결함으로써, 다회용 컵의 이용율을 조금이나마 제고시키고. 이로 인한 일회용 폐기물을 실질적으로 절감할 수 있는 대책을 제안하였다. 본 연구에서는 테이크아웃 이용자들의 텀블러 현황을 살펴보고, 직접 필드 리서치를 통하여 다양한 터치포인트들을 정리하였다. 또한 실제 텀블러 이용자들의 그룹 인터뷰를 통하여 텀블러를 이용하며 실질적으로 겪게 되는 불편함을 알아보기로 하였다. 이를 토대로 텀블러 사용자들의 휴대성 문제를 해결함으로서& 텀블러 이용률을 조금이나마 재고시키고자 하였다.
Large-Signal Output Equivalent Circuit Modeling for RF MOSFET IC Simulation
Hong, Seoyoung,Lee, Seonghearn The Institute of Electronics and Information Engin 2015 Journal of semiconductor technology and science Vol.15 No.5
An accurate large-signal BSIM4 macro model including new empirical bias-dependent equations of the drain-source capacitance and channel resistance constructed from bias-dependent data extracted from S-parameters of RF MOSFETs is developed to reduce $S_{22}$-parameter error of a conventional BSIM4 model. Its accuracy is validated by finding the much better agreement up to 40 GHz between the measured and modeled $S_{22}$-parameter than the conventional one in the wide bias range.
GRADIENT EXPLOSION FREE ALGORITHM FOR TRAINING RECURRENT NEURAL NETWORKS
SEOYOUNG HONG,HYERIN JEON,BYUNGJOON LEE,CHOHONG MIN 한국산업응용수학회 2020 Journal of the Korean Society for Industrial and A Vol.24 No.4
Exploding gradient is a widely known problem in training recurrent neural networks. The explosion problem has often been coped with cutting off the gradient norm by some fixed value. However, this strategy, commonly referred to norm clipping, is an ad hoc approach to attenuate the explosion. In this research, we opt to view the problem from a different perspective, the discrete-time optimal control with infinite horizon for a better understanding of the problem. Through this perspective, we fathom the region at which gradient explosion occurs. Based on the analysis, we introduce a gradient-explosion-free algorithm that keeps the training process away from the region. Numerical tests show that this algorithm is at least three times faster than the clipping strategy.
냉각시스템 예지보전을 위한 분류기와 예측기의 혼합 적용에 관한 연구
홍서영(Seoyoung Hong),황호진(Ho-Jin Hwang) (사)한국CDE학회 2024 한국CDE학회 논문집 Vol.29 No.3
Cooling systems are essential in both residential and industrial environments, where failures can reduce production efficiency and pose safety risks. Ensuring the reliability and continuous operation through predictive maintenance is increasingly important. This study aims to provide efficient and high-performance maintenance by developing a comprehensive 3-stage algorithm that combines classifiers and predictors. Utilizing data from the ASHRAE RP-1043 project, we validated and compared the performance of various machine learning models. Our findings indicate that Support Vector Machine (SVM), XGBoost, and RandomForest models exhibit the highest performance. Consequently, we propose a fault classifier based on SVM and XGBoost, along with a remaining useful life (RUL) predictor utilizing RandomForest. By leveraging these insights, we developed a three-stage predictive maintenance algorithm that effectively combines fault classification and RUL prediction. This integrated approach enhances the ability to predict and prevent cooling system failures, ensuring their continuous and reliable operation. Our research contributes significantly to the field of predictive maintenance, providing a practical solution for maintaining the efficiency and safety of cooling systems, thereby supporting their critical role in various settings.