• 한국컴퓨터정보학회논문지
• /
• 제25권5호
• /
• pp.187-197
• /
• 2020

최근 신경망 기반의 학습 알고리즘인 딥 러닝 기술의 발전으로 인해 텍스트의 문맥을 고려한 문서 임베딩 모델이 다양하게 고안되었으며, 특히 대량의 텍스트 데이터를 사용하여 학습을 수행한 사전 학습 언어 모델을 사용하여 분석 문서의 벡터를 추론하는 방식의 임베딩이 활발하게 연구되고 있다. 하지만 기존의 사전 학습 언어 모델을 사용하여 새로운 텍스트에 대한 임베딩을 수행할 경우 해당 텍스트가 가진 고유한 정보를 충분히 활용하지 못한다는 한계를 가지며, 이는 특히 텍스트가 가진 토큰의 수에 큰 영향을 받는 것으로 알려져 있다. 이에 본 연구에서는 다수의 토큰을 포함한 장문 텍스트의 정보를 최대한 활용하여 해당 텍스트의 벡터를 도출할 수 있는 자기 지도 학습 기반의 사전 학습 언어 모델 미세 조정 방법을 제안한다. 또한, 제안 방법론을 실제 뉴스 기사에 적용하여 문서 벡터를 도출하고 이를 활용하여 뉴스의 카테고리 분류 실험을 수행하는 외부적인 임베딩 평가를 수행함으로써, 제안 방법론과 기존 문서 임베딩 모델과의 성능을 평가하였다. 그 결과 제안 방법론을 통해 도출된 벡터가 텍스트의 고유 정보를 충분히 활용함으로써, 문서의 특성을 더욱 정확하게 표현할 수 있음을 확인하였다.

• 대한공간정보학회지
• /
• 제22권1호
• /
• pp.47-54
• /
• 2014

휴대용 단말에 간편하게 구현 가능하도록 GPS의 코드의사거리 관측치 기반의 정밀단독측위(PPP; Precise Point Positioning) 알고리즘을 개발하고 그 성능을 검증하였다. PPP에 필요한 기본 모델로 그룹 딜레이, 상대성 효과, 위성안테나 위상중심오프셋 보정모델을 적용하였다. 위성 궤도와 시계오차는 IGS(International GNSS Service) 공식 산출물에 최적의 알고리즘을 통해 보간하고, 대류권과 전리층 오차는 각각 과학기술용 GPS 자료처리 소프트웨어로 산출한 참값과 GIM(Global Ionosphere Model)을 사상함수를 적용해 시선방향 오차로 변환해 적용하였다. 개발된 알고리즘을 4일간 테스트한 결과 수평오차는 0.8~1.6m, 수직오차는 1.6~2.2m 수준으로 나타났다. 이는 DGPS 측위결과와 유사한 성능으로 향후 PPP 알고리즘의 추가개선이 이루어질 경우 다양한 측량 및 위치기반서비스 분야에 활용 가능할 것으로 기대된다.

• 전기전자학회논문지
• /
• 제20권3호
• /
• pp.279-284
• /
• 2016

This paper presents a two stage L-band power amplifier realized with a $0.32{\mu}m$ Silicon-On-Insulator (SOI) CMOS technology. To overcome a low breakdown voltage limit of MOSFET, stacked-FET structures are employed, where three transistors in the first stage amplifier and four transistors in the second stage amplifier are connected in series so that their output voltage swings are added in phase. The stacked-FET structures enable the proposed amplifier to achieve a 21.5 dB small-signal gain and 15.7 dBm output 1-dB compression power at 1.9 GHz with a 122 mA DC current from a 4 V supply. The amplifier delivers a 19.7 dBm. This paper presents a two stage L-band power amplifier realized with a $0.32{\mu}m$ Silicon-On-Insulator (SOI) CMOS technology. To overcome a low breakdown voltage limit of MOSFET, stacked-FET structures are employed, where three transistors in the first stage amplifier and four transistors in the second stage amplifier are connected in series so that their output voltage swings are added in phase. The stacked-FET structures enable the proposed amplifier to achieve a 21.5 dB small-signal gain and 15.7 dBm output 1-dB compression power at 1.9 GHz with a 122 mA DC current from a 4 V supply. The amplifier delivers a 19.7 dBm saturated output power with a 16 % maximum Power Added Efficiency (PAE). A bond wire fine tuning technology enables the amplifier a 23.67 dBm saturated output power with a 20.4 % maximum PAE. The die area is $1.9mm{\times}0.6mm$.

• 한국재료학회지
• /
• 제19권12호
• /
• pp.637-643
• /
• 2009

In this study, we analyzed the effect of silicon oxynitride matrix on the optical properties of Au nanoparticles dispersed on composite film and explored the effectiveness of the silicon in fine tuning the refractive index of the composite film for applications in optical waveguide devices. The atomic fraction of nitrogen in $SiO_xN_y$ films was controlled by varying the relative flow ratio of nitrogen gas in reactive sputtering and was evaluated optically using an effective medium theory with Bruggeman geometry consisting of a random mixture between $SiO_2$ and $Si_3N_4$. The Au nanoparticles were embedded in the $SiO_xN_y$ matrix by employing the alternating deposition technique and clearly showed an absorption peak due to the excitation of surface plasmon. With increasing nitrogen atomic fraction in the matrix, the surface plasmon resonance wavelength shifted to a longer wavelength (a red-shift) with an enhanced resonance absorption. These characteristics were interpreted using the Maxwell-Garnett effective medium theory. The formation of a guided mode in a slab waveguide consisting of 3 $\mu$m thick Au:$SiO_xN_y$ nanocomposite film was confirmed at the telecommunication wavelength of 1550 nm by prism coupler method and compared with the case of using $SiO_2$ matrix. The use of $SiO_xN_y$ matrix provides an effective way of controlling the mode confinement while maintaining or even enhancing the surface plasmon resonance properties.

• 한국재료학회지
• /
• 제5권6호
• /
• pp.650-656
• /
• 1995

이동통신용 고주파 표면탄성파(SAW) 필터를 36$^{\circ}$Y-X LiTaO$_3$압전기판과 IIDT 전극구조를 이용하여 제작하였다. 제작한 SAW 필터의 중심주파수는 주파수 응답 특성 측정 결과 설계한 중심주파수보다 낮아짐이 관찰되었다. 이러한 단점을 보완하고 미세한 주파수의 조절을 가능하게 하기 위해 SiN$_{x}$ 유전박막을 보호막으로 증착하여 이에 따른 주파수 특성 변화를 관찰하였다. SAW파장에 대한 SiN$_{x}$ 유전 박막의 두께비를 증가시킬수록 SAW 진행 속도가 증가하여, 제조한 필터의 중심주파수를 높게 이동시킬 수 있었다. 그러나 유전박막의 두께가 증가할수록 필터의 삽입손실이 증가하는 문제점이 존재하였다.

• 한국지능시스템학회논문지
• /
• 제9권5호
• /
• pp.555-565
• /
• 1999

복잡하고 비선형적인 시스템의 규칙베이스 퍼지모델링을 위하여 퍼지시스템의 최적 동정알고리즘을 연구한다. 비선형 시스템은 퍼지모델의 입력변수와 퍼지 입력공간 분할에 의한 구조동정과 파라미터 동정을 통해 표현된다. 본 논문에서 규칙베이스 퍼지모델링은 비선형 시스템을 위해 퍼지추론방법과 두 종류의 최적화 이론의 결합에 의한 하이브리드 구졸를 이용하여 시스템 구조와 파라미터동정을 수행한다. 퍼지모델의 추론방법은 간략추론 및 선형추론에 의한다. 제안된 하이브리드 최적 동정 알고리즘은 유전자 알고리즘과 개선된 콤플렉스 방법을 이용한다. 여기서 유전자 알고리즘은 전반부 퍼지규칙의 멤버쉽함수의 초기 파라미터들을 결정하기 위해 사용되고 강력한 자동동조 알고리즘인 개선된 콤플렉스 방법은 정교한 파라미터들을 얻기 위해 수행된다. 따라서 최적 퍼지모델을 위해 전반부 파라미터 동정에는 하이브리드형의 최적 알고리즘을 이용하고 후반부 동정에는 최소자승법을 이용한다. 또한 학습과 테스트 데이터에 의해 생성된 퍼지모델의 성능결과 사이의 상호균형을 얻기 위해 하중계수를 가지는 합성 성능지수를 제안한다. 제안된 모델의 성능평가를 위해 두가지 수치적 예를이용한다.

• Journal of Construction Engineering and Project Management
• /
• 제1권3호
• /
• pp.18-27
• /
• 2011

Although information communication technology (ICT) is long regard as very useful tool in today's construction engineering and project management environment, organizations must not only operate based upon its original setting, but also requires on-going observation, additional features and fine-tuning actions before the desirable outcome can be achieved. However, it is a very common phenomenon that organizations purchase the licensed "off-the-shelf-software" package and customize it to suit their own business need. Due to the incapability of such software and inefficient customization, the possible result is making that ICT tool not user-friendly and sometimes the whole system becomes obsolete. The purpose of this paper is to review and report those actions taken (between February 2006 and December 2010) by a construction organization to enhance the performance of its Enterprise Resource Planning (ERP) system launched in December 2002. Such actions include: improving data inputting method; removing the transition bottleneck; introducing crystallization function; revising the organization's "Delegation and Limits of Authority"; publishing the "League Table" amongst users; integrating the 3D Mode into the system and upgrading hardware. Whilst the ultimate goals of such system are well beyond the time limit of this research study, an obvious interim result, achieved by this case studied organization, was winning a landmark project worth US$500 million after the ERP system was functioned properly and effectively. Their experience and success becomes an exemplar which can be borrowed by those companies, from managerial perspectives and as a roadmap, planning to adopt information technology (IT) strategy and use ICT tool in the construction engineering and project management framework. Singapore, where public housing provisions have been a major concern of their citizens as the building stock gets older.

• 대한전기학회논문지:시스템및제어부문D
• /
• 제53권6호
• /
• pp.415-424
• /
• 2004

In this paper, we introduce and investigate a new category of rule-based fuzzy inference system based on Information Granulation(IG). The proposed rule-based fuzzy modeling implements system structure and parameter identification in the efficient form of “If..., then...” statements, and exploits the theory of system optimization and fuzzy implication rules. The form of the fuzzy rules comes with three types of fuzzy inferences: a simplified one that involves conclusions that are fixed numeric values, a linear one where the conclusion part is viewed as a linear function of inputs, and a regression polynomial one as the extended type of the linear one. By the nature of the rule-based fuzzy systems, these fuzzy models are geared toward capturing relationships between information granules. The form of the information granules themselves becomes an important design features of the fuzzy model. Information granulation with the aid of HCM(Hard C-Means) clustering algorithm hell)s determine the initial parameters of rule-based fuzzy model such as the initial apexes of the membership functions and the initial values of polynomial function being used in the Premise and consequence Part of the fuzzy rules. And then the initial Parameters are tuned (adjusted) effectively with the aid of the improved complex method(ICM) and the standard least square method(LSM). In the sequel, the ICM and LSM lead to fine-tuning of the parameters of premise membership functions and consequent polynomial functions in the rules of fuzzy model. An aggregate objective function with a weighting factor is proposed in order to achieve a balance between performance of the fuzzy model. Numerical examples are included to evaluate the performance of the proposed model. They are also contrasted with the performance of the fuzzy models existing in the literature.

• 한국운동역학회지
• /
• 제22권4호
• /
• pp.429-435
• /
• 2012

The purpose of this study was to determine the biomechanical effect of wearing carbon nanotube-based insole on cushioning and muscle tuning during drop landing. Twenty male university students(age: $21.2{\pm}1.5yrs$, height: $175.4{\pm}4.7cm$, weight: $70.2{\pm}5.8kg$) who have no musculoskeletal disorder were recruited as the subjects. Average axial strain, average shear strain, inversion angle, linear velocity, angular velocity, vertical GRF and loading rate were determined for each trial. For each dependent variable, a one-way analysis of variance(ANOVA) with repeated measures was performed to test if significant difference existed among different three conditions(p<.05). The results showed that Average axial strain of line 4 was significantly less in CNT compared with EVA and PU during IP phase. The average shear strain was less in CNT compared with EVA and PU during other phases. The inversion angle was increased in CNT compared with EVA and PU during all phase. In linear velocity, angular velocity, vertical GRF and loading rate, there were no significant difference between the three groups. This result seems that fine particle of carbon nanotube couldn't make geometric form which can absolve impact force by increasing density through eliminating voids of forms. Thus, searching for methods that keep voids of forms may play a pivotal role in developing of insole. This has led to suggestions of the need for further biomechanical analysis to these factors.

• 한국약용작물학회지
• /
• 제23권2호
• /
• pp.138-143
• /
• 2015

This study was carried out to evaluate the preventive effect of three forms of Korean ginseng roots (fresh, white and red) against bisphenol A (BPA) toxicity in mouse male germ cells (GC-2spd, TM3, TM4). ROS (reactive oxygen species) generation were measured by DCF-DA (2',7'-dichlorohydrofluorescein diacetate) assay. Also, semi-quantitative reverse transcriptase polymerase chain reaction (RT-PCR) was performed to quantify the mRNA expression levels of apoptosis-related genes, Bax (pro-apoptotic gene) and Bcl2 (anti-apoptotic gene). ROS generation was increased by $50{\mu}M$ BPA, but definitely decreased by treatment with Korean ginseng extracts (fresh, white and red) in mouse male germ cells. In especial, Korean fresh ginseng extract reduced significantly ROS production to normal control. In addition, Korean fresh and white ginseng extracts suppressed the apoptosis of mouse male germ cells by fine-tuning mRNA levels of apoptotic genes changed by BPA. In general, Korean fresh ginseng extract was more effective than white ginseng extract for reducing BPA-induced oxidative stress and apoptosis in mouse male germ cells. Therefore, Korean fresh and white ginseng may help to alleviate biphenol A toxicity in mouse male germ cells.