• Proceedings of the Korean Information Science Society Conference
• /
• 2010.06c
• /
• pp.380-384
• /
• 2010

기후온난화, 국지성 호우 및 대규모 태풍으로 인한 피해가 증대되면서 사회 경제적 손실 또한 날로 증가하고 있어 재해로 인한 피해 발생가능성을 효율적으로 예측하는 모델을 통한 선제적 대응이 필요하다. 재난 재해의 위험성 분석 방법은 주로 확률 통계기법을 기반으로 하는 연구가 주류를 이루었으나, 본 논문에서는 포착된 현상의 데이터를 이용해 그 데이터를 지배하는 경험적 규칙성을 학습하고 획득하는데 다른 기법보다 탁월한 성능을 가진 신경망 모델을 적용하여 자연재해 피해예측 모델을 연구하였다. 1991년부터 2005년 사이에 우리나라에서 발생한 자연재해의 피해자료와 기상개황 자료를 이용하여 지역별 자연재해로 인한 피해를 예측하는 신경망 모델은 우리나라 232개 행정구역에 대하여 누적강우량과 최대풍속, 그리고 재해사상 발생 5일 이내의 선행강우량을 입력변수로 하고 총 피해액을 출력변수로 한다. 또한 학습을 통한 최적의 해를 찾기 위해 신경망의 매개변수 학습률, 모멘텀, 편의값을 유전자알고리즘으로 결정하여 학습을 수행 하였다.

• Annual Conference of KIPS
• /
• 2024.05a
• /
• pp.604-606
• /
• 2024

컴퓨터 비전 응용은 우리 생활에서 중요한 역할을 한다. 현재, 대규모 모델의 등장으로 딥 러닝의 훈련 및 운행 비용이 급격히 상승하고 있다. 자원이 제한된 환경에서는 일부 AI 프로그램을 실행할 수 없게 되므로, 경량화 연구가 필요하다. YOLOv8 은 현재 주요 목표 검출 모델 중 하나이며, 본 논문은 다중 스케일 Ghost 컨볼루션 모듈을 사용하여 구축된 새로운 YOLOv8-pose-msg 키포인트 검출 모델을 제안한다. 다양한 사양에서 새 모델의 매개변수 양은 최소 34% 감소할 수 있으며, 최대 59%까지 감소할 수 있다. 종합적인 검출 성능은 비교적 대규모 데이터셋에서 원래의 수준을 유지할 수 있으며, 소규모 데이터셋에서의 키포인트 검출은 30% 이상 증가할 수 있다. 동시에 최대 25%의 훈련 및 추론 시간을 절약할 수 있다.

• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.618-621
• /
• 2023

최근 거대언어모델(LLM)이 기계 번역 및 기계 독해를 포함한 다양한 문제들에서 높은 성능을 보이고 있다. 특히 프롬프트 기반의 대규모 언어 모델은 사고사슬 방식으로 적절한 프롬프팅을 통해 원하는 형식의 답변을 생성할 수 있으며 자연어 추론 단계에서도 높은 정확도를 보여주고 있다. 그러나 근본적으로 LLM의 매개변수에 질문에 관련된 지식이 없거나 최신 정보로 업데이트 되지 않은 경우 추론이 어렵다. 이를 해결하기 위해, 본 연구는 검색문서와 생성모델의 상호작용을 통해 답변하는 한국어 질의응답 모델을 제안한다. 검색이 어려운 경우 생성형 모델을 통해 질문과 관련된 문장을 생성하며, 이는 다시 검색모델과 추론 과정에서 활용된다. 추가로 "판단불가"라는 프롬프팅을 통해 모델이 답변할 수 없는 경우를 스스로 판단하게 한다. 본 연구결과에서 GPT3를 활용한 사고사슬 모델이 63.4의 F1 점수를 보여주며 생성형 모델과 검색모델의 융합이 적절한 프롬프팅을 통해 오픈-도메인 질의응답에서 성능의 향상을 보여준다.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.3 no.3
• /
• pp.15-29
• /
• 2001

For the pertinent use of NMT method, both characteristics of joints (JRC, JCS and ${\phi}_r$) and characteristics of rock mass (Q-Value) must be investigated carefully. The main objective of the study presented is to investigate how sensitive the predicted behaviour of an underground excavation is to various realistic assumptions about some input parameter for the jointed rock mass. Joint pattern in the tunnel is predicted by statistical approach (chi-square test). In this paper, sensitivity studies involving in joint characteristics were carried out. The parametric studies involving change in Barton-Bandis joint model have shown that JCS is relatively insensitive to JRC and ${\phi}_r$. An increase in JRC value may not, according to the Barton-Bandis model, necessarily lead to a decrease in displacement. The importance of dilation in predicting the behaviour of a rock mass around an excavation is emphasized from a comparison of the Barton-Bandis joint behaviour model with the Mohr-Coulomb model. The Barton-Bandis model predicted higher stress, which allow for the build-up of stress caused by dilatant behaviour.

• Journal of the Korean GEO-environmental Society
• /
• v.18 no.12
• /
• pp.47-57
• /
• 2017

In this study, a predictive analysis was conducted on sediment disaster hazard area by selecting six research areas (Chuncheon, Seongnam, Sejong, Daejeon, Miryang and Busan) among the urban sediment disaster preliminary focus management area. The models that were used in the analysis were the existing models (SINMAP and TRIGRS) that are commonly used in predicting sediment disasters as well as the program developed through this study (LSMAP). A comparative analysis was carried out on the results as a means to review the applicability of the developed model. The parameters used in the predictions of sediment disaster hazard area were largely classified into topographic, soil, forest physiognomy and rainfall characteristics. A predictive analysis was carried out using each of the models, and it was found that the analysis using SINMAP, compared to LSMAP and TRIGRS, resulted in a prediction of a wider hazard zone. These results are considered to be due to the difference in analysis parameters applied to each model. In addition, a comparison between LSMAP, where the forest physiognomy characteristics were taken into account, and TRIGRS showed that similar tendencies were observed within a range of -0.04~2.72% for the predicted hazard area. This suggests that the forest physiognomy characteristics of mountain areas have diverse impacts on the stability of slopes, and serve as an important parameter in predicting sediment disaster hazard area.

• Journal of Korea Water Resources Association
• /
• v.55 no.10
• /
• pp.823-833
• /
• 2022

This study aims to improve the runoff modeling accuracy of a basin using Hydrological Simulation Program-FORTRAN (HSPF) model by considering nonhomogeneous characteristics of a basin. By entering classified values according to the various types of land cover and soil to the parameters in HSPF-roughness coefficient (NSUR), infiltration (INFILT), and evapotranspiration (LZETP)- the heterogeneity of the Yongdam Dam basin was reflected in the model. The results were analyzed and compared with the one where the parameters were set as a single value throughout the basin. The flow rate and water quality simulation results showed improved results when classified parameters were used by land cover and soil type than when single values were used. The parameterization changed not only the flow rate, but also the composition ratio of each hydrologic components such as surface runoff, baseflow, and evapotranspiration, which shows the impact of the value set to a parameter on the entire hydrological process. This implies the importance of considering the heterogeneous characteristics of the land cover and soil of the basin when setting the parameters in a model.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.36 no.6
• /
• pp.391-398
• /
• 2023

In this paper, a penalized maximum likelihood estimation (PMLE) method that applies a penalty to increase the accuracy of a basis-screening-based Kriging model (BSKM) is introduced. The maximum order and set of basis functions used in the BSKM are determined according to their importance. In this regard, the cross-validation error (CVE) for the basis functions is employed as an indicator of importance. When constructing the Kriging model (KM), the maximum order of basis functions is determined, the importance of each basis function is evaluated according to the corresponding maximum order, and finally the optimal set of basis functions is determined. This optimal set is created by adding basis functions one by one in order of importance until the CVE of the KM is minimized. In this process, the KM must be generated repeatedly. Simultaneously, hyper-parameters representing correlations between datasets must be calculated through the maximum likelihood evaluation method. Given that the optimal set of basis functions depends on such hyper-parameters, it has a significant impact on the accuracy of the KM. The PMLE method is applied to accurately calculate hyper-parameters. It was confirmed that the accuracy of a BSKM can be improved by applying it to Branin-Hoo problem.

• Asia-Pacific Journal of Business Venturing and Entrepreneurship
• /
• v.17 no.4
• /
• pp.73-86
• /
• 2022

Interest and investment in agri-food ventures is increasing worldwide. Agri-food venture startups are also increasing in Korea, but there are constraint factors due to industrial characteristics, and infrastructure and investment for growth are insufficient. This study confirmed the effect of technology commercialization capability and competitive strategy on the growth prospects of agri-food venture companies and the mediating effect of business model innovation. And, through comparison with previous studies, the industrial characteristics were identified by confirming the difference between all venture companies and agri-food ventures. For this, only agri-food ventures data were extracted from the original data of the Research on the Precision Status of Venture Firms 2021. And empirical analysis was conducted. As a result, manufacturing capability, cost leadership and product differentiation had a positive(+) effect on growth prospects. The mediating effect of business model innovation between manufacturing capacity, product differentiation and growth prospects was verified. Through the research results, factors affecting the long-term growth prospects of agri-food ventures were identified in consideration of industrial characteristics. In addition, practical implications in connection with business model innovation of agri-food ventures were presented. In future research, it is necessary to develop an objective measurement tool and to conduct a detailed analysis according to business history and company size.

• Asia-Pacific Journal of Business Venturing and Entrepreneurship
• /
• v.16 no.2
• /
• pp.79-93
• /
• 2021

The question of why do you want to start a "start?" Is the most basic step in trying to do something. In other words, previous studies have shown that the degree of confidence in an individual's decision affects the setting of a specific purpose. Based on this, this study aims to provide basic data for deriving the direction of entrepreneurship education in college students by analyzing the effects of goal-oriented behavioral model on college students' self-determination and intention to start a business through the 5 factor model. To achieve the purpose of the study, a self-report questionnaire was conducted from October 01 to November 11, 2019 for university students attending located in Gyeonggi-do, Seoul. A total of 150 questionnaires were distributed, and 125 parts were used for the final analysis, except 25 parts with insincere responses or errors. Data were analyzed using SPSS Win 24, and reliability, validity analysis, frequency analysis, One-way ANOVA and regression analysis were performed, and three-step regression analysis and Sobel verification were performed for mediating effects. The summary of the study is as follows. First, the influence of university students' goal-oriented behavioral model on self-determination showed that attitudes, subjective norms, and perceived behavioral controls had statistically significant positive effects, and positive and negative expectations were statistically significant. Did not affect. Therefore, the higher the attitude, subjective norms, and perceived behavioral control, the higher the university students' self-determination. Second, the influence of college students' goal-oriented behavioral model on the intention to start a business was as follows.). As a result, the higher the perceived behavioral control and positive expectation, the higher the intention to start up. Third, regression model 1 showed that the behavioral control and positive expectation sentiment among the goal-oriented behavioral model had a significant positive influence on the college students' intention to start a business. Affected. Regression model II added the parameters of the 5 factor model, which increased 2.5% of explanatory power than the first regression model. Perceived behavioral control and positive expectations had a statistically significant positive effect, negative expectations had a statistically significant negative effect, and among the 5 factor model, openness had a statistically significant positive (+) Affected. From these results, it can be seen that the Big Five personality factors have a mediating effect on the relationship between goal-oriented behavior model and intention to start up. This study confirmed that the goal-oriented behavioral model of college students is an important variable in implementing self-determination and intention to start a business. In addition, by using his Big 5 personality factors as positive feedback, he has proved to play an important role by identifying the mediation role that can be set, planned and utilized to plan and achieve his life. The result of this study is that college students are interested in the intention of individual start-ups, so they are not freed from difficult employment difficulties. It is intended to provide basic data useful in the age of creation of government.

• Journal of Korea Water Resources Association
• /
• v.54 no.11
• /
• pp.955-968
• /
• 2021

A water resource plan is routinely based on a natural flow and can be estimated using observed streamflow data or a long-term continuous rainfall-runoff model. However, the watershed with the natural flow is very limited to the upstream area of the dam. In particular, for the ungauged watershed, a rainfall-runoff model is established for the gauged watershed, and the model is then applied to the ungauged watershed by transferring the associated parameters. In this study, the GR4J rainfall-runoff model is mainly used to regionalize the parameters that are estimated from the 14 dam watershed via an optimization process. In terms of optimizing the parameters, the Bayesian approach was applied to consider the uncertainty of parameters quantitatively, and a number of parameter samples obtained from the posterior distribution were used for the regionalization. Here, the relationship between the estimated parameters and the topographical factors was first identified, and the dependencies between them are effectively modeled by a Copula function approach to obtain the regionalized parameters. The predicted streamflow with the use of regionalized parameters showed a good agreement with that of the observed with a correlation of about 0.8. It was found that the proposed regionalized framework is able to effectively simulate streamflow for the ungauged watersheds by the use of the regionalized parameters, along with the associated uncertainty, informed by the basin characteristics.