• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.11a
• /
• pp.749-750
• /
• 2020

링크 예측은 주어진 그래프에서 가까운 미래에 발생할 가능성이 높은 새로운 링크를 예측하는 문제이다. 본 논문에서는 유사도 기반 링크 예측의 정확도를 향상시키는 전처리 기법을 제안한다. 제안하는 기법은 유사도 기반으로 예측한 링크들을 그래프에 추가하고, 이 추가된 링크들을 포함하는 그래프를 기반으로 다시 새로운 링크들을 예측하여 추가하는 점진적 추가 방식을 채택한다. 실세계 데이터를 이용한 실험을 통하여, 제안하는 전처리 기법이 기존 링크 예측의 정확도를 향상시킬 수 있는 것을 확인하였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.11a
• /
• pp.759-760
• /
• 2020

최근 사용자, 상품, 그리고 상품의 메타 정보 사이의 관계를 표현한 지식 그래프 (knowledge graph) 가 추천 시스템 분야에서 많은 관심을 받고 있으며 활발히 이용되고 있다. 하지만 기존의 지식 그래프는 각 노드 (사용자, 상품, 메타 정보 등) 사이의 단순한 사실 관계만을 표현하고 있으며, 이는 사용자의 선호도를 정확히 파악하는 데 한계가 있다. 본 논문에서는 지식 그래프의 정보 부족 문제를 보완하기 위해 각 상품에 남겨진 텍스트 리뷰를 감정 분석 (sentiment analysis) 하고, 이를 각 노드 간의 선호도 정보로 활용하여 지식 그래프를 구축하는 방법을 제안한다.

• Proceedings of the Korean Institute of Building Construction Conference
• /
• 2020.11a
• /
• pp.151-152
• /
• 2020

When concrete is exposed to fire, the thermal decomposition of hydrates of Portland cement paste results in critical damage to the concrete structure of a building. Recently, nanosilica arose as the effective nano-additive which can enhance the thermal resistance of the cementitious materials. However, the mechanism of the enhancement was not elucidated specifically. In this study, we investigated the properties of calcium silicate hydrates(C-S-H)of the nanosilica incorporated cement paste after heating to different heating temperatures (200℃, 500℃, and 800℃) by 29Si nuclear magnetic resonance. The results showed that the polymerization of C-S-H of nanosilica incorporated samples was larger than ordinary cement paste after heating to 200℃, and C-S-H formed during heating process to 500℃ due to the pozzolanic reaction during heating process.

• Proceedings of the Korean Institute of Building Construction Conference
• /
• 2021.05a
• /
• pp.72-73
• /
• 2021

For the development of high-fluidity concrete using low-binder, The effect of the use of the developed acrylic viscosity agent on the physical properties of concrete evaluated. The amount acrylic viscosity agent used was 1.5%, 1.7%, and 2.0% based on the binder amount of 400kg/m3, and slump flow test, slump flow 500mm arrival time measurement, air volume measurement, and U-Box passing test were conducted to determine the effect of the physical properties of concrete. it was judged that 1.5% of the acrylic viscosity agent used in high-fluidity concrete using low-binder was most suitable.

• Proceedings of the Korean Institute of Building Construction Conference
• /
• 2020.11a
• /
• pp.81-82
• /
• 2020

Cement paste is the basic material constituting concrete and the most basic data of workability. This study quantitatively evaluated the measured data using a ring flow and a rotary viscometer to estimate the flow of cement paste, and also evaluated the flow and rheology over time. For this, the current work has studied admixtures that affect the fluidity of cement paste. As a result of the experiment, since fluidity and plastic viscosity are inversely proportional to each other, more experimental studies are needed to obtain high fluidity and high viscosity at the same time.

• Proceedings of the Korean Institute of Building Construction Conference
• /
• 2021.05a
• /
• pp.10-11
• /
• 2021

As concrete is used in many construction works, the use of aggregates is increasing. However, supply and demand of high-quality aggregates has become difficult recently, and although circular aggregates that recycle construction waste are used, the performance of concrete, such as liquidity and strength, are being reduced due to defective aggregates. As a result, quality tests such as sieve analysis test are conducted, but a lot of waste occurs such as time and manpower. To solve this problem, this study was conducted to measure the assembly rate of fine aggregate, which accounts for about 35% of the concrete volume, using Deep Learning.

• Proceedings of the Korean Institute of Building Construction Conference
• /
• 2020.11a
• /
• pp.27-28
• /
• 2020

In this paper, we investigated the performance of synthetic graphene oxide-functionalized carbon nanotubes (GF) to promote cement hydration and increase the mechanical properties of cement paste. The enhancement effect of GF on the various properties of cement paste was evaluated via the mechanical strength, X-ray diffraction, and heat of hydration of cement paste. The results clearly showed that GF incorporation into cement paste promotes the early hydration of cement paste, generates more hydration products, which results in the mechanical improvement of cement paste. The compressive and splitting tensile strength were increased by 32.17% and 17.31%, respectively, compared to ordinary Portland cement at 28 days of hydration.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2020.07a
• /
• pp.389-390
• /
• 2020

배포된 소프트웨어는 고객의 새로운 요구사항을 반영하면서 기능을 추가하거나 수정한다. 수정된 소프트웨어는 문제없이 동작하지만, 내부의 구조는 점점 더 복잡해지면서 유지보수가 어려워지고, 수정 시 버그 발생 확률이 높아지게 된다. 이에 따라 잘못된 코드 구조를 개선하고, 소프트웨어의 품질을 향상시켜야한다. 본 논문에서는 개발된 ERP 근태관리 시스템이 지속적인 요구사항을 반영하면서 발생한 잘못된 코드 구조인 코드 스멜을 탐지하고 분석하는 연구를 진행하였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.11a
• /
• pp.423-425
• /
• 2021

전기 자동차 에너지 시스템을 개발하기 위해서는 열 유체-전자기장 연성해석이 필요하다. 이를 위해 열 유체와 전자기장 각각의 해석에 특화된 패키지들을 사용하고, 이 패키지의 실험 결과를 저장 장치를 통해 주고받는다. 하지만 이는 저장장치 I/O 횟수를 늘려서 전체 계산 과정을 느리게 만들 수 있다. 데이터 교환을 MPI 패키지를 이용하여 DRAM 을 통해 이루어지도록 만들 수 있지만, 이는 계산 결과의 지속성을 악화시킨다. 즉, 열 유체 및 전자기장 연산과정에서 속도와 결과의 지속성은 상충적인 관계 갖는다. 본 연구는 이러한 관계를 실험적으로 분석하고 데이터의 지속성을 완화했을 때 이에 의해 얻을 수 있는 성능 이익을 분석한다.

• Proceedings of the Korean Institute of Building Construction Conference
• /
• 2021.11a
• /
• pp.50-51
• /
• 2021

Recently, as concrete is used in many construction works in Korea, the use of aggregates is also increasing. However, the depletion of aggregate resources is making it difficult to supply and demand high-quality aggregates, and the use of defective aggregates is causing problems such as poor performance such as the liquidity and strength of concrete pouring out in the field. As a result, quality tests such as sieve analysis test is conducted on their own, but this study was conducted to improve time and manpower by using the CNN-based Deep Learning Model for the fineness modulus.