• 한국정보과학회:학술대회논문집
• /
• 한국정보과학회 2005년도 가을 학술발표논문집 Vol.32 No.2 (2)
• /
• pp.523-525
• /
• 2005

본 논문에서는 가중치가 부여된 나이브 베이지안 분류자와 스팸 메일의 특성을 이용한 주소 유효성 검사를 결합하여 필터링하는 방식의 스팸 메일 필터링 시스템을 제안하였다. 주소 유효성 검사를 통해 스팸 메일을 효율적으로 필터링 할 수 있으며, 나이브 베이지안 분류자에 가중치를 부여함으로써 더욱 효과적인 분류를 할 수 있다. 또한, 각 요인의 중요도에 따라 다른 비중을 부여함으로써 메일의 특성을 고려한 필터링 환경을 구현하였다. 실험에서는 제안하는 요인들이 실제로 필터링 성능 향상에 어떤 영향을 미치는지 살펴보고 최적의 시스템 성능을 측정하였다.

• KDI Journal of Economic Policy
• /
• 제38권2호
• /
• pp.1-20
• /
• 2016

In this paper, we conduct uniform inference of two widely used versions of the Phillips curve, specifically the random-walk Phillips curve and the New-Keynesian Phillips curve (NKPC). For both specifications, we propose a potentially time-varying natural unemployment (NAIRU) to address the uncertainty surrounding the inflation-unemployment trade-off. The inference is conducted through the construction of what is known as the uniform confidence band (UCB). The proposed methodology is then applied to point-ahead inflation forecasting for the Korean economy. This paper finds that the forecasts can benefit from conducting UCB-based inference and that the inference results have important policy implications.

• 한국정보과학회:학술대회논문집
• /
• 한국정보과학회 2004년도 가을 학술발표논문집 Vol.31 No.2 (1)
• /
• pp.343-345
• /
• 2004

높은 효율성과 시스템 자원을 세일하게 제어할 수 있는 편리성을 제공하기 위해서 소프트웨어의 안전성에 대한 책임을 개발자가 지게하는 C 언어의 특성으로 인해서 버퍼 오버플로우, 포맷 스트링 기법 등을 이용한 소프트웨어 공격이 계속 나타나고 있다. 지금까지 알려진 소프트웨어 공격 기법의 다수가 버퍼 오버프로우 기법을 이용한 것이어서 지금까지의 연구는 주로 버퍼 오버플로우 공격 방지 및 탐지에 집중되어 있어 다른 공격 기법에 적용하는 데는 한계가 있었다. 본 논문에서는 소프트웨어 공격의 궁극적인 목적이 제어흐름을 변경시키는 것이라는 것을 바탕으로 프로그램의 제어흐름이 정상적인 범위를 벗어날 경우 이를 공격으로 탐지하는 새로운 기법을 제안하고 기존 연구 결과들과 비교하였다.

• 방송공학회논문지
• /
• 제27권7호
• /
• pp.1011-1020
• /
• 2022

Recent high-performance panoptic segmentation models are based on transformer architectures. However, transformer-based panoptic segmentation methods are basically slower than convolution-based methods, since the attention mechanism in the transformer requires quadratic complexity w.r.t. image resolution. Also, sine and cosine computation for positional embedding in the transformer also yields a bottleneck for computation time. To address these problems, we adopt three modules to speed up the inference runtime of the transformer-based panoptic segmentation. First, we perform channel-level reduction using depth-wise separable convolution for inputs of the transformer decoder. Second, we replace sine and cosine-based positional encoding with convolution operations, called conv-embedding. We also apply a separable self-attention to the transformer encoder to lower quadratic complexity to linear one for numbers of image pixels. As result, the proposed model achieves 44% faster frame per second than baseline on ADE20K panoptic validation dataset, when we use all three modules.

• Communications for Statistical Applications and Methods
• /
• 제30권6호
• /
• pp.551-560
• /
• 2023

Recently, there has been significant research into the recognition of human activities using three-dimensional sequential skeleton data captured by the Kinect depth sensor. Many of these studies employ deep learning models. This study introduces a novel feature selection method for this data and analyzes it using machine learning models. Due to the high-dimensional nature of the original Kinect data, effective feature extraction methods are required to address the classification challenge. In this research, we propose using the first four moments as predictors to represent the distribution of joint sequences and evaluate their effectiveness using two datasets: The exergame dataset, consisting of three activities, and the MSR daily activity dataset, composed of ten activities. The results show that the accuracy of our approach outperforms existing methods on average across different classifiers.

• 한국독성학회:학술대회논문집
• /
• 한국독성학회 2006년도 추계학술대회
• /
• pp.65-74
• /
• 2006

Modem drug discovery requires rapid pharmacokinetic evaluation of chemically diverse compounds for early candidate selection. This demands the development of analytical methods that offer high-throughput of samples. Naturally, liquid chromatography / tandem mass spectrometry (LC-MS/MS) is choice of the analytical method because of its superior sensitivity and selectivity. As a result of the short analysis time(typically 3-5min) by LC-MS/MS, sample preparation has become the rate- determining step in the whole analytical cycle. Consequently tremendous efforts are being made to speed up and automate this step. In a typical automated 96-well SPE(solid-phase extraction) procedure, plasma samples are transferred to the 96-well SPE plate, internal standard and aqueous buffer solutions are added and then vacuum is applied using the robotic liquid handling system. It takes only 20-90 min to process 96 samples by automated SPE and the analyst is physically occupied for only approximately 10 min. Recently, the ultra-high flow rate liquid chromatography (turbulent-flow chromatography)has sparked a huge interest for rapid and direct quantitation of drugs in plasma. There is no sample preparation except for sample aliquotting, internal standard addition and centrifugation. This type of analysis is achieved by using a small diameter column with a large particle size(30-5O ${\mu}$m) and a high flow rate, typically between 3-5 ml/min. Silica-based monolithic HPLC columns contain a novel chromatographic support in which the traditional particulate packing has been replaced with a single, continuous network (monolith) of pcrous silica. The main advantage of such a network is decreased backpressure due to macropores (2 ${\mu}$m) throughout the network. This allows high flow rates, and hence fast analyses that are unattainable with traditional particulate columns. The reduction of particle diameter in HPLC results in increased column efficiency. use of small particles (<2 urn), however, requires p.essu.es beyond the traditional 6,000 psi of conventional pumping devices. Instrumental development in recent years has resulted in pumping devices capable of handling the requirements of columns packed with small particles. The staggered parallel HPLC system consists of four fully independent binary HPLC pumps, a modified auto sampler, and a series of switching and selector valves all controlled by a single computer program. The system improves sample throughput without sacrificing chromatographic separation or data quality. Sample throughput can be increased nearly four-fold without requiring significant changes in current analytical procedures. The process of Bioanalytical Method Validation is required by the FDA to assess and verify the performance of a chronlatographic method prior to its application in sample analysis. The validation should address the selectivity, linearity, accuracy, precision and stability of the method. This presentation will provide all overview of the work required to accomplish a full validation and show how a chromatographic method is suitable for toxirokinetic sample analysis. A liquid chromatography/tandem mass spectrometry (LC-MS/MS) method developed to quantitate drug levels in dog plasma will be used as an example of tile process.

• 한국정보과학회논문지:컴퓨팅의 실제 및 레터
• /
• 제11권1호
• /
• pp.92-99
• /
• 2005

본 논문에서는 SystemC를 이용하여 네트웍 SOC에 적용이 가능한 상위 계층 설계 방법을 제안한다. 본 방식은 실제 양산되고 있는 네트웍 SOC를 기준 플랫폼으로 하여 NAT 라우터에서 보다 높은 변환율을 얻기 위한 최적의 하드웨어 계수 결정을 목표로 한다. 네트웍 SOC에 내장된 고속 이더넷 MAC, 전용 I)MA, 시스템 모듈들은 트랜잭션 레벨에서 SystemC를 이용하여 모델링되었다. 고속 이더넷 제어기 모델은 실제 Verilog RTL의 동작을 사이클 단위로 측정한 결과를 토대로 동작이 세부 조정되었다. SystemC 환경의 NAT 변환율은 기준 플랫폼 검증 보드상의 측정 결과와 비교하여 $\pm$10% 이내의 오차를 보였고, RTL 시뮬레이션보다 100배 이상의 속도 이득을 보였다. 본 모델은 NAT 라우터에서 성능 저하의 원인을 찾는 SOC 구조 탐색을 위해 사용될 수 있다.

• Nuclear Engineering and Technology
• /
• 제56권3호
• /
• pp.880-885
• /
• 2024

For Safety Assisted Engineering works, real-time simulators have emerged as a mandatory tool among all the key actors involved in the nuclear industry (utilities, designers and safety authorities). EDF, Electricité de France, as the leading worldwide nuclear power plant operator, has a crucial need for efficient and updated simulation tools for training, operating and safety analysis support. This paper will present the work performed at EDF/DT to develop a new generation of engineering simulator to fulfil these tasks. The project is called SiRENE, which is the acronym of Re-hosted Engineering Simulator in French. The project has been economically challenging. Therefore, to benefit from existing tools and experience, the SiRENE project combines: - A part of the process issued from the operating fleet training full-scope simulator. - An improvement of the simulator prediction reliability with the integration of High-Fidelity models, used in Safety Analysis. These High-Fidelity models address Nuclear Steam Supply System code, with CATHARE thermal-hydraulics system code and neutronics, with COCCINELLE code. - And taking advantage of the last generation and improvements of instructor station. The intensive and challenging uses of the new SiRENE engineering simulator are also discussed. The SiRENE simulator has to address different topics such as verification and validation of operating procedures, identification of safety paths, tests of I&C developments or modifications, tests on hydraulics system components (pump, valve etc.), support studies for Probabilistic Safety Analysis (PSA). etc. It also emerges that SiRENE simulator is a valuable tool for self-training of the newcomers in EDF nuclear engineering centers. As a modifiable tool and thanks to a skillful team managing the SiRENE project, specific and adapted modifications can be taken into account very quickly, in order to provide the best answers for our users' specific issues. Finally, the SiRENE simulator, and the associated configurations, has been distributed among the different engineering centers at EDF (DT in Lyon, DIPDE in Marseille and CNEPE in Tours). This distribution highlights a strong synergy and complementarity of the different engineering institutes at EDF, working together for a safer and a more profitable operating fleet.

• Journal of Astronomy and Space Sciences
• /
• 제34권2호
• /
• pp.139-151
• /
• 2017

This paper presents an overview of deep space orbit determination software (DSODS), as well as validation and verification results on its event prediction capabilities. DSODS was developed in the MATLAB object-oriented programming environment to support the Korea Pathfinder Lunar Orbiter (KPLO) mission. DSODS has three major capabilities: celestial event prediction for spacecraft, orbit determination with deep space network (DSN) tracking data, and DSN tracking data simulation. To achieve its functionality requirements, DSODS consists of four modules: orbit propagation (OP), event prediction (EP), data simulation (DS), and orbit determination (OD) modules. This paper explains the highest-level data flows between modules in event prediction, orbit determination, and tracking data simulation processes. Furthermore, to address the event prediction capability of DSODS, this paper introduces OP and EP modules. The role of the OP module is to handle time and coordinate system conversions, to propagate spacecraft trajectories, and to handle the ephemerides of spacecraft and celestial bodies. Currently, the OP module utilizes the General Mission Analysis Tool (GMAT) as a third-party software component for high-fidelity deep space propagation, as well as time and coordinate system conversions. The role of the EP module is to predict celestial events, including eclipses, and ground station visibilities, and this paper presents the functionality requirements of the EP module. The validation and verification results show that, for most cases, event prediction errors were less than 10 millisec when compared with flight proven mission analysis tools such as GMAT and Systems Tool Kit (STK). Thus, we conclude that DSODS is capable of predicting events for the KPLO in real mission applications.

• 한국지반공학회논문집
• /
• 제40권3호
• /
• pp.33-39
• /
• 2024

구조물의 상태를 파악하기 위한 균열조사는 정밀안전 진단에 필수적인 검사 항목이다. 그러나 육안으로 이루어지는 균열조사 방식은 현장 상황의 변화에 따라 주관적으로 수행될 수 있다. 이러한 육안검사의 한계를 극복하기 위해 본 연구에서는, ResNet, FPN, Mask R-CNN을 백본(Backbone), 넥(Neck), 헤드(head)로 구성한 합성곱 신경망을 바탕으로, 이미지 데이터에서의 콘크리트 균열 탐지를 자동화하고. 그 성능을 IoU 값을 바탕으로 분석하였다. 해석에 사용된 데이터는 총 1,203개의 이미지 데이터로 구성하였으며, 이 중 70%를 훈련(Training)에, 20%를 검증(Validation)에, 그리고 10%의 데이터를 시험(Testing)에 사용하였다. 시험 결과의 평균 IoU값은 95.83%로 산정되었고, 또한 이미지 내 균열이 전혀 탐지되지 않는 경우는 존재하지 않아, 본 연구에 가정한 모델이 콘크리트의 균열 탐지를 성공적으로 수행하는 것을 확인하였다.