• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.80-85
• /
• 2023

기계 번역에서 품질 검증은 정답 문장 없이 기계 번역 시스템에서 생성된 번역의 품질을 자동으로 추정하는 것을 목표로 한다. 일반적으로 이 작업은 상용화된 기계 번역 시스템에서 후처리 모듈 역할을 하여 사용자에게 잠재적인 번역 오류를 경고한다. 품질 검증의 하위 작업인 치명적인 오류 탐지는 번역의 오류 중에서도 정치, 경제, 사회적으로 문제를 일으킬 수 있을 만큼 심각한 오류를 찾는 것을 목표로 한다. 본 논문은 치명적인 오류의 유무를 분류하는 것을 넘어 문장에서 치명적인 오류가 존재하는 부분을 제시하기 위한 새로운 데이터셋과 모델을 제안한다. 이 데이터셋은 거대 언어 모델을 활용하는 구축 방식을 채택하여 오류의 구체적인 범위를 표시한다. 또한, 우리는 우리의 데이터를 효과적으로 활용할 수 있는 다중 작업 학습 모델을 제시하여 오류 범위 탐지에서 뛰어난 성능을 입증한다. 추가적으로 언어 모델을 활용하여 번역 오류를 삽입하는 데이터 증강 방법을 통해 보다 향상된 성능을 제시한다. 우리의 연구는 기계 번역의 품질을 향상시키고 치명적인 오류를 줄이는 실질적인 해결책을 제공할 것이다.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.47 no.4
• /
• pp.105-111
• /
• 2010

This paper proposes the recursive algorithm for passive proximity detection system based on signal strength. The system is designed to be used in the smart medicine chest in order to provide location-based service for the senior personnel. Due to the system profile, single receiver and uni-direction communication are applied over the signal attenuation model for the determination of user existence within certain proximity. The performance of conventional methods is subjective to the sight between the transmitter and receiver unless the direction of target is known. To appreciate the temporal and spatial locality of human subjects, the authors present exponential moving average (EMA) to compensate the unexpected position error from the direction and/or environment. By using optimal parameter, the experiments with EMA algorithm demonstrates 32.26% (maximum 40.80%) reduction in average of the error probability with 50% of consecutive sight in time.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2003.05a
• /
• pp.514-517
• /
• 2003

Oxygen supply is one of the most basic things in human body. Breathing is controlled by the lungs' stationary function and the respiratory center which is in the mesulla oblongata. Nothing but, the external breathing that air movement between the lungs and atmosphere and the internal breathing that cellular air movement between the hemoglobin and a single cell. The adult's number of times of the respirations is about 15∼20 per 1 minute, but it depends ages, exercise, temperature, disease, etc. The important thing in detecting the respiration is that doing it in object person's resting time. Detecting the respiration have to be done without attracting any attention of object person. In present using method is detecting the pulse with catching an object person's wrist and observing the object person's movement. In this paper, we propose the mobile respiration detection diagnostic system using ultrasound sensing method that does not be influenced by the inertia error and the pressure error.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.27 no.10
• /
• pp.68-77
• /
• 1990

In this paper, a neural patten recognition method for the automatic circuit diagram reading system is proposed. The proposed procedure to recognize a deformed logic symbols is composed of three stages: feature detection, log mapping, and pattern classification. In the feature detection stage, a modified competitive learning algorithm where each pattern has the inhibition weight as well as the activation weight is developed. The global information of hand-written logic symbols is obtained by the feature detection neural network having both the inhibition and activation weights. The obtained global data is then transformed into a log space by the conformal mapping where according to the Schwartz's theory about the human visual signal process-ing, the degree of rotation and the scale change are mapped into the translation change. Logic symbols are finally classified by a three layer perceptron trained by the error back propagation algorithm. The computer simulation demonstrates that the proposed multistage neural network system can recognize well the deformed patterns of hand-written logic circuit diagrams.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.10
• /
• pp.35-42
• /
• 2020

In this study, we propose a crack detection method using limited data with a U-Net based image inpainting technique that is a modified unsupervised anomaly detection method. Concrete cracking occurs due to a variety of causes and is a factor that can cause serious damage to the structure in the long term. In general, crack investigation uses an inspector's visual inspection on the concrete surfaces, which is less objective in judgment and has a high possibility of human error. Therefore, a method with objective and accurate image analysis processing is required. In recent years, the methods using deep learning have been studied to detect cracks quickly and accurately. However, when the amount of crack data on the building or infrastructure to be inspected is small, existing crack detection models using it often show a limited performance. Therefore, in this study, an unsupervised anomaly detection method was used to augment the data on the object to be inspected, and as a result of learning using the data, we confirmed the performance of 98.78% of accuracy and 82.67% of harmonic average (F1_Score).

• YAKHAK HOEJI
• /
• v.42 no.5
• /
• pp.473-479
• /
• 1998

Lovastatin (LOVA), a fungal metabolite isolated from cultures of Aspergillus terreus, is a competitive HMG-CoA reductase inhibitor used for the treatment of primary hyper cholesterolemia, and has also been shown to suppress growth in a variety of non-glioma tumor cell lines. A sensitive reversed-phase high-perfonnance liquid chromatographic method with ultraviolet (UV) absorbance detection has been developed to quantitate LOVA in human plasma and urine samples using liquid-liquid extraction procedure. Baseline separation of LOVA and internal standard, simvastatin was achieved on a Novapak $C_{18}$ analytical column with a mobile phase containing 0.025M $NaH_2PO_4$: CAN (35:65, v/v%), adjusted pH to 4.5. The flow rate was set at 1.5ml/min, and the column effluent was monitored by a UV detection at 238nm. The limit of quantification was determined to be 0.5${\mu}$g/ml while extraction efficiency of LOVA ranged from 73.4-82.9% at LOVA concentrations of 0.5 to 10${\mu}$g/ml. Good linearity with correlation coefficients greater than 0.999 was obtained in the range of LOVA concentrations from 0.5 to 10${\mu}$g/ml. The accuracy and the precision were proven excellent with relative standard deviation (RSD, %) and relative error (RE, %) of less than 4.2 and 4.0, respectively. Intraday precision, evaluated at five LOVA concentrations (0.5, 1, 2, 5, 10${\mu}$g/ml) and expressed as RSD ranged from 0-1.82% while the interday precision at the same concentrations ranged from 0.7-10.5%. The analytical method described was then successfully employed for the determination of LOVA concentrations in plasma samples obtained during a phase II clinical trial using high doses of LOVA (30-40mg/kg/day). This method could be further utilized for the ongoing pharmacolkinetic studies and therapeutic drug monitoring of the high-dose LOVA therapy in adenocarcinoma patients.

• The Journal of the Korea Contents Association
• /
• v.11 no.7
• /
• pp.234-238
• /
• 2011

This paper will evaluate the usefulness of 3D target of CBCT by comparing human body's posture and position when simulated treatment is being carried out as well as human phantom posture and position using CBCT which is applying OBI. From the Rando Phantom which is located in the datum point moved in parallel and rotationary direction using CBCT. Then the mean and standard deviation difference on images location difference that are acquired were compared with real the Rando Phantom' moved distance. To make a plan of simulated treatment with the same procedure of real radiation therapy, we are going to setup the Rando Phantom. With an assumption that the position is set in accurate place, we measured the setup errors accroding to the change of the translation and rotation. Tests are repeated 10 times to get the standard deviation of the error values. The variability in couch shift after positioning equivalent to average residual error showed lateral $0.2{\pm}0.2$mm, longitudinal $0.4{\pm}0.3$mm, vertical $-0.4{\pm}0.1$mm. The average rotation erroes target localization after simulated $0.4{\pm}0.2$ mm, $0.3{\pm}0.3$ mm, and $0.3{\pm}0.4$ mm. The detection error by rotation is $0{\sim}0.6^{\circ}$ CBCT 3D/3D matching using the Rando Phantom minimized the errors by realizing accurate matching during simulated treatment and patient caring.

• Journal of Advanced Navigation Technology
• /
• v.23 no.6
• /
• pp.584-588
• /
• 2019

Wireless body area network (WBAN) is a networking technology that enables early detection of abnormal health conditions, real-time medical monitoring, and telemedicine support systems. The internet of things (IoT) for healthcare, which has become an issue recently, is one of the most promising areas for improving the quality of human life. It must meet the high QoS requirements of the medical communication system like any other communication system. Therefore, the bit error rate (BER) threshold was chosen to accommodate the QoS requirements of the WBAN communication system. In this paper, we calculated BER performance of WBAN channel using IR-UWB PPM modulation and analyzed link budget and system margin of WBAN according to various system parameters.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.5 no.4
• /
• pp.39-49
• /
• 2009

Nowadays, according to rapid development of computing environments, information processing and analysis system are very interesting research area. As a viewpoint of data preparation-processing-analysis in knowledge technology, the goal of automated information system is to satisfy high reliability and confidence and to minimize of human-administrator intervention. In addition, we expect the system which can deal with problem and abnormal error effectively as a fault detection and fault tolerance. In this paper, we design a monitoring system as follows. A productive monitoring information from various systems has unstructured forms and characteristics and crawls informative data by conditions and gathering rules. For representing of monitering information which requested by administrator, running-status can be able to check dynamically and systematic like connection/closed status in real-time. Our proposed system can easily correct and processing for monitoring information from various type of server and support to make objective judgement and analysis of administrator under operative target of information system. We implement semi-realtime monitering system using AJAX technology for dynamic browsing of web information and information processing using XML and XPATH. We apply our system to SMS server for checking running status and the system shows that has high utility and reliability.

• Korean Journal of Construction Engineering and Management
• /
• v.25 no.2
• /
• pp.69-80
• /
• 2024

Currently rebar spacing inspection is carried out by human inspectors who heavily rely on their individual experience, lacking a guarantee of objectivity and accuracy in the inspection process. In addition, if incorrectly placed rebars are identified, the inspector need to correct them. Recently, laser scanning and AR technologies have been widely used because of their merits of measurement accuracy and visualization. This study proposes a technology for rebar spacing inspection and fixing by combining laser scanning and AR technology. First, scan data acquisition of rebar layers is performed and the raw scan data is processed. Second, AR-based visualization and fixing are performed by comparing the design model with the model generated from the scan data. To verify the developed technique, performance comparison test is conducted by comparing with existing drawing-based method in terms of inspection time, error detection rate, cognitive load, and situational awareness ability. It is found from the result of the experiment that the AR-based rebar inspection and fixing technology is faster than the drawing-based method, but there was no significant difference between the two groups in error identification rate, cognitive load, and situational awareness ability. Based on the experimental results, the proposed AR-based rebar spacing inspection and fixing technology is expected to be highly useful throughout the construction industry.