• Journal of KIISE:Computing Practices and Letters
• /
• v.13 no.2
• /
• pp.100-109
• /
• 2007

Object recognition of service robots is very important for most of services such as delivery, and errand. Conventional methods are based on the geometric models in static industrial environments, but they have limitations in indoor environments where the condition is changable and the movement of service robots occur because the interesting object can be occluded or small in the image according to their location. For solving these uncertain situations, in this paper, we propose the method that exploits observed objects as context information for predicting interesting one. For this, we propose the method for modeling domain knowledge in probabilistic frame by adopting Bayesian networks and ontology together, and creating knowledge model dynamically to extend reasoning models. We verify the performance of our method through the experiments and show the merit of inductive reasoning in the probabilistic model

• The KIPS Transactions:PartA
• /
• v.16A no.5
• /
• pp.403-410
• /
• 2009

In this paper, we present an on-line quiz authoring software that helps teachers create an intelligent on-line quiz. It is designed to give each student appropriate diagnostic report using Bayesian inference networks that represent the relationships among knowledge-items. Once the authors design and edit quizzes in quiz authoring page, the authoring tool automatically produces a knowledge-model based on Bayesian inference network, on-line quizzes, and student report pages. It turns out that the on-line quizzes generated by this tool help students identify their weak parts of subject, make learning strategies for the next learning steps and carry out supplementary learning for their weak knowledge-items.

• Journal of Navigation and Port Research
• /
• v.47 no.6
• /
• pp.324-330
• /
• 2023

Aids to navigation, which provide information about a ship's position, direction, and the location of obstacles, are crucial for uninterrupted maritime services. This study aimed to analyze accidents involving aids to navigation that resulted in service disruptions and identify the key factors associated with these accidents. Aids to navigation accident data from 2000 to 2022 were utilized to achieve this. We categorized accidents by accident type, cause, region, season, and type of navigation aid and established a network through correlation analysis. Bayesian networks based on aids to navigation accidents were assigned prior probabilities, and the factors that increased the probability of accidents for different types of aids to navigation were identified. The findings can be used to infer the causes of unreported aids to navigation accidents and serve as foundational data for the prevention of such accidents.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.42 no.1
• /
• pp.11-21
• /
• 2022

Drought occurs extensively over a long period and causes great socio-economic damage. Since drought risk consists of social, environmental, physical, and economic factors along with meteorological and hydrological factors, it is important to quantitatively identify their impacts on drought risk. This study investigated the relationship among drought hazard, vulnerability, response capacity, and risk in Chungcheongbuk-do using a structural equation model and evaluated their impacts on drought risk using Bayesian networks. We also performed sensitivity analysis to investigate how the factors change drought risk. Overall results showed that Chungju-si had the highest risk of drought. The risk was calculated as the largest even when the hazard and response capacity were changed. However, when the vulnerability was changed, Eumseong-gun had the greatest risk. The sensitivity analysis showed that Jeungpyeong-gun had the highest sensitivity, and Jecheon-si, Eumseong-gun, and Okcheon-gun had highest individual sensitivities with hazard, vulnerability, and response capacity, respectively. This study concluded that it is possible to identify impact factors on drought risk using regional characteristics, and to prepare appropriate drought countermeasures considering regional drought risk.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.21 no.5
• /
• pp.515-523
• /
• 2015

This study reviewed 612 DP LOP(Loss of Position) incident reports which submitted to IMCA from 2001~2010 and identified 103 human error caused incidents and classified it through HFACS. And, this study analysis of conditional probability of human error on DP LOP incidents through application of bayesian network. As a result, all 103 human error related DP LOP incidents were caused by unsafe acts, and among unsafe acts 70 incidents(68.0 %) were related to skill based error which are the largest proportion of human error causes. Among skill based error, 60(58.3%) incidents were involved inadvertent use of controls and 8(7.8%) incidents were involved omitted step in procedure. Also, 21(20.8%) incidents were involved improper maneuver because of decision error. Also this study identified that unsafe supervision(68%) is effected as the largest latent causes of unsafe acts through application to bayesian network. As a results, it is identified that combined analysis of HFACS and bayesian network are useful tool for human error analysis. Based on these results, this study suggest 9 recommendations such as polices, interpersonal interaction, training etc. to prevent and mitigate human errors during DP operations.

• Proceedings of the Korean Nuclear Society Conference
• /
• 1998.05a
• /
• pp.264-270
• /
• 1998

This research broadens the prime concern of nuclear power plant operations from safe performance to both economic and safe performance. First emergency diesel generator is identified as one of main contributors for the lost plant availability through the review of plants forced outage records. The framework of an integrated architecture for performing modern on-line condition for operational availability improvement is configured in this work. For the development of the comprehensive sensor networks for complex target systems, an integrated methodology incorporating a structural hierarchy, a functional hierarchy, and a fault-system matrix is formulated. The second part of our research is development of intelligent diagnosis and maintenance advisory system, which employs Bayesian Belief networks (BBNs) as a high level reasoning tool incorporating inherent uncertainty use in probabilistic inference. Our prototype diagnosis algorithms are represented explicitly through topological symbols and links between them in a causal direction. As new evidence from sensor network development is entered into the model especially, our advisory of system provides operational advice concerning both availability and safety, so that the operator is able to determine the likely modes, diagnose the system state, locate root causes, and take the most advantageous action. Thereby, this advice improves operational availability

• Journal of KIISE:Computing Practices and Letters
• /
• v.16 no.10
• /
• pp.995-999
• /
• 2010

Every year, traffic accidents and traffic congestion have been rapidly increasing, Although the roadway design and signal system have been improved to relieve traffic accidents, traffic casualties and property damage do not decrease. This paper develops a real-time traffic accident detection and analysis system (RTADAS): In the proposed system, we aim to precisely detect traffic accidents at different design and flow of intersections, However, because the data collected from intersections have uncertainty and complicated causal dependency between them, we construct probability-based networks for correct accident detection.

• Journal of Information Processing Systems
• /
• v.18 no.2
• /
• pp.188-196
• /
• 2022

Recently, neural architecture search (NAS) has received increasing attention as it can replace human experts in designing the architecture of neural networks for different tasks and has achieved remarkable results in many challenging tasks. In this study, a path-based computation neural architecture encoder (PCE) was proposed. Our PCE first encodes the computation of information on each path in a neural network, and then aggregates the encodings on all paths together through an attention mechanism, simulating the process of information computation along paths in a neural network and encoding the computation on the neural network instead of the structure of the graph, which is more consistent with the computational properties of neural networks. We performed an extensive comparison with eight encoding methods on two commonly used NAS search spaces (NAS-Bench-101 and NAS-Bench-201), which included a comparison of the predictive capabilities of performance predictors and search capabilities based on two search strategies (reinforcement learning-based and Bayesian optimization-based) when equipped with different encoders. Experimental evaluation shows that PCE is an efficient encoding method that effectively ranks and predicts neural architecture performance, thereby improving the search efficiency of neural architectures.

• Geomechanics and Engineering
• /
• v.33 no.6
• /
• pp.611-624
• /
• 2023

Soil properties make it attractive as a building material due to its mechanical strength, aesthetically appearance, plasticity, and low cost. However, it is frequently necessary to improve and stabilize the soil mechanical properties with binders. Soil-cement is applied for purposes ranging from housing to dams, roads and foundations. Unconfined compression strength (UCS) and split tensile strength (CD) are essential mechanical parameters for ascertaining the aptitude of soil-cement for a given application. However, quantifying these parameters requires specimen preparation, testing, and several weeks. Methodologies that allowed accurate estimation of mechanical parameters in shorter time would represent an important advance in order to ensure shorter deliverable timeline and reduce the amount of laboratory work. In this work, an extensive campaign of UCS and CD tests was carried out in a sandy soil from the Leiria region (Portugal). Then, using the machine learning tool Neural Pattern Recognition of the MATLAB software, a prediction of these two parameters based on six input parameters was made. The results, especially those obtained with resource to a Bayesian regularization-backpropagation algorithm, are frankly positive, with a forecast success percentage over 90% and very low root mean square error (RMSE).

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2007.11a
• /
• pp.11-14
• /
• 2007

인간과 컴퓨터간의 상호교류 하는데 있어서 감정 인식은 필수라 하겠다. 그래서 본 논문에서는 음성 신호 및 얼굴 영상을 BL(Bayesian Learning)과 PCA(Principal Component Analysis)에 적용하여 5가지 감정 (Normal, Happy, Sad, Anger, Surprise) 으로 패턴 분류하였다. 그리고 각각 신호의 단점을 보완하고 인식률을 높이기 위해 결정 융합 방법과 특징 융합 방법을 이용하여 감정융합을 실행하였다. 결정 융합 방법은 각각 인식 시스템을 통해 얻어진 인식 결과 값을 퍼지 소속 함수에 적용하여 감정 융합하였으며, 특정 융합 방법은 SFS(Sequential Forward Selection)특정 선택 방법을 통해 우수한 특정들을 선택한 후 MLP(Multi Layer Perceptron) 기반 신경망(Neural Networks)에 적용하여 감정 융합을 실행하였다.