• KIPS Transactions on Software and Data Engineering
• /
• v.9 no.3
• /
• pp.91-100
• /
• 2020

Unlike the existing Visual Question Answering(VQA) problems, the new Visual Commonsense Reasoning(VCR) problems require deep common sense reasoning for answering questions: recognizing specific relationship between two objects in the image, presenting the rationale of the answer. In this paper, we propose a novel deep neural network model, KG_VCR, for VCR problems. In addition to make use of visual relations and contextual information between objects extracted from input data (images, natural language questions, and response lists), the KG_VCR also utilizes commonsense knowledge embedding extracted from an external knowledge base called ConceptNet. Specifically the proposed model employs a Graph Convolutional Neural Network(GCN) module to obtain commonsense knowledge embedding from the retrieved ConceptNet knowledge graph. By conducting a series of experiments with the VCR benchmark dataset, we show that the proposed KG_VCR model outperforms both the state of the art(SOTA) VQA model and the R2C VCR model.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.20 no.1
• /
• pp.70-85
• /
• 2021

Deep learning methodology, which has been actively studied in recent years, has improved the performance of artificial intelligence. Accordingly, systems utilizing deep learning have been proposed in various industries. In traffic systems, spatio-temporal graph modeling using GNN was found to be effective in predicting traffic speed. Still, it has a disadvantage that the model is trained inefficiently due to the memory bottleneck. Therefore, in this study, the road network is clustered through the graph clustering algorithm to reduce memory bottlenecks and simultaneously achieve superior performance. In order to verify the proposed method, the similarity of road speed distribution was measured using Jensen-Shannon divergence based on the analysis result of Incheon UTIC data. Then, the road network was clustered by spectrum clustering based on the measured similarity. As a result of the experiments, it was found that when the road network was divided into seven networks, the memory bottleneck was alleviated while recording the best performance compared to the baselines with MAE of 5.52km/h.

• Journal of Institute of Control, Robotics and Systems
• /
• v.22 no.6
• /
• pp.477-482
• /
• 2016

This paper presents a simple approximation-based design approach for consensus tracking of heterogeneous second-order nonlinear systems under a directed network. All nonlinearities of followers are assumed to be unknown and non-identical. In the controller design procedure, graph-independent error surfaces are used and an unimplementable intermediate controller for each follower is designed at the first design step. Then, by adding and subtracting a graph-based term at the second step, the actual controller for each follower is designed by using one neural network employed to estimate a lumped and distributed nonlinearity. Therefore, the proposed local controller for each follower has a simpler structure than existing approximation-based consensus tracking controllers for multi-agent systems with unmatched nonlinearities.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.8
• /
• pp.319-330
• /
• 2021

Recently, in the field of open domain natural language question answering, multi-task, multi-hop question answering has been studied extensively. In this paper, we propose a novel deep neural network model using hierarchical graphs to answer effectively such multi-task, multi-hop questions. The proposed model extracts different levels of contextual information from multiple paragraphs using hierarchical graphs and graph neural networks, and then utilize them to predict answer type, supporting sentences and answer spans simultaneously. Conducting experiments with the HotpotQA benchmark dataset, we show high performance and positive effects of the proposed model.

• International Journal of Contents
• /
• v.10 no.1
• /
• pp.23-28
• /
• 2014

When developing a classifier using various objective functions, it is important to compare the performances of the classifiers. Although there are statistical analyses of objective functions for classifiers, simulation results can provide us with direct comparison results and in this case, a comparison criterion is considerably critical. A Receiver Operating Characteristics (ROC) graph is a simulation technique for comparing classifiers and selecting a better one based on a performance. In this paper, we adopt the ROC graph to compare classifiers trained by mean-squared error, cross-entropy error, classification figure of merit, and the n-th order extension of cross-entropy error functions. After the training of feed-forward neural networks using the CEDAR database, the ROC graphs are plotted to help us identify which objective function is better.

• Journal of Information Processing Systems
• /
• v.15 no.3
• /
• pp.520-537
• /
• 2019

This paper proposes a system that can detect the data leakage pattern using a convolutional neural network based on defining the behaviors of leaking data. In this case, the leakage detection scenario of data leakage is composed of the patterns of occurrence of security logs by administration and related patterns between the security logs that are analyzed by association relationship analysis. This proposed system then detects whether the data is leaked through the convolutional neural network using an insider malicious behavior graph. Since each graph is drawn according to the leakage detection scenario of a data leakage, the system can identify the criminal insider along with the source of malicious behavior according to the results of the convolutional neural network. The results of the performance experiment using a virtual scenario show that even if a new malicious pattern that has not been previously defined is inputted into the data leakage detection system, it is possible to determine whether the data has been leaked. In addition, as compared with other data leakage detection systems, it can be seen that the proposed system is able to detect data leakage more flexibly.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.11a
• /
• pp.574-576
• /
• 2022

최근 여러 분야에서 그래프 신경망(graph neural network, GNN)이 활발히 연구되고 있다. 하지만 지금까지 대부분의 GNN 연구는 단일 GNN 모델의 성능을 향상하는 데 집중되었다. 본 논문에서는 앙상블(ensemble) 기법의 대표적 기법인 그래디언트 부스팅(gradient boosting)을 이용하여 GNN의 앙상블 모델을 만드는 방법을 제안한다. 제안 방법은 앞서 만들어진 GNN의 오차를 경사 하강법(gradient descent)을 이용하여 감소시키는 방향으로 다음 GNN을 생성한다. 이 과정을 반복하여 GNN의 최종 앙상블 모델을 얻는다. 실험에서 GNN의 대표적인 모델인 그래프 합성곱 신경망(graph convolutional network, GCN)에 제안 방법을 적용하여 앙상블 모델을 생성한 결과, 단일 GCN 모델에 비해 노드 분류 정확도가 11.3%p까지 증가하였음을 확인하였다.

• Steel and Composite Structures
• /
• v.44 no.2
• /
• pp.241-254
• /
• 2022

The cables in a cable-stayed bridge are critical load-carrying parts. The potential damage to cables should be identified early to prevent disasters. In this study, an efficient deep learning model is proposed for the damage identification of cables using both a multi-layer perceptron (MLP) and a graph neural network (GNN). Datasets are first generated using the practical advanced analysis program (PAAP), which is a robust program for modeling and analyzing bridge structures with low computational costs. The model based on the MLP and GNN can capture complex nonlinear correlations between the vibration characteristics in the input data and the cable system damage in the output data. Multiple hidden layers with an activation function are used in the MLP to expand the original input vector of the limited measurement data to obtain a complete output data vector that preserves sufficient information for constructing the graph in the GNN. Using the gated recurrent unit and set2set model, the GNN maps the formed graph feature to the output cable damage through several updating times and provides the damage results to both the classification and regression outputs. The model is fine-tuned with the original input data using Adam optimization for the final objective function. A case study of an actual cable-stayed bridge was considered to evaluate the model performance. The results demonstrate that the proposed model provides high accuracy (over 90%) in classification and satisfactory correlation coefficients (over 0.98) in regression and is a robust approach to obtain effective identification results with a limited quantity of input data.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.7
• /
• pp.89-98
• /
• 2024

In this paper, we propose a system for recommending lecture videos at the chapter level, addressing the balance between accuracy and processing speed in chapter-level video recommendations. Specifically, it has been observed that enhancing recommendation accuracy reduces processing speed, while increasing processing speed decreases accuracy. To mitigate this trade-off, a hybrid approach is proposed, utilizing techniques such as TF-IDF, k-means++ clustering, and Graph Neural Networks (GNN). The approach involves pre-constructing clusters based on chapter similarity to reduce computational load during recommendations, thereby improving processing speed, and applying GNN to the graph of clusters as nodes to enhance recommendation accuracy. Experimental results indicate that the use of GNN resulted in an approximate 19.7% increase in recommendation accuracy, as measured by the Mean Reciprocal Rank (MRR) metric, and an approximate 27.7% increase in precision defined by similarities. These findings are expected to contribute to the development of a learning system that recommends more suitable video chapters in response to learners' queries.

• Proceedings of the IEEK Conference
• /
• 2009.05a
• /
• pp.372-374
• /
• 2009

In this research, an obstacle avoidance method is proposed. The common usage of a robot is indoor and the obstacles to the indoor robot is studied. The accurate detection of direction after overcoming the obstacles is necessary for performance of autonomous navigation and mission project. The sensors such as Laser, Ultrasound, PSD can be used to measure the obstacles. In this research, a PSD sensor is used to detect obstacles. It detects the height and width of obstacles located on the floor. Before measuring the obstacles, a calibration of the sensor was done and it produced a better accuracy. We have plotted an error graph using data obtained from the repeated experiments. The graph is fitted to a polynomial curve. The polynomial equation is used for the robot navigation. And in this research, a model of the error of the direction of the robot after overcoming obstacles was obtained also. The prototype of the obstacle and the error of the direction after overcoming the obstacles are modelled using a neural networks. The input of the neural network composed with the height of the obstacles, the speed of robot, the direction of wheels and the error of the direction. To implement the suggested algorithm, we set up a robot which is operated by a notebook computer. Experiment showed the suggested algorithm performed well.