• Journal of the Korea Convergence Society
• /
• v.11 no.7
• /
• pp.9-18
• /
• 2020

The research objectives investigates main research topics and trends in the information and communication technology(ICT) field, Korea using LDA(Latent Dirichlet Allocation), one of the topic modeling techniques. The experimental dataset of ICT research and development(R&D) project of 5,200 was acquired through matching with the EZone system of IITP after downloading R&D project dataset from NTIS(National Science and Technology Information Service) during recent five years. Consequently, our finding was that the majority research topics were found as intelligent information technologies such as AI, big data, and IoT, and the main research trends was hyper realistic media. Finally, it is expected that the research results of topic modeling on the national R&D foundation dataset become the powerful information about establishment of planning and strategy of future's research and development in the ICT field.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.32 no.2
• /
• pp.201-211
• /
• 2022

According to the recent change in the cybersecurity paradigm, research on anomaly detection methods using machine learning and deep learning techniques, which are AI implementation technologies, is increasing. In this study, a comparative study on data preprocessing techniques that can improve the anomaly detection performance of a GRU (Gated Recurrent Unit) neural network-based intrusion detection model using NGIDS-DS (Next Generation IDS Dataset), an open dataset, was conducted. In addition, in order to solve the class imbalance problem according to the ratio of normal data and attack data, the detection performance according to the oversampling ratio was compared and analyzed using the oversampling technique applied with DCGAN (Deep Convolutional Generative Adversarial Networks). As a result of the experiment, the method preprocessed using the Doc2Vec algorithm for system call feature and process execution path feature showed good performance, and in the case of oversampling performance, when DCGAN was used, improved detection performance was shown.

• Journal of Intelligence and Information Systems
• /
• v.29 no.2
• /
• pp.241-265
• /
• 2023

A corporate insolvency prediction model serves as a vital tool for objectively monitoring the financial condition of companies. It enables timely warnings, facilitates responsive actions, and supports the formulation of effective management strategies to mitigate bankruptcy risks and enhance performance. Investors and financial institutions utilize default prediction models to minimize financial losses. As the interest in utilizing artificial intelligence (AI) technology for corporate insolvency prediction grows, extensive research has been conducted in this domain. However, there is an increasing demand for explainable AI models in corporate insolvency prediction, emphasizing interpretability and reliability. The SHAP (SHapley Additive exPlanations) technique has gained significant popularity and has demonstrated strong performance in various applications. Nonetheless, it has limitations such as computational cost, processing time, and scalability concerns based on the number of variables. This study introduces a novel approach to variable selection that reduces the number of variables by averaging SHAP values from bootstrapped data subsets instead of using the entire dataset. This technique aims to improve computational efficiency while maintaining excellent predictive performance. To obtain classification results, we aim to train random forest, XGBoost, and C5.0 models using carefully selected variables with high interpretability. The classification accuracy of the ensemble model, generated through soft voting as the goal of high-performance model design, is compared with the individual models. The study leverages data from 1,698 Korean light industrial companies and employs bootstrapping to create distinct data groups. Logistic Regression is employed to calculate SHAP values for each data group, and their averages are computed to derive the final SHAP values. The proposed model enhances interpretability and aims to achieve superior predictive performance.

• International Journal of Computer Science & Network Security
• /
• v.24 no.9
• /
• pp.30-40
• /
• 2024

Building an accurate 3-D spatial road network model has become an active area of research now-a-days that profess to be a new paradigm in developing Smart roads and intelligent transportation system (ITS) which will help the public and private road impresario for better road mobility and eco-routing so that better road traffic, less carbon emission and road safety may be ensured. Dealing with such a large scale 3-D road network data poses challenges in getting accurate elevation information of a road network to better estimate the CO₂ emission and accurate routing for the vehicles in Internet of Vehicle (IoV) scenario. Clustering and regression techniques are found suitable in discovering the missing elevation information in 3-D spatial road network dataset for some points in the road network which is envisaged of helping the public a better eco-routing experience. Further, recently Explainable Artificial Intelligence (xAI) draws attention of the researchers to better interprete, transparent and comprehensible, thus enabling to design efficient choice based models choices depending upon users requirements. The 3-D road network dataset, comprising of spatial attributes (longitude, latitude, altitude) of North Jutland, Denmark, collected from publicly available UCI repositories is preprocessed through feature engineering and scaling to ensure optimal accuracy for clustering and regression tasks. K-Means clustering and regression using Support Vector Machine (SVM) with radial basis function (RBF) kernel are employed for 3-D road network analysis. Silhouette scores and number of clusters are chosen for measuring cluster quality whereas error metric such as MAE ( Mean Absolute Error) and RMSE (Root Mean Square Error) are considered for evaluating the regression method. To have better interpretability of the Clustering and regression models, SHAP (Shapley Additive Explanations), a powerful xAI technique is employed in this research. From extensive experiments , it is observed that SHAP analysis validated the importance of latitude and altitude in predicting longitude, particularly in the four-cluster setup, providing critical insights into model behavior and feature contributions SHAP analysis validated the importance of latitude and altitude in predicting longitude, particularly in the four-cluster setup, providing critical insights into model behavior and feature contributions with an accuracy of 97.22% and strong performance metrics across all classes having MAE of 0.0346, and MSE of 0.0018. On the other hand, the ten-cluster setup, while faster in SHAP analysis, presented challenges in interpretability due to increased clustering complexity. Hence, K-Means clustering with K=4 and SVM hybrid models demonstrated superior performance and interpretability, highlighting the importance of careful cluster selection to balance model complexity and predictive accuracy.

• Journal of KIISE
• /
• v.42 no.2
• /
• pp.220-226
• /
• 2015

Semantic role labeling (SRL) systems determine the semantic role labels of the arguments of predicates in natural language text. An SRL system usually needs to perform four tasks in sequence: Predicate Identification (PI), Predicate Classification (PC), Argument Identification (AI), and Argument Classification (AC). In this paper, we use the Korean Propbank to develop our Korean semantic role labeling system. We describe our Korean semantic role labeling system that uses sequence labeling with structured Support Vector Machine (SVM). The results of our experiments on the Korean Propbank dataset reveal that our method obtains a 97.13% F1 score on Predicate Identification and Classification (PIC), and a 76.96% F1 score on Argument Identification and Classification (AIC).

• Journal of Sensor Science and Technology
• /
• v.29 no.1
• /
• pp.19-26
• /
• 2020

Electrocardiogram (ECG) classification has become an essential task of modern day wearable devices, and can be used to detect cardiovascular diseases. State-of-the-art Artificial Intelligence (AI)-based ECG classifiers have been designed using various artificial neural networks (ANNs). Despite their high accuracy, ANNs require significant computational resources and power. Herein, three different ANNs have been compared: multilayer perceptron (MLP), convolutional neural network (CNN), and spiking neural network (SNN) only for the ECG classification. The ANN model has been developed in Python and Theano, trained on a central processing unit (CPU) platform, and deployed on a PYNQ-Z2 FPGA board to validate the model using a Jupyter notebook. Meanwhile, the hardware accelerator is designed with Overlay, which is a hardware library on PYNQ. For classification, the MIT-BIH dataset obtained from the Physionet library is used. The resulting ANN system can accurately classify four ECG types: normal, atrial premature contraction, left bundle branch block, and premature ventricular contraction. The performance of the ECG classifier models is evaluated based on accuracy and power. Among the three AI algorithms, the SNN requires the lowest power consumption of 0.226 W on-chip, followed by MLP (1.677 W), and CNN (2.266 W). However, the highest accuracy is achieved by the CNN (95%), followed by MLP (76%) and SNN (90%).

• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.10a
• /
• pp.931-934
• /
• 2019

Research using artificial intelligence to generate captions for an image has been studied extensively. However, these systems are unable to create creative stories that include more than one sentence based on image content. A story is a better way that humans use to foster social cooperation and develop social norms. This paper proposes a framework that can generate a relatively short story to describe based on the context of an image. The main contributions of this paper are (1) An unsupervised framework which uses recurrent neural network structure and encoder-decoder model to construct a short story for an image. (2) A huge English novel dataset, including horror and romantic themes that are manually collected and validated. By investigating the short stories, the proposed model proves that it can generate more creative contents compared to existing intelligent systems which can produce only one concise sentence. Therefore, the framework demonstrated in this work will trigger the research of a more robust AI story writer and encourages the application of the proposed model in helping story writer find a new idea.

• Journal of Korea Multimedia Society
• /
• v.24 no.12
• /
• pp.1673-1681
• /
• 2021

Item cold start is a well studied problem in the research field of recommender systems. Still, many existing collaborative filters cannot recommend items accurately when only a few user-item interaction data are available for newly introduced items (Cold items). We propose a interaction feature prediction method to mitigate item cold start problem. The proposed method predicts the interaction features that collaborative filters can calculate for the cold items. For prediction, in addition to content features of the cold-items used by state-of-the-art methods, our method exploits the interaction features of k-nearest content neighbors of the cold-items. An attention network is adopted to extract appropriate information from the interaction features of the neighbors by examining the contents feature similarity between the cold-item and its neighbors. Our evaluation on a real dataset CiteULike shows that the proposed method outperforms state-of-the-art methods 0.027 in Recall@20 metric and 0.023 in NDCG@20 metric.

• Journal of Korea Multimedia Society
• /
• v.25 no.2
• /
• pp.206-214
• /
• 2022

The way to measure the extent of damage from floods and droughts is to identify changes in the extent of water systems. In order to effectively grasp this at a glance, satellite images are used. KOMPSAT-5 uses Synthetic Aperture Radar (SAR) to capture images regardless of weather conditions such as clouds and rain. In this paper, various deep learning models are applied to perform semantic segmentation of the water system in this SAR image and the performance is compared. The models used are U-net, V-Net, U2-Net, UNet 3+, PSPNet, Deeplab-V3, Deeplab-V3+ and PAN. In addition, performance comparison was performed when the data was augmented by applying elastic deformation to the existing SAR image dataset. As a result, without data augmentation, U-Net was the best with IoU of 97.25% and pixel accuracy of 98.53%. In case of data augmentation, Deeplab-V3 showed IoU of 95.15% and V-Net showed the best pixel accuracy of 96.86%.

• Annual Conference on Human and Language Technology
• /
• 2022.10a
• /
• pp.219-224
• /
• 2022

최근 딥러닝이 상식 정보를 추론하지 못하거나, 해석 불가능하다는 한계점을 보완하기 위해 지식 그래프를 기반으로 자연어 텍스트를 생성하는 연구가 중요하게 수행되고 있다. 그러나 이를 위해서 대량의 지식 그래프와 이에 대응되는 문장쌍이 요구되는데, 이를 구축하는 데는 시간과 비용이 많이 소요되는 한계점이 존재한다. 또한 하나의 그래프에 다수의 문장을 생성할 수 있기에 구축자 별로 품질 차이가 발생하게 되고, 데이터 균등성에 문제가 발생하게 된다. 이에 본 논문은 공개된 지식 그래프인 디비피디아를 활용하여 전문가의 도움 없이 자동으로 데이터를 쉽고 빠르게 구축하는 방법론을 제안한다. 이를 기반으로 KoBART와 mBART, mT5와 같은 한국어를 포함한 대용량 언어모델을 활용하여 문장 생성 실험을 진행하였다. 실험 결과 mBART를 활용하여 미세 조정 학습을 진행한 모델이 좋은 성능을 보였고, 자연스러운 문장을 생성하는데 효과적임을 확인하였다.