• Journal of Korea Multimedia Society
• /
• v.23 no.2
• /
• pp.135-145
• /
• 2020

Recently, deep-learning based methods for low-light image enhancement accomplish great success through supervised learning. However, they still suffer from the lack of sufficient training data due to difficulty of obtaining a large amount of low-/normal-light image pairs in real environments. In this paper, we propose an unsupervised learning approach for single low-light image enhancement using the bright channel prior (BCP), which gives the constraint that the brightest pixel in a small patch is likely to be close to 1. With this prior, pseudo ground-truth is first generated to establish an unsupervised loss function. The proposed enhancement network is then trained using the proposed unsupervised loss function. To the best of our knowledge, this is the first attempt that performs a low-light image enhancement through unsupervised learning. In addition, we introduce a self-attention map for preserving image details and naturalness in the enhanced result. We validate the proposed method on various public datasets, demonstrating that our method achieves competitive performance over state-of-the-arts.

• The Korean Journal of Applied Statistics
• /
• v.29 no.6
• /
• pp.1007-1023
• /
• 2016

This paper explores the conditions and potential of newly designed and tried methodology of big data analysis that apply to Korean history subject matter. In order to advance them, we need to pay more attention to quantitative analysis methodologies over pre-existing qualitative analysis. To obtain our new challenge, I propose 'digital history' methods along with associated disciplines such as linguistics and computer science, data science and statistics, and visualization techniques. As one example, I apply interdisciplinary convergence approaches to the principle and mechanism of elite reproduction during the Korean medieval age. I propose how to compensate for a lack of historical material by applying a semi-supervised learning method, how to create a database that utilizes text-mining techniques, how to analyze quantitative data with statistical methods, and how to indicate analytical outcomes with intuitive visualization.

• Journal of KIISE
• /
• v.44 no.7
• /
• pp.686-691
• /
• 2017

Telemarketing has become the center of marketing action of the industry in the information society. Recently, machine learning has emerged in many areas, especially, financial prediction. Financial data consists of lots of unlabeled data in most parts, and therefore, it is difficult for humans to perform their labeling. In this paper, we propose a fusion method of semi-supervised learning for automatic labeling of unlabeled data to predict telemarketing. Specifically, we integrate labeling results of label propagation and co-training with a decision tree. The data with lower reliabilities are removed, and the data are extracted that have consistent label from two labeling methods. After adding them to the training set, a decision tree is learned with all of them. To confirm the usefulness of the proposed method, we conduct the experiments with a real telemarketing dataset in a Portugal bank. Accuracy of the proposed method is 83.39%, which is 1.82% higher than that of the conventional method, and precision of the proposed method is 19.37%, which is 2.67% higher than that of the conventional method. As a result, we have shown that the proposed method has a better performance as assessed by the t-test.

• Journal of Internet Computing and Services
• /
• v.22 no.1
• /
• pp.13-22
• /
• 2021

Recently network based attack technologies are rapidly advanced and intelligent, the limitations of existing signature-based intrusion detection systems are becoming clear. The reason is that signature-based detection methods lack generalization capabilities for new attacks such as APT attacks. To solve these problems, research on machine learning-based intrusion detection systems is being actively conducted. However, in the actual network environment, attack samples are collected very little compared to normal samples, resulting in class imbalance problems. When a supervised learning-based anomaly detection model is trained with such data, the result is biased to the normal sample. In this paper, we propose to overcome this imbalance problem through One-Class Anomaly Detection using an auto encoder. The experiment was conducted through the NSL-KDD data set and compares the performance with the supervised learning models for the performance evaluation of the proposed method.

• International Journal of Computer Science & Network Security
• /
• v.22 no.2
• /
• pp.272-282
• /
• 2022

Between 2014 and 2019, the US lost more than 2.1 billion USD to phishing attacks, according to the FBI's Internet Crime Complaint Center, and COVID-19 scam complaints totaled more than 1,200. Phishing attacks reflect these awful effects. Phishing websites (PWs) detection appear in the literature. Previous methods included maintaining a centralized blacklist that is manually updated, but newly created pseudonyms cannot be detected. Several recent studies utilized supervised machine learning (SML) algorithms and schemes to manipulate the PWs detection problem. URL extraction-based algorithms and schemes. These studies demonstrate that some classification algorithms are more effective on different data sets. However, for the phishing site detection problem, no widely known classifier has been developed. This study is aimed at identifying the features and schemes of SML that work best in the face of PWs across all publicly available phishing data sets. The Scikit Learn library has eight widely used classification algorithms configured for assessment on the public phishing datasets. Eight was tested. Later, classification algorithms were used to measure accuracy on three different datasets for statistically significant differences, along with the Welch t-test. Assemblies and neural networks outclass classical algorithms in this study. On three publicly accessible phishing datasets, eight traditional SML algorithms were evaluated, and the results were calculated in terms of classification accuracy and classifier ranking as shown in tables 4 and 8. Eventually, on severely unbalanced datasets, classifiers that obtained higher than 99.0 percent classification accuracy. Finally, the results show that this could also be adapted and outperforms conventional techniques with good precision.

• Korean Journal of Cognitive Science
• /
• v.26 no.2
• /
• pp.147-165
• /
• 2015

Through short-answer questions, we can reflect the depth of students' understanding and higher-order thinking skills. Scoring for short-answer questions may take long time and may be an issue on consistency of grading. To alleviate such the suffering, automated scoring systems are widely used in Europe and America, but are in the initial stage in research in Korea. In this paper, we propose a semi-automatic scoring system for short Korean free-text responses using semi-supervised learning. First of all, based on the similarity score between students' answers and model answers, the proposed system grades students' answers and the scored answers with high reliability have been included in the model answers through the thorough test. This process repeats until all answers are scored. The proposed system is used experimentally in Korean and social studies in Nationwide Scholastic Achievement Test. We have confirmed that the processing time and the consistency of grades are promisingly improved. Using the system, various assessment methods have got to be developed and comparative studies need to be performed before applying to school fields.

• Smart Media Journal
• /
• v.8 no.2
• /
• pp.46-57
• /
• 2019

In recent research, deep learning-based methods have achieved state-of-the-art performance in various computer vision tasks. However, these methods are commonly supervised, and require huge amounts of annotated data to train. Acquisition of data demands an additional costly effort, particularly for the tasks where it becomes challenging to obtain large amounts of data considering the time constraints and the requirement of professional human diligence. In this paper, we present a data level synthetic sampling solution to learn from small and imbalanced data sets using Generative Adversarial Networks (GANs). The reason for using GANs are the challenges posed in various fields to manage with the small datasets and fluctuating amounts of samples per class. As a result, we present an approach that can improve learning with respect to data distributions, reducing the partiality introduced by class imbalance and hence shifting the classification decision boundary towards more accurate results. Our novel method is demonstrated on a small dataset of 2789 tomato plant disease images, highly corrupted with class imbalance in 9 disease categories. Moreover, we evaluate our results in terms of different metrics and compare the quality of these results for distinct classes.

• Steel and Composite Structures
• /
• v.52 no.3
• /
• pp.293-312
• /
• 2024

Researchers are actively investigating the potential for utilizing alternative materials in construction to tackle the environmental and economic challenges linked to traditional concrete-based materials. Nevertheless, conventional laboratory methods for testing the mechanical properties of concrete are both costly and time-consuming. The limitations of traditional models in predicting the tensile strength of concrete composited with geopolymer have created a demand for more advanced models. Fortunately, the increasing availability of data has facilitated the use of machine learning methods, which offer powerful and cost-effective models. This paper aims to explore the potential of several machine learning methods in predicting the tensile strength of geopolymer concrete under different curing conditions. The study utilizes a dataset of 221 tensile strength test results for geopolymer concrete with varying mix ratios and curing conditions. The effectiveness of the machine learning models is evaluated using additional unseen datasets. Based on the values of loss functions and evaluation metrics, the results indicate that most models have the potential to estimate the tensile strength of geopolymer concrete satisfactorily. However, the Takagi Sugeno fuzzy model (TSF) and gene expression programming (GEP) models demonstrate the highest robustness. Both the laboratory tests and machine learning outcomes indicate that geopolymer concrete composed of 50% fly ash and 40% ground granulated blast slag, mixed with 10 mol of NaOH, and cured in an oven at 190°F for 28 days has superior tensile strength.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.38 no.1
• /
• pp.1-6
• /
• 2001

In this paper we reanalyze Kohonen‘s learning vector quantizing (LVQ) Learning rule which is based on Hcbb’s learning rule with a view to a gradient descent method. Kohonen's LVQ can be classified into two algorithms according to 6learning mode: unsupervised LVQ(ULVQ) and supervised LVQ(SLVQ). These two algorithms can be represented as gradient descent methods, if target values of output neurons are generated properly. As a result, we see that the LVQ learning method is a special case of a gradient descent method and also that LVQ is represented by a generalized percetron-like LVQ(PLVQ).

• Smart Structures and Systems
• /
• v.30 no.6
• /
• pp.557-569
• /
• 2022

Vibration-based structural health monitoring (SHM) is crucial for the dynamic maintenance of civil building structures to protect property security and the lives of the public. Analyzing these vibrations with modern artificial intelligence and deep learning (DL) methods is a new trend. This paper proposed an unsupervised deep learning method based on a convolutional autoencoder (CAE), which can overcome the limitations of conventional supervised deep learning. With the convolutional core applied to the DL network, the method can extract features self-adaptively and efficiently. The effectiveness of the method in detecting damage is then tested using a benchmark model. Thereafter, this method is used to detect damage and instant disaster events in a rubber bearing-isolated gymnasium structure. The results indicate that the method enables the CAE network to learn the intact vibrations, so as to distinguish between different damage states of the benchmark model, and the outcome meets the high-dimensional data distribution characteristics visualized by the t-SNE method. Besides, the CAE-based network trained with daily vibrations of the isolating layer in the gymnasium can precisely recover newly collected vibration and detect the occurrence of the ground motion. The proposed method is effective at identifying nonlinear variations in the dynamic responses and has the potential to be used for structural condition assessment and safety warning.