• The Journal of Bigdata
• /
• v.6 no.2
• /
• pp.39-49
• /
• 2021

Recently, large and diverse weather data are being collected by sensors from various sources. Efforts to predict the concentration of fine dust through machine learning are being made everywhere, and this study intends to compare PM10 and PM2.5 prediction models using data from 840 outdoor air meters installed throughout the city. Information can be provided in real time by predicting the concentration of fine dust after 5 minutes, and can be the basis for model development after 10 minutes, 30 minutes, and 1 hour. Data preprocessing was performed, such as noise removal and missing value replacement, and a derived variable that considers temporal and spatial variables was created. The parameters of the model were selected through the response surface method. XGBoost, Random Forest, and Deep Learning (Multilayer Perceptron) are used as predictive models to check the difference between fine dust concentration and predicted values, and to compare the performance between models.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.19 no.5
• /
• pp.911-918
• /
• 2024

This study developed and evaluated the performance of a machine learning-based model for predicting cucumber fruit set using cucumber growth data. In this study, plant height, node number, internode length, stem thickness, leaf length, leaf width, leaf count, and female flower count were used as independent variables, and the fruit set was set as the dependent variable to develop a prediction model. Various machine learning algorithms, including Linear Regression, Random Forest, XGBoost, Support Vector Regression (SVR), and K-Nearest Neighbors (KNN), were applied, and model performance was evaluated based on Mean Squared Error (MSE) and the coefficient of determination (R²). As a result, the Random Forest algorithm demonstrated the best performance, with an MSE of 3.91 and an R² of 0.828, effectively capturing the non-linear relationships in the cucumber growth data. In particular, the Random Forest model showed robustness against outliers and proved to be highly effective in predicting fruit set.

• Journal of Multimedia Information System
• /
• v.6 no.4
• /
• pp.209-216
• /
• 2019

Over the past decade, researchers were able to solve complex medical problems as well as acquire deeper understanding of entire issue due to the availability of machine learning techniques, particularly predictive algorithms and automatic recognition of patterns in medical imaging. In this study, a technique called transfer learning has been utilized to classify Magnetic Resonance (MR) images by a pre-trained Convolutional Neural Network (CNN). Rather than training an entire model from scratch, transfer learning approach uses the CNN model by fine-tuning them, to classify MR images into Alzheimer's disease (AD), mild cognitive impairment (MCI) and normal control (NC). The performance of this method has been evaluated over Alzheimer's Disease Neuroimaging (ADNI) dataset by changing the learning rate of the model. Moreover, in this study, in order to demonstrate the transfer learning approach we utilize different pre-trained deep learning models such as GoogLeNet, VGG-16, AlexNet and ResNet-18, and compare their efficiency to classify AD. The overall classification accuracy resulted by GoogLeNet for training and testing was 99.84% and 98.25% respectively, which was exceptionally more than other models training and testing accuracies.

• Journal of the Korean Society of Radiology
• /
• v.83 no.6
• /
• pp.1208-1218
• /
• 2022

In the medical field, artificial intelligence has been used in various ways with many developments. However, most artificial intelligence technologies are developed so that one model can perform only one task, which is a limitation in designing the complex reading process of doctors with artificial intelligence. Multi-task learning is an optimal way to overcome the limitations of single-task learning methods. Multi-task learning can create a model that is efficient and advantageous for generalization by simultaneously integrating various tasks into one model. This study investigated the concepts, types, and similar concepts as multi-task learning, and examined the status and future possibilities of multi-task learning in the medical research.

• Annual Conference on Human and Language Technology
• /
• 2015.10a
• /
• pp.115-118
• /
• 2015

Neural Machine Translation (NMT) 모델은 단일 신경망 구조만을 사용하는 End-to-end 방식의 기계번역 모델로, 기존의 Statistical Machine Translation (SMT) 모델에 비해서 높은 성능을 보이고, Feature Engineering이 필요 없으며, 번역 모델 및 언어 모델의 역할을 단일 신경망에서 수행하여 디코더의 구조가 간단하다는 장점이 있다. 그러나 NMT 모델은 출력 언어 사전(Target Vocabulary)의 크기에 비례해서 학습 및 디코딩의 속도가 느려지기 때문에 출력 언어 사전의 크기에 제한을 갖는다는 단점이 있다. 본 논문에서는 NMT 모델의 출력 언어 사전의 크기 제한 문제를 해결하기 위해서, 입력 언어는 단어 단위로 읽고(Encoding) 출력 언어를 문자(Character) 단위로 생성(Decoding)하는 방법을 제안한다. 출력 언어를 문자 단위로 생성하게 되면 NMT 모델의 출력 언어 사전에 모든 문자를 포함할 수 있게 되어 출력 언어의 Out-of-vocabulary(OOV) 문제가 사라지고 출력 언어의 사전 크기가 줄어들어 학습 및 디코딩 속도가 빨라지게 된다. 실험 결과, 본 논문에서 제안한 방법이 영어-일본어 및 한국어-일본어 기계번역에서 기존의 단어 단위의 NMT 모델보다 우수한 성능을 보였다.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.17 no.5
• /
• pp.173-187
• /
• 2018

This paper proposes a computer vision and deep learning-based technique for surveillance camera system for vehicle counting as one part of parking lot management system. We applied the You Only Look Once version 2 (YOLOv2) detector and come up with a deep convolutional neural network (CNN) based on YOLOv2 with a different architecture and two models. The effectiveness of the proposed architecture is illustrated using a publicly available Udacity's self-driving-car datasets. After training and testing, our proposed architecture with new models is able to obtain 64.30% mean average precision which is a better performance compare to the original architecture (YOLOv2) that achieved only 47.89% mean average precision on the detection of car, truck, and pedestrian.

• Journal of the Korea Convergence Society
• /
• v.12 no.10
• /
• pp.27-35
• /
• 2021

Automatic post editing is a research field that aims to automatically correct errors in machine translation results. This research is mainly being focus on high resource language pairs, such as English-German. Recent APE studies are mainly adopting transfer learning based research, where pre-training language models, or translation models generated through self-supervised learning methodologies are utilized. While translation based APE model shows superior performance in recent researches, as such researches are conducted on the high resource languages, the same perspective cannot be directly applied to the low resource languages. In this work, we apply two transfer learning strategies to Korean-English APE studies and show that transfer learning with translation model can significantly improves APE performance.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.21 no.9
• /
• pp.1744-1751
• /
• 2017

We are striving to strengthen environmental regulations and reduce household waste in all countries around the world. Korea is also striving for the circulation of energy resources by enacting laws to promote resource saving and recycling. The government has implemented an empty bottle deposit system for the recycling of empty bottles, but there is a limit to the collection through manpower and the reverse vending machine is not localized. In this paper, we propose a recyclable bottle recognition and classification algorithm which is essential in the reverser vending machine to promote energy resource circulation. The proposed algorithm is a complex identification algorithm using OpenCV and CNN(Convolution Neural Network). In order to evaluate the effectiveness of the proposed algorithm, we implement a classification system that operates in an reverse vending machine, so that it can easily acquire information about bottles and reverse vending machine in various devices.

• Journal of Korea Society of Industrial Information Systems
• /
• v.27 no.2
• /
• pp.71-82
• /
• 2022

In order to replace the sampling tensile test of products produced in the tab welding process, which is one of the automotive battery manufacturing processes, vision inspectors are currently being developed and used. However, the vision inspection has the problem of inspection position error and the cost of improving it. In order to solve these problems, there are recent cases of applying deep learning technology. As one such case, this paper tries to examine the usefulness of applying Faster R-CNN, one of the deep learning technologies, to existing product inspection. The images acquired through the existing vision inspection machine are used as training data and trained using the Faster R-CNN ResNet101 V1 1024x1024 model. The results of the conventional vision test and Faster R-CNN test are compared and analyzed based on the test standards of 0% non-detection and 10% over-detection. The non-detection rate is 34.5% in the conventional vision test and 0% in the Faster R-CNN test. The over-detection rate is 100% in the conventional vision test and 6.9% in Faster R-CNN. From these results, it is confirmed that deep learning technology is very useful for detecting welding error of lead tabs in automobile batteries.

• Journal of Internet Computing and Services
• /
• v.23 no.4
• /
• pp.87-94
• /
• 2022

Recently, with the development of information and communication infrastructure, the number of Internet access devices is rapidly increasing. Smartphones, laptops, computers, and even IoT devices are receiving information and communication services through Internet access. Since most of the device operating environment consists of web (WEB), it is vulnerable to web cyber attacks using web shells. When the web shell is uploaded to the web server, it is confirmed that the attack frequency is high because the control of the web server can be easily performed. As the damage caused by the web shell occurs a lot, each company is responding to attacks with various security devices such as intrusion prevention systems, firewalls, and web firewalls. In this case, it is difficult to detect, and in order to prevent and cope with web shell attacks due to these characteristics, it is difficult to respond only with the existing system and security software. Therefore, it is an automated defense system through the collection and analysis of web shells based on artificial intelligence machine learning that can cope with new cyber attacks such as detecting unknown web shells in advance by using artificial intelligence machine learning and deep learning techniques in existing security software. We would like to propose about. The machine learning-based web shell defense system model proposed in this paper quickly collects, analyzes, and detects malicious web shells, one of the cyberattacks on the web environment. I think it will be very helpful in designing and building a security system.