• Journal of Internet Computing and Services
• /
• v.24 no.5
• /
• pp.17-27
• /
• 2023

Today, as AI (Artificial Intelligence) technology develops and its practicality increases, it is widely used in various application fields in real life. At this time, the AI model is basically learned based on various statistical properties of the learning data and then distributed to the system, but unexpected changes in the data in a rapidly changing data situation cause a decrease in the model's performance. In particular, as it becomes important to find drift signals of deployed models in order to respond to new and unknown attacks that are constantly created in the security field, the need for lifecycle management of the entire model is gradually emerging. In general, it can be detected through performance changes in the model's accuracy and error rate (loss), but there are limitations in the usage environment in that an actual label for the model prediction result is required, and the detection of the point where the actual drift occurs is uncertain. there is. This is because the model's error rate is greatly influenced by various external environmental factors, model selection and parameter settings, and new input data, so it is necessary to precisely determine when actual drift in the data occurs based only on the corresponding value. There are limits to this. Therefore, this paper proposes a method to detect when actual drift occurs through an Anomaly analysis technique based on XAI (eXplainable Artificial Intelligence). As a result of testing a classification model that detects DGA (Domain Generation Algorithm), anomaly scores were extracted through the SHAP(Shapley Additive exPlanations) Value of the data after distribution, and as a result, it was confirmed that efficient drift point detection was possible.

• Journal of Korean Elementary Science Education
• /
• v.39 no.1
• /
• pp.117-132
• /
• 2020

The purpose of this study is to investigate elementary school teachers' awareness of Artificial Intelligence (AI) and find out how to apply it in elementary science education. The survey was conducted online and involved 95 teachers working in the metropolitan area. The results of this study are as follows. First, teachers need to learn about the general characteristics of AI and how to apply it to education. Second, science classes had the highest preference for AI among elementary school subjects. Third, the preference for AI application by elementary science field was 68.4% for earth and space, 54.7% for exercise and energy, 32.6% for matter, 27.4% for life. Fourth, AI-based Science Education (AISE) teaching- learning strategies were developed based on AI characteristics and the changing perspective of elementary science education, AISE's teaching-learning strategies are five: 'automation', 'individualization', 'diversification', 'cooperation' and 'creativity' and teachers can use them in teaching design, class practice and evaluation stages. Finally, the creative problem-solving Doing Thinking Making Sharing (DTMS) model was devised to implement the creativity strategy in AISE. This model consists of four-steps teaching courses: Doing, Thinking, Making and Sharing based on the empirical learning theory. In the future, follow-up research is needed to verify the effectiveness of this model by applying it to elementary science education.

• Journal of the Semiconductor & Display Technology
• /
• v.22 no.3
• /
• pp.84-89
• /
• 2023

In water resource management, data prediction is performed using artificial intelligence, and companies, governments, and institutions continue to attempt to efficiently manage resources through this. LSTM is a model specialized for processing time series data, which can identify data patterns that change over time and has been attempted to predict groundwater level data. However, groundwater level data can cause sen-sor errors, missing values, or outliers, and these problems can degrade the performance of the LSTM model, and there is a need to improve data quality by processing them in the pretreatment stage. Therefore, in pre-dicting groundwater data, we will compare the LSTM model with the MSE and the model after normaliza-tion through distribution, and discuss the important process of analysis and data preprocessing according to the comparison results and changes in the results.

• Journal of Platform Technology
• /
• v.11 no.6
• /
• pp.3-12
• /
• 2023

In today's system operation, it is difficult to detect failures and take immediate action in the case of a shortage of manpower compared to the number of equipment or failures in vulnerable time zones, which can lead to delays in failure recovery. In addition, various algorithms exist to detect abnormal symptom data, and it is important to select an appropriate algorithm for each problem. In this paper, an ensemble-based isolation forest model was used to efficiently detect multivariate point anomalies that deviated from the mean distribution in the data set generated to predict system failure and minimize service interruption. And since significant changes in memory space usage are observed together with changes in CPU usage, the problem is solved by using LSTM-Auto Encoder for a collective anomaly in which another feature exhibits an abnormal pattern according to a change in one by comparing two or more features. did In addition, evaluation indicators are set for the performance evaluation of the model presented in this study, and then AI model evaluation is performed.

• The Plant Pathology Journal
• /
• v.38 no.4
• /
• pp.395-402
• /
• 2022

To predict rice blast, many machine learning methods have been proposed. As the quality and quantity of input data are essential for machine learning techniques, this study develops three artificial neural network (ANN)-based rice blast prediction models by combining two ANN models, the feed-forward neural network (FFNN) and long short-term memory, with diverse input datasets, and compares their performance. The Blast_Weathe long short-term memory r_FFNN model had the highest recall score (66.3%) for rice blast prediction. This model requires two types of input data: blast occurrence data for the last 3 years and weather data (daily maximum temperature, relative humidity, and precipitation) between January and July of the prediction year. This study showed that the performance of an ANN-based disease prediction model was improved by applying suitable machine learning techniques together with the optimization of hyperparameter tuning involving input data. Moreover, we highlight the importance of the systematic collection of long-term disease data.

• Journal of The Korean Association of Information Education
• /
• v.25 no.5
• /
• pp.705-712
• /
• 2021

This study aims to develop artificial intelligence education programs that can be easily applied in elementary schools according to the trend of the times called artificial intelligence. The training program designed the purpose and direction based on the analysis results of the needs of 70 elementary school teachers according to the steps of the ADDIE model. According to the survey, elementary school students developed a machine learning education program to set sound data as the theme of the most accessible in their daily lives and to learn the principles of artificial intelligence in solving problems using sound data in real life. These days, when the need for artificial intelligence education emerges, elementary machine learning education programs that solve daily life problems based on sound data developed in this study will lay the foundation for elementary artificial intelligence education.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2021.07a
• /
• pp.41-44
• /
• 2021

In this paper, we propose a monitoring system that can monitor gas leakage concentrations in real time and forecast the amount of gas leaked after one minute. When gas leaks happen, they typically lead to accidents such as poisoning, explosion, and fire, so a monitoring system is needed to reduce such occurrences. Previous research has mainly been focused on analyzing explosion characteristics based on gas types, or on warning systems that sound an alarm when a gas leak occurs in industrial areas. However, there are no studies on creating systems that utilize specific gas explosion characteristic analysis or empirical urban gas data. This research establishes a deep learning model that predicts the gas explosion risk level over time, based on the gas data collected in real time. In order to determine the relative risk level of a gas leak, the gas risk level was divided into five levels based on the lower explosion limit. The monitoring platform displays the current risk level, the predicted risk level, and the amount of gas leaked. It is expected that the development of this system will become a starting point for a monitoring system that can be deployed in urban areas.

• 한국정보교육학회:학술대회논문집
• /
• 2021.08a
• /
• pp.7-11
• /
• 2021

This study designs block-based machine learning education program using sound data that can be easily applied in elementary schools. The education program designed its goals and directions based on the results of a demand analysis conducted on 70 elementary school teachers in advance according to the ADDIE model. Scratch in Machine Learning for Kids was used for block-based programming, and the education program was designed to discover regularity of data values using sound data, learn the principles of artificial intelligence, and improve computational thinking in the programming process. In a later study, the education program needs to verify what changes there are in attitudes and computational thinking about artificial intelligence.

• Journal of Information Processing Systems
• /
• v.19 no.5
• /
• pp.673-687
• /
• 2023

Convolutional neural network-based deep learning technology is the most commonly used in image identification, but it requires large-scale data for training. Therefore, application in specific fields in which data acquisition is limited, such as in the military, may be challenging. In particular, the identification of ground weapon systems is a very important mission, and high identification accuracy is required. Accordingly, various studies have been conducted to achieve high performance using small-scale data. Among them, the ensemble method, which achieves excellent performance through the prediction average of the pre-trained models, is the most representative method; however, it requires considerable time and effort to find the optimal combination of ensemble models. In addition, there is a performance limitation in the prediction results obtained by using an ensemble method. Furthermore, it is difficult to obtain the ensemble effect using models with imbalanced classification accuracies. In this paper, we propose a transfer learning-based feature fusion technique for heterogeneous models that extracts and fuses features of pre-trained heterogeneous models and finally, fine-tunes hyperparameters of the fully connected layer to improve the classification accuracy. The experimental results of this study indicate that it is possible to overcome the limitations of the existing ensemble methods by improving the classification accuracy through feature fusion between heterogeneous models based on transfer learning.

• Annual Conference of KIPS
• /
• 2023.05a
• /
• pp.510-512
• /
• 2023

In order to enhance a model's capability for detecting facial expressions, this research suggests a pipeline that makes use of the GradCAM component. The patching module and the pseudo-labeling module make up the pipeline. The patching component takes the original face image and divides it into four equal parts. These parts are then each input into a 2Dconvolutional layer to produce a feature vector. Each picture segment is assigned a weight token using GradCAM in the pseudo-labeling module, and this token is then merged with the feature vector using principal component analysis. A convolutional neural network based on transfer learning technique is then utilized to extract the deep features. This technique applied on a public dataset MMI and achieved a validation accuracy of 96.06% which is showing the effectiveness of our method.