실제세계에서 데이터 수집의 비용과 한계를 고려할 때, 시뮬레이션 생성 환경은 데이터 생성 과 다양한 시도에 있어 효율적인 대안이다. 이 연구에서는 Unity ML Agent를 로그라이크 장 르에 적합한 강화학습 모델로 구현하였다. 간단한 게임에Agent를 이식하고, 이 Agent가 적을 인식하고 대응하는 과정을 코드로 작성하였다. 초기 모델은 조준사격의 한계를 보였으나 RayPerceptionSensor-Component2D를 통해 Agent의 센서 정보를 직접 제공함으로써, Agent가 적을 감지하고 조준 사격을 하는 능력을 관찰할 수 있었다. 결과적으로, 개선된 모델 은 평균3.81배 향상된 성능을 보여주었으며, 이는 Unity ML Agent가 로그라이크 장르에서 강화학습을 통한 데이터 수집이 가능함을 입증한다.

Existing reinforced concrete buildings with seismically deficient column details affect the overall behavior depending on the failure type of column. This study aims to develop and validate a machine learning-based prediction model for the column failure modes (shear, flexure-shear, and flexure failure modes). For this purpose, artificial neural network (ANN), K-nearest neighbor (KNN), decision tree (DT), and random forest (RF) models were used, considering previously collected experimental data. Using four machine learning methodologies, we developed a classification learning model that can predict the column failure modes in terms of the input variables using concrete compressive strength, steel yield strength, axial load ratio, height-to-dept aspect ratio, longitudinal reinforcement ratio, and transverse reinforcement ratio. The performance of each machine learning model was compared and verified by calculating accuracy, precision, recall, F1-Score, and ROC. Based on the performance measurements of the classification model, the RF model represents the highest average value of the classification model performance measurements among the considered learning methods, and it can conservatively predict the shear failure mode. Thus, the RF model can rapidly predict the column failure modes with simple column details.

The government is implementing a policy to expand eco-friendly energy as a power source. However, the output of new and renewable energy is not constant. It is difficult to stably adjust the power supply to the power demand in the power system. Therefore, the government predicts day-ahead the amount of renewable energy generation to cope with the output volatility caused by the expansion of renewable energy. It is a system that pays a settlement amount if it transitions within a certain error rate the next day. In this paper, Machine Learning was used to study the prediction of power generation within the error rate.

PURPOSES : To enhance the accuracy of predicting the compressive strength of practical concrete mixtures, this study aimed to develop a machine learning model by utilizing the most commonly employed curing age, specifically, the 28-day curing period. The training dataset consisted of concrete mixture sample data at this curing age, along with samples subjected to a total load not exceeding 2,350 kg. The objective was to train a machine learning model to create a more practical predictive model suitable for real-world applications. METHODS : Three machine learning models—random forest, gradient boosting, and AdaBoost—were selected. Subsequently, the prepared dataset was used to train the selected models. Model 1 was trained using concrete sample data from the 28th curing day, followed by a comprehensive analysis of the results. For Model 2, training was conducted using data from the 28th day of curing, focusing specifically on instances where the total load was 2,350 kg or less. The results were systematically analyzed to determine the most suitable machine learning model for predicting the compressive strength of concrete. RESULTS : The machine learning model trained on concrete sample data from the 28th day of curing with a total weight of 2,350 kg or less exhibited higher accuracy than the model trained on weight-unrestricted data from the 28th day of curing. The models were evaluated in terms of accuracy, with the gradient boosting, AdaBoost, and random forest models demonstrating high accuracy, in that order. CONCLUSIONS : Machine learning models trained using concrete mix data based on practical and real-world scenarios demonstrated a higher accuracy than models trained on impractical concrete mix data. This case illustrates the significance of not only the quantity but also the quality of the data during the machine learning training process. Excluding outliers from the data appears to result in better accuracy for machine learning models. This underscores the importance of using high-quality and practical mixed concrete data for reliable and accurate model training.

PURPOSES : In this study, an optimal model for compressive strength prediction was derived by learning and directly comparing several machine learning models based on the same data. METHODS : Approximately 478 pieces of concrete compressive strength data were obtained to compare the performance of the machine learning models. In addition, five machine learning models were trained based on the obtained data. The performance of the learned model was compared using three performance indicators. Finally, the performance of the model trained using additional data was reviewed. RESULTS : As a result of comparing the performance of machine learning models, the XGB(eXtra Gradient Boost) model showed the best performance. In addition, as a result of the verification based on additional data, highly reliable results can be obtained if the XGB model is used to predict the compressive strength of concrete. CONCLUSIONS : If a concrete strength prediction model is derived based on a machine learning model, a highly reliable model can be derived.

This study investigated the convergence of content and language integrated learning, translanguaging, and global citizenship education in an EFL tertiary English class. Conceptualized within translanguaging as an assemblage for meaning-making, machine translation was incorporated into the course in a way that EFL bilinguals could fully avail themselves of their linguistic repertoire for the learning of global citizenship and language. The analyses of thirty-three students’ response essays and survey results demonstrate the success of MT as both a scaffold for bridging language-content gaps and a tool for language acquisition. Design features, perceived as important, were a careful introduction and training on MT use and teacher feedback on MT-assisted writing. Survey results emphasize the crucial role of the students’ L1 in meaning-making. The study offers a practical guide for educators interested in using MT in L2 writing instruction and encourages further research on the theoretical and pedagogical applications of translanguaging in diverse EFL contexts.

PURPOSES : Construction cost estimates are important information for business feasibility analysis in the planning stage of road construction projects. The quality of current construction cost estimates are highly dependent on the expert's personal experience and skills to estimate the arithmetic average construction cost based on past cases, which makes construction cost estimates subjective and unreliable. An objective approach in construction cost estimation shall be developed with the use of machine learning. In this study, past cases of road projects were analyzed and a machine learning model was developed to produce a more accurate and time-efficient construction cost estimate in teh planning stage. METHODS : After conducting case analysis of 100 road construction, a database was constructed including the road construction's details, drawings, and completion reports. To improve the construction cost estimation, Mallow's Cp. BIC, Adjusted R methodology was applied to find the optimal variables. Consequently, a plannigs-stage road construction cost estimation model was developed by applying multiple regression analysis, regression tree, case-based inference model, and artificial neural network (ANN, DNN). RESULTS : The construction cost estimation model showed excellent prediction performance despite an insufficient amount of learning data. Ten cases were randomly selected from the data base and each developed machine learning model was applied to the selected cases to calculate for the error rate, which should be less than 30% to be considered as acceptable according to American Estimating Association. As a result of the analysis, the error rates of all developed machine learning models were found to be acceptable with values rangine from 17.3% to 26.0%. Among the developed models, the ANN model yielded the least error rate. CONCLUSIONS : The results of this study can help raise awareness of the importance of building a systematic database in the construction industry, which is disadvantageous in machine learning and artificial intelligence development. In addition, it is believed that it can provide basic data for research to determine the feasibility of construction projects that require a large budget, such as road projects.

Numerous factors contribute to the deterioration of reinforced concrete structures. Elevated temperatures significantly alter the composition of the concrete ingredients, consequently diminishing the concrete's strength properties. With the escalation of global CO2 levels, the carbonation of concrete structures has emerged as a critical challenge, substantially affecting concrete durability research. Assessing and predicting concrete degradation due to thermal effects and carbonation are crucial yet intricate tasks. To address this, multiple prediction models for concrete carbonation and compressive strength under thermal impact have been developed. This study employs seven machine learning algorithms—specifically, multiple linear regression, decision trees, random forest, support vector machines, k-nearest neighbors, artificial neural networks, and extreme gradient boosting algorithms—to formulate predictive models for concrete carbonation and thermal impact. Two distinct datasets, derived from reported experimental studies, were utilized for training these predictive models. Performance evaluation relied on metrics like root mean square error, mean square error, mean absolute error, and coefficient of determination. The optimization of hyperparameters was achieved through k-fold cross-validation and grid search techniques. The analytical outcomes demonstrate that neural networks and extreme gradient boosting algorithms outshine the remaining five machine learning approaches, showcasing outstanding predictive performance for concrete carbonation and thermal effect modeling.

인공지능은 4차 산업혁명의 프레임이 소개된 이후 점차 보편적인 기술로 자리를 잡아가고 있으며, 인공지능 관련 특허 출원도 크게 증가하고 있다. 최근에는 특허 생태계가 출원 건수 위주의 양적 경쟁에서 고품질의 특허 확보라는 질적 경쟁으로 패러다임이 변화되면서, 저품질 특허로 인한 비용 손실에 관심이 높아지고 있다. 이러한 배경으로 본 연구에서는 머신러닝과 Doc2Vec 알고리즘을 활용하여 특허 품질을 예측하는 방법을 제안하고자 한다. 본 연구를 위해 WIPO에서 정의한 CPC 코드를 활용하여 미국 특허청(USPTO)에 등록된 인공지능 관련 특허 데이터를 추출하였고, 이를 통해 정형 데이터 기반 19개 변수, 비정형 데이터 기반 7개 변수를 개발하였다. 특히, 새롭게 제안하는 Doc2Vec 알고리즘을 이용한 제목과 초록 텍스트 유사도 변수는 고품질 특허를 예측하는데 영향을 미칠 것으로 판단된다. 이에 유사도 변수의 효과를 확인하기 위해 유사도 변수를 포함한 앙상블 기반 머신러닝 모델과 포함하지 않은 모델을 개발하여 비교하였다. 실험 결과, 유사도 변수를 포함한 모델이 AUC 0.013, f1-score 0.025가 높게 나타나 더 우수한 성능을 보였다. 이는 유사도 변수가 고품질 특허 예측에 기여한다는 것을 시사한다. 또한, SHAP을 이용하여 블랙박스 형태의 머신러닝 변수 영향도를 설명하였다. 본 연구를 통해 핵심 기술 분야인 인공지능과 같은 영역에서 특허의 품질을 예측하고, 고품질 특허 개발을 장려함으로써 사회적 가치를 실현하는 데 기여할 수 있을 것으로 기대한다.

Nuclear facilities present the important task related to the migration and retention of radioactive contaminants such as cesium (Cs), strontium (Sr), and cobalt (Co) for unexpected events in various environmental conditions. The distribution coefficient (Kd) is important factor for understanding these contaminants mobility, influenced by environmental variables. This study focusses the prediction of Kd values for radionuclides within solid phase groups through the application of machine-learning models trained on experimental data and open source data from Japan atomic energy agency. Three machine-learning models, such as the convolutional neural network, artificial neural network, and random forest, were trained for prediction model of the distribution coefficient (Kd). Fourteen input variables drawn from the database and experimental data, including parameters such as initial concentration, solid-phase characteristics, and solution conditions, served as the basis for model training. To enhance model performance, these variables underwent preprocessing steps involving normalization and log transformation. The performances of the models were evaluated using the coefficient of determination. These results showed that the environmental media, initial radionuclide concentration, solid phase properties, and solution conditions were significant variables for Kd prediction. These models accurately predict Kd values for different environmental conditions and can assess the environmental risk by analyzing the behavior of radionuclides in solid phase groups. The results of this study can improve safety analyses and longterm risk assessments related to waste disposal and prevent potential hazards and sources of contamination in the surrounding environment.

Over the years, in the field of safety assessment of geological disposal system, system-level models have been widely employed, primarily due to considerations of computational efficiency and convenience. However, system-level models have their limitations when it comes to phenomenologically simulating the complex processes occurring within disposal systems, particularly when attempting to account for the coupled processes in the near-field. Therefore, this study investigates a machine learning-based methodology for incorporating phenomenological insights into system-level safety assessment models without compromising computational efficiency. The machine learning application targeted the calculation of waste degradation rates and the estimation of radionuclide flux around the deposition holes. To develop machine learning models for both degradation rates and radionuclide flux, key influencing factors or input parameters need to be identified. Subsequently, process models capable of computing degradation rates and radionuclide flux will be established. To facilitate the generation of machine learning data encompassing a wide range of input parameter combinations, Latin-hypercube sampling will be applied. Based on the predefined scenarios and input parameters, the machine learning models will generate time-series data for the degradation rates and radionuclide flux. The time-series data can subsequently be applied to the system-level safety assessment model as a time table format. The methodology presented in this study is expected to contribute to the enhancement of system-level safety assessment models when applied.

Conducting a TSPA (Total System Performance Assessment) of the entire spent nuclear fuel disposal system, which includes thousands of disposal holes and their geological surroundings over many thousands of years, is a challenging task. Typically, the TSPA relies on significant efforts involving numerous parts and finite elements, making it computationally demanding. To streamline this process and enhance efficiency, our study introduces a surrogate model built upon the widely recognized U-network machine learning framework. This surrogate model serves as a bridge, correcting the results from a detailed numerical model with a large number of small-sized elements into a simplified one with fewer and large-sized elements. This approach will significantly cut down on computation time while preserving accuracy comparable to those achieved through the detailed numerical model.

Fouling is an inevitable problem in membrane water treatment plant. It can be measured by trans-membrane pressure (TMP) in the constant flux operation, and chemical cleaning is carried out when TMP reaches a critical value. An early fouilng alarm is defined as warning the critical TMP value appearance in advance. The alarming method was developed using one of machine learning algorithms, decision tree, and applied to a ceramic microfiltration (MF) pilot plant. First, the decision tree model that classifies the normal/abnormal state of the filtration cycle of the ceramic MF pilot plant was developed and it was then used to make the early fouling alarm method. The accuracy of the classification model was up to 96.2% and the time for the early warning was when abnormal cycles occurred three times in a row. The early fouling alram can expect reaching a limit TMP in advance (e.g., 15-174 hours). By adopting TMP increasing rate and backwash efficiency as machine learning variables, the model accuracy and the reliability of the early fouling alarm method were increased, respectively.

A machine learning-based algorithms have used for constructing species distribution models (SDMs), but their performances depend on the selection of backgrounds. This study attempted to develop a noble method for selecting backgrounds in machine-learning SDMs. Two machine-learning based SDMs (MaxEnt, and Random Forest) were employed with an example species (Spodoptera litura), and different background selection methods (random sampling, biased sampling, and ensemble sampling by using CLIMEX) were tested with multiple performance metrics (TSS, Kappa, F1-score). As a result, the model with ensemble sampling predicted the widest occurrence areas with the highest performance, suggesting the potential application of the developed method for enhancing a machine-learning SDM.

A causality exists between insect density and plant health, where plant health is affected by both the plant’s potential and environmental factors. In other words, causality is possible between insect density and environmental factors, allowing for the analysis of insect density based on these environmental factors. Machine learning enables studying insect density alongside environmental factors, providing insights into the causality between insects, the environment, and plant health. Machine learning is a methodology that involves the design of models by learning patterns from input data. This study aims to predict F. occidentalis density by sampling environmental factors and applying them to machine learning models.

모기는 감염병을 매개하는 종으로 전염병 확산 억제를 위해서는 개체수의 감시와 정확한 예측이 필요하다. 본 연구에서는 모기 개체수 및 기상 및 현장 자료를 활용해 모기 개체수 머신러닝 모델을 개발하였다. 모기 개체수는 디지털 모기 측정기(Digital Mosquito Monitoring System, DMS)의 2015 년~2022년의 5월~10월의 자료를 활용하였다. 기상 자료는 기온, 강수량, 풍속, 습도를 사용하였으며, 현장 조사 자료는 현장을 명목척도와 서열척도로 나누어 기록하여, 명목 척도의 경우 원핫 인코딩으 로 변환해 수치화하여 사용하였다. 분석에 사용된 머신러닝 모델은 Artificial Neural Network, Random Forest, Gradient Boosting Machine, Support Vector Machine이며 성능지표로 R2, RMSE를 사용하였다. 연구 결과, Gradient Boosting 모델이 R2 0.4, RMSE 22.45로 가장 좋은 성능을 나타냈다. 현장 조사 자료 를 분석에 활용하였을 때 R2는 증가하였고, RMSE는 감소하였다. 본 연구 결과 모기 개체수에 현장 조사 자료가 예측 정확도를 향상시킬 수 있음을 확인하였다.