• 한국건축시공학회:학술대회논문집
• /
• 한국건축시공학회 2021년도 가을 학술논문 발표대회
• /
• pp.30-31
• /
• 2021

This study aims to build a deep learning model that can predict the value of concrete mixing properties according to a given concrete strength value. A model was created for a total of 1,291 concrete data, including 8 characteristics related to concrete mixing elements and environment, and the compressive strength of concrete. As the deep learning model, DNN-3L-256N, which showed the best performance on the prior study, was used. The average value for each characteristic of the data set was used as the initial input value. In results, in the case of 'curing temperature', which had a narrow range of values in the existing data set, showed the lowest error rate with less than 1% error based on MAE. The highest error rate with an error of 12 to 14% for fly and bfs.

• International journal of advanced smart convergence
• /
• 제9권2호
• /
• pp.90-104
• /
• 2020

It is difficult for non-experts to build machine learning (ML) models at the level that satisfies their needs. Deep learning models are even more challenging because it is unclear how to improve the model, and a trial-and-error approach is not feasible since training these models are time-consuming. To assist these novice users, we examined how interactive and explainable feedback while training a deep learning network can contribute to model performance and users' satisfaction, focusing on the data preparation process. We conducted a user study with 31 participants without expertise, where they were asked to improve the accuracy of a deep learning model, varying feedback conditions. While no significant performance gain was observed, we identified potential barriers during the process and found that interactive and explainable feedback provide complementary benefits for improving users' understanding of ML. We conclude with implications for designing an interface for building ML models for novice users.

• Earthquakes and Structures
• /
• 제25권6호
• /
• pp.469-479
• /
• 2023

A probabilistic seismic damage analysis is an essential procedure to identify seismically vulnerable structures, prioritize the seismic retrofit, and ultimately minimize the overall seismic risk. To assess the seismic risk of multiple structures within a region, a large number of nonlinear time-history structural analyses must be conducted and studied. As a result, each assessment requires high computing resources. To overcome this limitation, we explore a deep learning-based metamodel to enable the prediction of the mean and the standard deviation of the seismic damage distribution of track-on steel-plate girder railway bridges in Korea considering the geometric variation. For machine learning training, nonlinear dynamic time-history analyses are performed to generate 800 high-fidelity datasets on the seismic response. Through intensive trial and error, the study is concentrated on developing an optimal machine learning architecture with the pre-identified variables of the physical configuration of the bridge. Additionally, the prediction performance of the proposed method is compared with a previous, well-defined, response surface model. Finally, the statistical testing results indicate that the overall performance of the deep-learning model is improved compared to the response surface model, as its errors are reduced by as much as 61%. In conclusion, the model proposed in this study can be effectively deployed for the seismic fragility and risk assessment of a region with a large number of structures.

• International journal of advanced smart convergence
• /
• 제8권4호
• /
• pp.75-81
• /
• 2019

Though recent advancement of deep learning methods have provided satisfactory results from large data domain, somehow yield poor performance on few-shot classification tasks. In order to train a model with strong performance, i.e. deep convolutional neural network, it depends heavily on huge dataset and the labeled classes of the dataset can be extremely humongous. The cost of human annotation and scarcity of the data among the classes have drastically limited the capability of current image classification model. On the contrary, humans are excellent in terms of learning or recognizing new unseen classes with merely small set of labeled examples. Few-shot learning aims to train a classification model with limited labeled samples to recognize new classes that have neverseen during training process. In this paper, we increase the backbone depth of the embedding network in orderto learn the variation between the intra-class. By increasing the network depth of the embedding module, we are able to achieve competitive performance due to the minimized intra-class variation.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권12호
• /
• pp.3855-3867
• /
• 2022

Name separation (splitting full names into surnames and given names) is not a tedious task in a multiethnic country because the procedure for splitting surnames and given names is ethnicity-specific. Malaysia has multiple main ethnic groups; therefore, separating Malaysian full names into surnames and given names proves a challenge. In this study, we develop a two-phase framework for Malaysian name separation using deep learning. In the initial phase, we predict the ethnicity of full names. We propose a recurrent neural network with long short-term memory network-based model with character embeddings for prediction. Based on the predicted ethnicity, we use a rule-based algorithm for splitting full names into surnames and given names in the second phase. We evaluate the performance of the proposed model against various machine learning models and demonstrate that it outperforms them by an average of 9%. Moreover, transfer learning and fine-tuning of the proposed model with an additional dataset results in an improvement of up to 7% on average.

• 웰빙융합연구
• /
• 제6권2호
• /
• pp.27-31
• /
• 2023

Purpose: Establishment of a real-time monitoring system for odor control in traditional markets in Gangwon-do and a system for linking prevention facilities. Research design, data and methodology: Build server and system logic based on data through real-time monitoring device (sensor-based). A temporary data generation program for deep learning is developed to develop a model for odor data. Results: A REST API was developed for using the model prediction service, and a test was performed to find an algorithm with high prediction probability and parameter values optimized for learning. In the deep learning algorithm for AI modeling development, Pandas was used for data analysis and processing, and TensorFlow V2 (keras) was used as the deep learning library. The activation function was swish, the performance of the model was optimized for Adam, the performance was measured with MSE, the model method was Functional API, and the model storage format was Sequential API (LSTM)/HDF5. Conclusions: The developed system has the potential to effectively monitor and manage odors in traditional markets. By utilizing real-time data, the system can provide timely alerts and facilitate preventive measures to control and mitigate odors. The AI modeling component enhances the system's predictive capabilities, allowing for proactive odor management.

• Korean Journal of Radiology
• /
• 제25권3호
• /
• pp.217-219
• /
• 2024

• 한국컴퓨터정보학회논문지
• /
• 제25권9호
• /
• pp.37-44
• /
• 2020

AI 기술은 법률, 특허, 금융, 국방의 의사결정지원 기술 형태로 발전하여 질병 진단과 법률 판정 등에 적용되고 있다. Deep Learning으로 실시간 정보를 검색하려면, Big data Analysis과 Deep Learning Algorithm이 필요하다. 본 논문에서는 Deep Learning 모델인 RNN(Recurrent Neural Network)을 이용하여 상위권 대학 진학률을 예측하고자 한다. 우선, 행정구역 사설학원 현황과 행정구역 연령별 학생 수를 분석하고 교육열이 높은 지역에 거주하는 학생이 상위권 대학 진학률이 높다는 사회 통념의 가설을 설정했다. 예측된 가설과 정부의 공공데이터를 활용하여 분석된 자료를 토대로 검증하고자 한다. 예측모델은 2015년부터 2017년까지의 데이터를 활용하여 상위권 진학률을 예상하도록 학습하고, 학습된 모델은 2018년 상위권 진학률을 예측한다. 교육특구지역의 상위권 진학률을 Deep Learning 모델인 RNN을 이용하여 예측 실험을 수행했다. 본 논문은 교육열이 높은 지역의 사설학원 현황, 연령별 학생 수에 미치는 영향에 대해서 가구소득, 사교육의 참여 비율을 분석하여 상위권 진학률의 상관관계를 정의한다.

• Composites Research
• /
• 제36권2호
• /
• pp.92-100
• /
• 2023

본 논문에서는 유리섬유 직물 강화 복합재 소재위에 Cross-Dipole 패턴이 배치된 정형적 Circuit Analog(CA) 전파 흡수 구조 설계를 위한 머신러닝 및 딥러닝 모델을 제시하였다. 제시된 모델은 Cross-Dipole 패턴의 형상에 따라서 Ku-band (12-18 GHz)에서의 전파흡수성능을 3차원 전자파 수치해석 없이 바로 계산할 수 있다. 이를 위하여 다양한 머신러닝 및 딥러닝 기술을 적용한 최적 학습 모델을 도출하고, 학습 모델이 계산한 결과를 3차원 전자파 수치해석결과로 얻은 전파흡수특성과 비교함으로써 각각의 모델 간의 성능의 비교우위를 평가하였다. 개발된 모델들은 대부분 수치해석결과와 유사한 계산결과를 보여주었지만, 그 중 Fully-Connected 모델이 가장 유사한 계산결과를 제공할 수 있음을 확인하였다.

• 한국정보통신학회:학술대회논문집
• /
• 한국정보통신학회 2021년도 추계학술대회
• /
• pp.555-556
• /
• 2021

우리는 패션 의류 이미지의 빠르고 정확한 분류를 달성하기 위해 최적화된 동적 붕괴 학습률과 개선된 모델 구조를 가진 딥 러닝 모델을 기반으로 하는 새로운 방법을 제안한다.