• 예술인문사회 융합 멀티미디어 논문지
• /
• 제8권7호
• /
• pp.759-768
• /
• 2018

국가 간 다양한 분야에서의 교류가 증가함에 따라서 개최되는 국제회의의 수도 증가하고 있다. 이로 인해, 동시통역의 수요도 증가하고 있기 때문에 많은 학습자들이 동시통역을 학습하고 있다. 동시통역은 많은 학습 시간이 필요하기 때문에, 통역을 학습하는 학생들은 개별 학습을 수행한다. 자율적인 학습법 중 대표적으로 문장구역 훈련법이 있으나, 백트래킹으로 인한 학습 효과 저하 및 원활한 학습을 위한 파트너의 도움이 필요하다는 단점이 존재한다. 이를 해결하기 위해서 컴퓨터 기반 문장구역 훈련 시스템이 제안되어 학습자들의 학습 능률을 향상시켰다. 하지만, 컴퓨터의 경우 이동성이 매우 낮기 때문에 학습자들이 정해진 공간에서만 학습을 진행할 수 있기 때문에 활용도가 저하될 수 있다. 본 논문에서는 이동성으로 인한 활용도 저하 문제를 해결하기 위해 스마트 단말 기반의 문장구역 훈련 시스템을 제안한다. 스마트 단말은 컴퓨터에 비해 낮은 처리 용량을 가지고 있기 때문에 크기가 큰 파일을 처리할 때 성능이 저하된다. 이를 해결하기 위한 여러 알고리즘들을 제안한다. 제안한 스마트 단말 기반 문장구역 훈련 시스템을 구현하고 기능을 검증하였다.

• 제어로봇시스템학회논문지
• /
• 제9권11호
• /
• pp.917-924
• /
• 2003

Recurrent neural networks have been successfully applied to communications channel equalization. Major disadvantages of gradient-based learning algorithms commonly employed to train recurrent neural networks are slow convergence rates and long training sequences required for satisfactory performance. In a high-speed communications system, fast convergence speed and short training symbols are essential. We propose decision feedback equalizers using a recurrent neural network trained with Kalman filtering algorithms. The main features of the proposed recurrent neural equalizers, utilizing extended Kalman filter (EKF) and unscented Kalman filter (UKF), are fast convergence rates and good performance using relatively short training symbols. Experimental results for two time-varying channels are presented to evaluate the performance of the proposed approaches over a conventional recurrent neural equalizer.

• Smart Structures and Systems
• /
• 제30권6호
• /
• pp.673-686
• /
• 2022

Deep learning algorithms for Structural Health Monitoring (SHM) have been extracting the interest of researchers and engineers. These algorithms commonly used loss functions and evaluation indices like the mean square error (MSE) which were not originally designed for SHM problems. An updated loss function which was specifically constructed for deep-learning-based structural damage detection problems has been proposed in this study. By tuning the coefficients of the loss function, the weights for damage localization and quantification can be adapted to the real situation and the deep learning network can avoid unnecessary iterations on damage localization and focus on the damage severity identification. To prove efficiency of the proposed method, structural damage detection using convolutional neural networks (CNNs) was conducted on a truss bridge model. Results showed that the validation curve with the updated loss function converged faster than the traditional MSE. Data augmentation was conducted to improve the anti-noise ability of the proposed method. For reducing the training time, the normalized modal strain energy change (NMSEC) was extracted, and the principal component analysis (PCA) was adopted for dimension reduction. The results showed that the training time was reduced by 90% and the damage identification accuracy could also have a slight increase. Furthermore, the effect of different modes and elements on the training dataset was also analyzed. The proposed method could greatly improve the performance for structural damage detection on both the training time and detection accuracy.

• 한국공간구조학회논문집
• /
• 제20권2호
• /
• pp.51-58
• /
• 2020

Recently, deep learning that is the most popular and effective class of machine learning algorithms is widely applied to various industrial areas. A number of research on various topics about structural engineering was performed by using artificial neural networks, such as structural design optimization, vibration control and system identification etc. When nonlinear semi-active structural control devices are applied to building structure, a lot of computational effort is required to predict dynamic structural responses of finite element method (FEM) model for development of control algorithm. To solve this problem, an artificial neural network model was developed in this study. Among various deep learning algorithms, a recurrent neural network (RNN) was used to make the time history response prediction model. An RNN can retain state from one iteration to the next by using its own output as input for the next step. An eleven-story building structure with semi-active tuned mass damper (TMD) was used as an example structure. The semi-active TMD was composed of magnetorheological damper. Five historical earthquakes and five artificial ground motions were used as ground excitations for training of an RNN model. Another artificial ground motion that was not used for training was used for verification of the developed RNN model. Parametric studies on various hyper-parameters including number of hidden layers, sequence length, number of LSTM cells, etc. After appropriate training iteration of the RNN model with proper hyper-parameters, the RNN model for prediction of seismic responses of the building structure with semi-active TMD was developed. The developed RNN model can effectively provide very accurate seismic responses compared to the FEM model.

• 한국공간구조학회논문집
• /
• 제23권4호
• /
• pp.81-88
• /
• 2023

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.

• Computers and Concrete
• /
• 제34권3호
• /
• pp.367-378
• /
• 2024

This study evaluates the direct tensile strength of ultra-high-performance concrete (UHPC) using tests. A total of 45 dogbone-shaped specimens are tested, with the test variables being the fiber volume fraction and notch length. The test results showed that the material properties of UHPC were largely dependent on the fiber volume fraction and compressive strength. When steel fibers with more than 1% fiber volume fraction are mixed in the manufacturing of UHPC, the tensile strength can be more than twice that of plain UHPC. In addition, the incorporation of steel fibers enabled the significant improvement of the initial cracking strength. However, the effect of the notch length on the tensile behavior was insignificant. An assessment of the direct tensile strength is conducted using machine-learning algorithms (ML). For evaluation of the direct tensile strength of UHPC using ML, a total of 98 test data, including 53 data from other research works and 45 data from this experimental program, were collected. In total, 67 data with a 70% confidence interval on a normal distribution curve were selected, with 47 data among 67 used for ML training and 20 data used for ML testing. As a result, the machine-learning algorithm with a steel fiber volume fraction predicted that the tensile strength has an average of 0.98 and the lowest values of regression evaluation metrics among analytical and ML-based models. It is considered that an ML-based model can help to predict a more accurate tensile strength of UHPC.

• International Journal of Computer Science & Network Security
• /
• 제21권2호
• /
• pp.148-157
• /
• 2021

The effectiveness of recommendation systems depends on the performance of the algorithms with which these systems are designed. The quality of the algorithms themselves depends on the quality of the strategies with which they were designed. These strategies differ from author to author. Thus, designing a good recommendation system means implementing the good strategies. It's in this context that several research works have been proposed on various strategies applied to algorithms to meet the needs of recommendations. Researchers are trying indefinitely to address this objective of seeking the qualities of recommendation algorithms. In this paper, we propose a new algorithm for recommending learning items. Learner performance predictions and collaborative recommendation methods are used as strategies for this algorithm. The proposed performance prediction model is based on convolutional neural networks (CNN). The results of the performance predictions are used by the proposed recommendation algorithm. The results of the predictions obtained show the efficiency of Deep Learning compared to the k-nearest neighbor (k-NN) algorithm. The proposed recommendation algorithm improves the recommendations of the learners' learning items. This algorithm also has the particularity of dissuading learning items in the learner's profile that are deemed inadequate for his or her training.

• 로봇학회논문지
• /
• 제15권2호
• /
• pp.197-204
• /
• 2020

In this paper, we present a learning platform for robotic grasping in real world, in which actor-critic deep reinforcement learning is employed to directly learn the grasping skill from raw image pixels and rarely observed rewards. This is a challenging task because existing algorithms based on deep reinforcement learning require an extensive number of training data or massive computational cost so that they cannot be affordable in real world settings. To address this problems, the proposed learning platform basically consists of two training phases; a learning phase in simulator and subsequent learning in real world. Here, main processing blocks in the platform are extraction of latent vector based on state representation learning and disentanglement of a raw image, generation of adapted synthetic image using generative adversarial networks, and object detection and arm segmentation for the disentanglement. We demonstrate the effectiveness of this approach in a real environment.

• 대한영상의학회지
• /
• 제81권6호
• /
• pp.1274-1289
• /
• 2020

최근, 의료 영상 분야에서 딥러닝은 가장 활발하게 연구되고 있는 기술 중 하나이다. 충분한 데이터와 최신의 딥러닝 알고리즘은 딥러닝 모델의 개발에 중요한 요소이다. 하지만 일반화된 최적의 딥러닝 모델을 개발하기 위해서는 데이터의 양과 최신의 딥러닝 알고리즘 외에도 많은 것을 고려해야 한다. 데이터 수집부터 가공, 전처리, 모델의 학습 및 검증, 경량화까지 모든 과정이 딥러닝 모델의 성능에 영향을 미칠 수 있기 때문이다. 본 종설에서는 의료 영상에 최적화된 딥러닝 모델을 위해 개발 과정 각각에서 고려해야 할 중요한 요소들을 살펴보고자 한다.

• 전자공학회논문지CI
• /
• 제39권1호
• /
• pp.23-34
• /
• 2002

본 논문에서는 진화전략을 이용하여 빠르게 학습하는 새로운 구조의 뉴로퍼지 시스템을 제안하고 제안한 시스템의 효용성을 입증하기 위하여 비선형 시스템 동정에 응용한 결과를 설명한다. 뉴로퍼지 시스템의 학습 방법으로는 지금까지 주로 변형된 오류역전파 알고리즘과 최적화 기법인 유전자 알고리즘이 많이 사용되어왔으나, 오류역전파 알고리즘은 학습시간이 많이 걸리며 유전자 알고리즘은 해를 유전형 형태로 표현함으로 인하여 미세한 탐색이 힘든 단점이 있었다. 본 논문에서 사용한 진화전력은 해를 표현형의 개체로 나타내어 실수형태로 진화하기 대문에 미세한 탐색이 가능하며 오류역전파 알고리즘에 비해 지역해에 빠질 가능성이 작고 속도가 빠른 장점이 있다. 제안한 뉴로퍼지 시스템을 비선형 시스템 동정에 적용한 결과 학습속도가 빠르며 학습결과도 우수함을 보았다.