• Journal of the Korea Institute of Military Science and Technology
• /
• v.27 no.4
• /
• pp.474-484
• /
• 2024

Reinforcement learning, which are also studied in the field of defense, face the problem of sample efficiency, which requires a large amount of data to train. Transfer learning has been introduced to address this problem, but its effectiveness is sometimes marginal because the model does not effectively leverage prior knowledge. In this study, we propose a stochastic initial state randomization(SISR) method to enable robust knowledge transfer that promote generalized and sufficient knowledge transfer. We developed a simulation environment involving a cooperative robot transportation task. Experimental results show that successful tasks are achieved when SISR is applied, while tasks fail when SISR is not applied. We also analyzed how the amount of state information collected by the agents changes with the application of SISR.

• Journal of the Semiconductor & Display Technology
• /
• v.23 no.1
• /
• pp.19-24
• /
• 2024

It is not an exaggeration to say that a single fire causes a lot of damage, so fires are one of the disaster situations that must be alerted as soon as possible. Various technologies have been utilized so far because preventing and detecting fires can never be completely accomplished with individual human efforts. Recently, deep learning technology has been developed, and fire detection systems using object detection neural networks are being actively studied. In this paper, we propose a new fire detection system that improves the previously studied fire detection system. We train the YOLOv8 model using refined datasets through improved labeling methods, derive results, and demonstrate the superiority of the proposed system by comparing it with the results of previous studies.

• Annual Conference of KIPS
• /
• 2024.05a
• /
• pp.294-297
• /
• 2024

Malware detection has become increasingly critical with the proliferation of end devices. To improve detection rates and efficiency, the research focus in malware detection has shifted towards leveraging machine learning and deep learning approaches. This shift is particularly relevant in the context of the widespread adoption of end devices, including smartphones, Internet of Things devices, and personal computers. Machine learning techniques are employed to train models on extensive datasets and evaluate various features, while deep learning algorithms have been extensively utilized to achieve these objectives. In this research, we introduce TabNet, a novel architecture designed for deep learning with tabular data, specifically tailored for enhancing malware detection techniques. Furthermore, the Synthetic Minority Over-Sampling Technique is utilized in this work to counteract the challenges posed by imbalanced datasets in machine learning. SMOTE efficiently balances class distributions, thereby improving model performance and classification accuracy. Our study demonstrates that SMOTE can effectively neutralize class imbalance bias, resulting in more dependable and precise machine learning models.

• The Journal of the Acoustical Society of Korea
• /
• v.40 no.5
• /
• pp.479-487
• /
• 2021

In this paper, we propose an Sound Event Detection (SED) model using self-training based on a noisy student model. The proposed SED model consists of two stages. In the first stage, a mean-teacher model based on an Residual Convolutional Recurrent Neural Network (RCRNN) is constructed to provide target labels regarding weakly labeled or unlabeled data. In the second stage, a self-training-based noisy student model is constructed by applying different noise types. That is, feature noises, such as time-frequency shift, mixup, SpecAugment, and dropout-based model noise are used here. In addition, a semi-supervised loss function is applied to train the noisy student model, which acts as label noise injection. The performance of the proposed SED model is evaluated on the validation set of the Detection and Classification of Acoustic Scenes and Events (DCASE) 2020 Challenge Task 4. The experiments show that the single model and ensemble model of the proposed SED based on the noisy student model improve F1-score by 4.6 % and 3.4 % compared to the top-ranked model in DCASE 2020 challenge Task 4, respectively.

• Journal of the Korean Society for Railway
• /
• v.17 no.3
• /
• pp.210-215
• /
• 2014

Various methods have been developed in an effort to increase the scheduled speeds of the urban railways. Reducing the train dwell times by extending door widths is one such method. However, there is thus far no domestic model of boarding and alighting that is appropriate to lead to boarding and alighting time reductions if the door width is extended. Foreign models are not suitable because human behaviors, which are important factors when assessing boarding and alighting times, differ from country to country. In this study, a boarding and alighting model for domestic urban railways is proposed and related equations and parameters are derived from measured and experimental data. The model can be employed to assess time reductions in Korean railroad system if the door widths are extended.

• Journal of the Earthquake Engineering Society of Korea
• /
• v.8 no.4
• /
• pp.63-69
• /
• 2004

In this paper, a nonlinear modeling of lead rubber bearings(LRBs) was presented by a neural network theory. An shaking table test for a scaled frame model, of which base was isolated by the LRBs, was performed to verify numerical accuracies of the neural network model. White noise and three types of seismic records were adoped as base loads of the shaking table in order to train and generalize the neural network in case of seismic loads, numerical results of the neural network model were evaluated according to different magnitudes of PGA. As results, it is concluded that the presented neural network model has given a good agreement with the experimental data in details and can be useful to a nonlinear modeling of LRBs within prescribed domains.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.37 no.4
• /
• pp.581-587
• /
• 2013

In a magnetic train set composed of more than two cars, the installation of dampers between cars is carefully considered for improving both the ride quality and the safety, particularly during curve negotiation. In this study, a dynamic simulation of the ride quality and curve negotiation of a Maglev vehicle was carried out. The dynamic model is developed based on multibody dynamics. The presented full vehicle multibody dynamic model integrates the electromagnet model and its control algorithm. By using this model, the effects of the dampers are numerically analyzed. The proposed damper is installed on the vehicle and tested to analyze its effects. In this study, the simulation and measured results of the vehicle behavior and ride quality are discussed.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.30 no.1
• /
• pp.55-61
• /
• 2017

Recently several researches have been conducted to develop sliding track system in which friction between concrete track and bridge slab has been reduced. This paper investigated shear load carrying capacity of lateral supporting concrete block which should be implemented to resist lateral load due to train in sliding track system. In order to evaluate shear load carrying capacity of lateral supporting concrete block, analytical model has been developed considering concrete friction and rebar dowel action along construction joint. The proposed model predicted test results on the shear load carrying capacity from literature conservatively by 13~23% because effect of aggregate interlock along crack surface was neglected. Since construction joint status is ambiguous on construction site, it can be concluded that the proposed model can be used for reasonable design of lateral supporting concrete block. Based on the proposed model, design proposal for lateral supporting concrete block has been established.

• Journal of Korean Society of Transportation
• /
• v.22 no.3 s.74
• /
• pp.167-177
• /
• 2004

The Hub-and-spoke strategy is widely used in the field of transportation. According to containerization and the development of transshipment technology, it is also introduced into European rail freight transportation. The objective of this article is to develop a service network design model for rail freight transportation based on the Hub-and-spoke strategy and efficient algorithms that can be applied to large-scale network. Although this model is for strategic decision, it includes not only general operational cost but also time-delay cost. The non-linearity of objective function due to time-delay factor is transformed into linearity by establishing train service variables by frequency. To solve large scale problem, this model used a heuristic method based on decomposition and three newly-developed algorithms. The new algorithms were examined with respect to four test problems base on the actual network of European rail freight and discussed the accuracy of solutions and the efficiency of proposed algorithms.

• Journal of Korean Society on Water Environment
• /
• v.24 no.1
• /
• pp.91-98
• /
• 2008

This work was performed to develop a model possible to predict the influent flow and influent components, which are one of main disturbances causing process problems at the operation of municipal wastewater treatment plant. In this study, artificial neural network (ANN) was used in order to develop a model that was able to predict the influent flow, $COD_{Mn}$, SS, TN 1 day-ahead, 2day-ahead and 3 day ahead. Multi-layer feed-forward back-propagation network was chosen as neural network type, and tanh-sigmoid function was used as activation function to transport signal at the neural network. And Levenberg-Marquart (LM) algorithm was used as learning algorithm to train neural network. Among 420 data sets except missing data, which were collected between 2005 and 2006 at field plant, 210 data sets were used for training, and other 210 data sets were used for validation. As result of it, ANN model for predicting the influent flow and components 1-3day ahead could be developed successfully. It is expected that this developed model can be practically used as follows: Detecting the fault related to effluent concentration that can be happened in the future by combining with other models to predict process performance in advance, and minimization of the process fault through the establishment of various control strategies based on the detection result.