• IEMEK Journal of Embedded Systems and Applications
• /
• v.15 no.1
• /
• pp.1-8
• /
• 2020

A derivation method for a quantized gradient for machine learning on an embedded system is proposed, in this paper. The proposed differentiation method induces the quantized gradient vector to an objective function and provides that the validation of the directional derivation. Moreover, mathematical analysis shows that the sequence yielded by the learning equation based on the proposed quantization converges to the optimal point of the quantized objective function when the quantized parameter is sufficiently large. The simulation result shows that the optimization solver based on the proposed quantized method represents sufficient performance in comparison to the conventional method based on the floating-point system.

• Communications for Statistical Applications and Methods
• /
• v.26 no.6
• /
• pp.591-610
• /
• 2019

Deep Learning is one of the machine learning methods to find features from a huge data using non-linear transformation. It is now commonly used for supervised learning in many fields. In particular, Convolutional Neural Network (CNN) is the best technique for the image classification since 2012. For users who consider deep learning models for real-world applications, Keras is a popular API for neural networks written in Python and also can be used in R. We try examine the parameter estimation procedures of Deep Neural Network and structures of CNN models from basics to advanced techniques. We also try to figure out some crucial steps in CNN that can improve image classification performance in the CIFAR10 dataset using Keras. We found that several stacks of convolutional layers and batch normalization could improve prediction performance. We also compared image classification performances with other machine learning methods, including K-Nearest Neighbors (K-NN), Random Forest, and XGBoost, in both MNIST and CIFAR10 dataset.

• Journal of Institute of Control, Robotics and Systems
• /
• v.17 no.8
• /
• pp.777-782
• /
• 2011

Neural network technique is widely employed in the fields of signal processing, control systems, pattern recognition, etc. Learning of neural networks is an important procedure to accomplish dynamic system modeling. This paper presents a novel learning approach for differential neural network models based on the Kalman-Bucy filter theory. We construct an augmented state vector including original neural state and parameter vectors and derive a state estimation rule avoiding gradient function terms which involve to the conventional neural learning methods such as a back-propagation approach. We carry out numerical simulation to evaluate the proposed learning approach in nonlinear system modeling. By comparing to the well-known back-propagation approach and Kalman-Bucy filtering, its superiority is additionally proved under stochastic system environments.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.33B no.2
• /
• pp.159-167
• /
• 1996

In this paper, a new competitive learning algorithm called dynamic competitive learning (DCL) is presented. DCL is a supervised learning mehtod that dynamically generates output neuraons and nitializes weight vectors from training patterns. It introduces a new parameter called LOG (limit of garde) to decide whether or not an output neuron is created. In other words, if there exist some neurons in the province of LOG that classify the input vector correctly, then DCL adjusts the weight vector for the neuraon which has the minimum grade. Otherwise, it produces a new output neuron using the given input vector. It is largely learning is not limited only to the winner and the output neurons are dynamically generated int he trining process. In addition, the proposed algorithm has a small number of parameters. Which are easy to be determined and applied to the real problems. Experimental results for patterns recognition of remote sensing data and handwritten numeral data indicate the superiority of dCL in comparison to the conventional competitive learning methods.

• Design & Manufacturing
• /
• v.16 no.4
• /
• pp.60-66
• /
• 2022

Recently, with the development of industry and the improvement of living standards, various wastes are generated along with the production of various products. Most of these wastes are used as containers for products, and plastic or aluminum is used. Various attempts are being made to automate the classification of these wastes due to the high labor cost, but most of them are solved by manpower due to the geometrical shape change due to the nature of the waste. In this study, in order to automate the waste sorting task, Deep Learning technology is applied to a robot system for waste sorting and a vision system for waste sorting to effectively perform sorting tasks according to the shape of waste. As a result of the experiment, a Deep Learning parameter suitable for waste sorting was selected. In addition, through various experiments, it was confirmed that 99% of wastes could be selected in individual & group image learning. It is expected that this will enable automation of the waste sorting operation.

• The Journal of Korea Robotics Society
• /
• v.18 no.2
• /
• pp.225-232
• /
• 2023

With the recent increase of multiple robots cooperating in smart manufacturing logistics environments, it has become very important how to predict the safety and efficiency of the individual tasks and dynamically assign them to the best one of available robots. In this paper, we propose a novel task policy learner based on deep relational reinforcement learning for predicting the safety and efficiency of tasks in a multi-robot manufacturing logistics environment. To reduce learning complexity, the proposed system divides the entire safety/efficiency prediction process into two distinct steps: the policy parameter estimation and the rule-based policy inference. It also makes full use of domain-specific knowledge for policy rule learning. Through experiments conducted with virtual dynamic manufacturing logistics environments using NVIDIA's Isaac simulator, we show the effectiveness and superiority of the proposed system.

• Educational Technology International
• /
• v.10 no.2
• /
• pp.63-83
• /
• 2009

The purpose of this study was to investigate the cognitive mechanism of e-Learning contents development projects on the basis of the Shared Mental Model theory perspective. To pursue the purpose, a theoretical model and several hypotheses were developed based on relevant literature. Thirty five (35) e-Learning contents development teams composed of 202 instructional designers from for-profit professional e-Learning companies in Korea were participated in this study. For the analyses of the fit of the Model and parameter estimations, Structural Equation Modeling (SEM) method was employed. As hypothesized, e-Learning contents development team members' interaction leads to higher SMMs which in turn facilitate member satisfaction within the team. Meanwhile, the frequency of interaction among team members decreases as projects progress.

• Journal of Platform Technology
• /
• v.11 no.4
• /
• pp.80-88
• /
• 2023

To construct a learning model in the field of artificial intelligence, a dataset should be collected and be delivered to the central server where the learning model is constructed. Federated learning is a machine learning method building a global learning model without transmitting data located in a client side in a collaborative manner. It can be used to protect privacy, and after constructing a local trained model on individual clients, the parameters of the local model are aggregated centrally to update the global model. In this paper, we reuse the existing learning parameter to improve federated learning, describe incremental federated learning. For this work, we do experiments using the federated learning framework named Flower, and evaluate the experiment results with regard to elapsed time and precision when executing optimization algorithms.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.34S no.7
• /
• pp.102-111
• /
• 1997

In this paper, a patern recognition method of surface mount device(SMD) IC using wavelet transform and neural network is proposed. We chose the feature parameter according to the characteristics of coefficient matrix which is obtained from four level discrete wavelet transform (DWT). These feature parameters are normalized and then used for the input vector of neural network which is capable of adapting the surroundings such as variation of illumination, arrangement of objects and translation. Experimental results show that when the same form of feature pattern, as is used for learning, is put into neural network and gained 100% rate ofrecognition irrespective of SMD IC kinds, location and variation of illumination. In the case of unused feature pattern for learning, the recognition rate is 85.9% under the similar surroundings, where as an average recognition rate is 96.87% for the case of reregulated value of illumination. Proosed method is relatively simple compared with the traditional space domain method in extracting the feature parameter and is also well suited for recognizing the pattern's class, position and existence. It can also shorten the processing tiem better than method extracting feature parameter with the use of discrete cosine transform(DCT) and adapt the surroundings such as variation of illumination, the arrangement and the translation of SMD IC.

• Nuclear Engineering and Technology
• /
• v.49 no.7
• /
• pp.1423-1430
• /
• 2017

Detonation cell width is an important parameter in hydrogen explosion assessments. The experimental data on gas detonation are statistically analyzed to establish a universal method to numerically predict detonation cell widths. It is commonly understood that detonation cell width, ${\lambda}$, is highly correlated with the characteristic reaction zone width, ${\delta}$. Classical parametric regression methods were widely applied in earlier research to build an explicit semiempirical correlation for the ratio of ${\lambda}/{\delta}$. The obtained correlations formulate the dependency of the ratio ${\lambda}/{\delta}$ on a dimensionless effective chemical activation energy and a dimensionless temperature of the gas mixture. In this paper, support vector regression (SVR), which is based on nonparametric machine learning, is applied to achieve functions with better fitness to experimental data and more accurate predictions. Furthermore, a third parameter, dimensionless pressure, is considered as an additional independent variable. It is found that three-parameter SVR can significantly improve the performance of the fitting function. Meanwhile, SVR also provides better adaptability and the model functions can be easily renewed when experimental database is updated or new regression parameters are considered.