• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2005년도 제36회 하계학술대회 논문집 D
• /
• pp.2681-2683
• /
• 2005

In this paper, we present the structure of fuzzy neural network(FNN) based on wavelet function, and apply this network structure to the identification of nonlinear systems. For adjusting the shape of membership function and the connection weights, the parameter learning method based on the gradient descent scheme is adopted. And an approach that uses adaptive learning rates is driven via a Lyapunov stability analysis to guarantee the fast convergence. Finally, to verify the efficiency of our network structure. we compare the Identification performance of proposed wavelet based fuzzy neural network(WFNN) with those of the FNN, the wavelet fuzzy model(WFM) and the wavelet neural network(WNN) through the computer simulation.

• 전자공학회논문지C
• /
• 제34C권9호
• /
• pp.61-67
• /
• 1997

The HMM-Net is an architecture for a neural network that implements a hidden markov model(HMM). The architecture is developed for the purpose of combining the classification power of neural networks with the time-domain modeling capability of HMMs. Criteria which are used for learning HMM_Net classifiers are maximum likelihood(ML), maximum mutual information (MMI), and minimization of mean squared error(MMSE). In this classifiers trained by the gradient descent algorithm with the above criteria. Experimental results for the isolated numbers from /young/to/koo/ show that in the binary inputs the performance of MMSE is better than the others, while in the fuzzy inputs the performance of MMI is better than the others.

• 한국정밀공학회지
• /
• 제23권9호
• /
• pp.84-91
• /
• 2006

The synthesis of optimal controllers for multivariable systems usually requires an accurate linear model of the plant dynamics. Real systems, however, contain nonlinearities and high-order dynamics that may be difficult to model using conventional techniques. This paper presents a novel loaming method for the synthesis of LQR controllers that doesn't require explicit modeling of the plant dynamics. This method utilizes the sign of Jacobian and gradient descent techniques to iteratively reduce the LQR objective function. It becomes easier and more convenient because it is relatively very easy to get the sign of Jacobian instead of its Jacobian. Simulations involving an overhead crane and a hydrofoil catamaran show that the proposed LQR-LC algorithm improves controller performance, even when the Jacobian information is estimated from input-output data.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1995년도 추계학술대회 논문집 학회본부
• /
• pp.207-209
• /
• 1995

The PID controller is widely used due to its fast response and robustness. But its performance is not so good compared with modem controllers such as adaptive, robust, fuzzy, neural controller. Therefore, it is natural to replace PID controller by modem controllers. But, the problem is that modem controller can not be easily applied to the real time process. Hence, this paper proposes such a structure that PID controller and Self-Learning Fuzzy Controller(SLFC) are in parallel with each other. The parameter of SLFC will be updated by gradient descent method using neuro - identifier. The usefulness of this hybrid controller will be proved by simulation results.

• 전력전자학회:학술대회논문집
• /
• 전력전자학회 1998년도 전력전자학술대회 논문집
• /
• pp.384-387
• /
• 1998

This paper presents a newly developed speed sensorless drive using RLS(Recursive Least Squares) based on Neural Network Training Algorithm. The proposed algorithm based on the RLS has just the time-varying learning rate, while the well-known back-propagation (or generalized delta rule) algorithm based on gradient descent has a constant learning rate. The number of iterations required by the new algorithm to converge is less than that of the back-propagation algorithm. The RLS based on NN is used to adjust the motor speed so that the neural model output follows the desired trajectory. This mechanism forces the estimated speed to follow precisely the actual motor speed. In this paper, a flux estimation strategy using filter concept is discussed. The theoretical analysis and experimental results to verify the effectiveness of the proposed analysis and the proposed control strategy are described.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 1998년도 추계학술대회 학술발표 논문집
• /
• pp.560-567
• /
• 1998

This paper concerned about a imputed color printed material image in camera to decrease noise and distortion by processing median filtering with input image to identical condition. Also this paper proposed the way of compares a normal printed material with an abnormal printed material color tone with trained a learning of the error back-propagation to block classification by extracting five place from identical block(3${\times}$3) of color printed material R, G, B value. As a representative algorithm of multi-layer perceptron the error Back-propagation technique used to solve complex problems. However, the Error Back-propagation is algorithm which basically used a gradient descent method which can be converged to local minimum and the Back Propagation train include problems, and that may converge in a local minimum rather than get a global minimum. The network structure appropriate for a given problem. In this paper, a good result is obtained by improve initial condition and adjust th number of hidden layer to solve the problem of real time process, learning and train.

• 대한원격탐사학회지
• /
• 제38권6_1호
• /
• pp.1329-1341
• /
• 2022

해양환경 및 해양생태계를 파괴하고 해양사고의 원인이 되는 해양쓰레기는 매년 늘어나고 있으나 그 중 해양침적쓰레기는 해저에 위치해 있어 파악과 수거에 어려움이 있다. 이에 효율적인 수거와 분포량 파악을 위해 수중촬영 이미지를 이용하여 폐그물과 폐밧줄을 대상으로 딥러닝 기반의 의미론적 분할을 실험하였다. 분할에는 최신 딥러닝 기법인 high-resolution network (HRNet)을 사용하고 최적화 알고리즘(optimizer) 별 성능 비교를 하였다. 분할 결과 그물에서는 adaptive moment estimation (Adam), Momentum, stochastic gradient descent(SGD) 순으로 F1 score=(86.46%, 86.20%, 85.29%), IoU=(76.15%, 75.74%, 74.36%) 이며, 밧줄은 F1 score=(80.49%, 80.48%, 77.86%), IoU=(67.35%, 67.33%, 63.75%)로 그물과 밧줄에서 모두 Adam의 결과가 가장 높게 나타났다. 연구 결과를 통해 optimizer 별 분할 성능 평가와 최신 딥러닝 기법의 해양침적쓰레기 분할에 대한 가능성을 확인하였다. 이에 따라 수중촬영 이미지를 통한 해양침적쓰레기 식별에 최신 딥러닝 기법을 적용시킴으로써 육안을 통한 식별보다 정확하고 효율적인 식별을 통해 해양침적쓰레기의 분포량 산정에 기여할 수 있을 것으로 사료된다.

• International Journal of Control, Automation, and Systems
• /
• 제6권5호
• /
• pp.755-766
• /
• 2008

This study presents a recurrent compensatory fuzzy neural network (RCFNN) for dynamic system identification. The proposed RCFNN uses a compensatory fuzzy reasoning method, and has feedback connections added to the rule layer of the RCFNN. The compensatory fuzzy reasoning method can make the fuzzy logic system more effective, and the additional feedback connections can solve temporal problems as well. Moreover, an online learning algorithm is demonstrated to automatically construct the RCFNN. The RCFNN initially contains no rules. The rules are created and adapted as online learning proceeds via simultaneous structure and parameter learning. Structure learning is based on the measure of degree and parameter learning is based on the gradient descent algorithm. The simulation results from identifying dynamic systems demonstrate that the convergence speed of the proposed method exceeds that of conventional methods. Moreover, the number of adjustable parameters of the proposed method is less than the other recurrent methods.

• International Journal of Computer Science & Network Security
• /
• 제23권10호
• /
• pp.115-128
• /
• 2023

Deep learning has been incorporating various optimization techniques motivated by new pragmatic optimizing algorithm advancements and their usage has a central role in Machine learning. In recent past, new avatars of various optimizers are being put into practice and their suitability and applicability has been reported on various domains. The resurgence of novelty starts from Stochastic Gradient Descent to convex and non-convex and derivative-free approaches. In the contemporary of these horizons of optimizers, choosing a best-fit or appropriate optimizer is an important consideration in deep learning theme as these working-horse engines determines the final performance predicted by the model. Moreover with increasing number of deep layers tantamount higher complexity with hyper-parameter tuning and consequently need to delve for a befitting optimizer. We empirically examine most popular and widely used optimizers on various data sets and networks-like MNIST and GAN plus others. The pragmatic comparison focuses on their similarities, differences and possibilities of their suitability for a given application. Additionally, the recent optimizer variants are highlighted with their subtlety. The article emphasizes on their critical role and pinpoints buttress options while choosing among them.

• 지능정보연구
• /
• 제22권2호
• /
• pp.127-142
• /
• 2016

딥러닝은 인공신경망(neural network)이라는 인공지능분야의 모형이 발전된 형태로서, 계층구조로 이루어진 인공신경망의 내부계층(hidden layer)이 여러 단계로 이루어진 구조이다. 딥러닝에서의 주요 모형은 합성곱신경망(convolutional neural network), 순환신경망(recurrent neural network), 그리고 심층신뢰신경망(deep belief network)의 세가지라고 할 수 있다. 그 중에서 현재 흥미로운 연구가 많이 발표되어서 관심이 집중되고 있는 모형은 지도학습(supervised learning)모형인 처음 두 개의 모형이다. 따라서 본 논문에서는 지도학습모형의 가중치를 최적화하는 기본적인 방법인 오류역전파 알고리즘을 살펴본 뒤에 합성곱신경망과 순환신경망의 구조와 응용사례 등을 살펴보고자 한다. 본문에서 다루지 않은 모형인 심층신뢰신경망은 아직까지는 합성곱신경망 이나 순환신경망보다는 상대적으로 주목을 덜 받고 있다. 그러나 심층신뢰신경망은 CNN이나 RNN과는 달리 비지도학습(unsupervised learning)모형이며, 사람이나 동물은 관찰을 통해서 스스로 학습한다는 점에서 궁극적으로는 비지도학습모형이 더 많이 연구되어야 할 주제가 될 것이다.