• Journal of Digital Contents Society
• /
• v.2 no.2
• /
• pp.103-115
• /
• 2001

This paper proposes the two algorithms which generate a minimal decision rule and approximate inference operation, adapted the rough set and the factor space theory in fuzzy knowledge base. The generation of the minimal decision rule is executed by the data classification technique and reduct applying the correlation analysis and the Bayesian theorem related attribute factors. To retrieve the specific object, this paper proposes the approximate inference method defining the membership function and the combination operation of t-norm in the minimal knowledge base composed of decision rule. We compare the suggested algorithms with the other retrieval theories such as possibility theory, factor space theory, Max-Min, Max-product and Max-average composition operations through the simulation generating the object numbers and the attribute values randomly as the memory size grows. With the result of the comparison, we prove that the suggested algorithm technique is faster than the previous ones to retrieve the object in access time.

• Journal of Korean Society for Quality Management
• /
• v.21 no.2
• /
• pp.156-161
• /
• 1993

Limiting distributions of Studentized test statistics have been shown for testing the slope parameter in a simple linear structural model. Since the limiting distribution of Studentized one appears to yield inaccurate inference, this paper suggests adjustment of critical value and normalization of the Studentized one. As results, we can have procedures for refined inference based on our approximate distrbution instead of the limiting distribution.

• Journal of the Korean Society of Industry Convergence
• /
• v.9 no.3
• /
• pp.173-178
• /
• 2006

For the design of multivariable fuzzy control systems decomposition of control rules is a efficent inference method since it alleviates the complexity of the problem. In some systems, however, inference error of the Gupta's decomposition method is inevitable because of its approximate nature. In this paper we define indices of applicability which decides whether the decomposition method can be applied to a multivariable fuzzy system or not.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1993.06a
• /
• pp.1141-1144
• /
• 1993

For the design of multivariable fuzzy control systems the decomposition of control rules is preferable since it alleviates the complexity of the problem. In some systems, however, inference error of the Gupta's decomposition method is inevitable because of its approximate nature. In this paper, we propose a new multivariable fuzzy controller with a coordinator which can reduce the inference error of the decomposition method by using an index of applicability.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1998.06a
• /
• pp.719-724
• /
• 1998

This paper presents a systematic developement of a formal approach to inference in approximate reasoning. We introduce some measures of similarity and discuss their properties. Using the concept of similarity index we formulate two methods for inferring from vague knowledge. In order to illustrate the effectiveness of the proposed technique we use it to develop a vowel recognition system.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.23 no.1
• /
• pp.143-159
• /
• 1998

This paper proposes a new type of fuzzy approximate reasoning method that combines a self organizing feature map and a fuzzy logic. Previous methods considered only input part to determine the number of fuzzy rules, while this paper considers both input and output parts simultaneously. Our approach proved to improve the inference performance. We also developed a new index for avoiding overlearning which guarantees more accurate results. Experimental results showed that our approach surpasses the performance of Takagi & Hayashi (1991) approach.

• International Journal of Reliability and Applications
• /
• v.8 no.1
• /
• pp.1-16
• /
• 2007

In this paper, we derive exact explicit expressions for the triple and quadruple moments of the lower record values from inverse the Weibull (IW) distribution. Next, we present and calculate the coefficients of the best linear unbiased estimates of the location and scale parameters of IW distribution (BLUEs) for different choices of the shape parameter and records size. We then use the higher order moments and the calculated BLUEs to compute the mean, variance, and the coefficients of skewness and kurtosis of certain linear functions of lower record values. By using the coefficients of the skewness and kurtosis, we develop approximate confidence intervals for the location and scale parameters of the IW distribution using Edgeworth approximate values and then compare them with the corresponding intervals constructed through Monte Carlo simulations. Finally, we apply the findings of the paper to some simulated data.

• Journal of Semiconductor Engineering
• /
• v.1 no.1
• /
• pp.31-37
• /
• 2020

We present an approximate DRAM architecture for energy-efficient deep learning. Our key premise is that by bounding memory errors to non-critical information, we can significantly reduce DRAM refresh energy without compromising recognition accuracy of deep neural networks. To validate the key premise, we make extensive Monte-Carlo simulations for several well-known convolutional neural networks such as LeNet, ConvNet and AlexNet with the input of MINIST, CIFAR-10, and ImageNet, respectively. We assume that the highest-order 8-bits (in single precision) and 4-bits (in half precision) are protected from retention errors under the proposed architecture and then, randomly inject bit-errors to unprotected bits with various bit-error-rates. Here, recognition accuracies of the above convolutional neural networks are successfully maintained up to the 10^-5-order bit-error-rate. We simulate DRAM energy during inference of the above convolutional neural networks, where the proposed architecture shows the possibility of considerable energy saving up to 10 ~ 37.5% of total DRAM energy.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1998.10a
• /
• pp.532-538
• /
• 1998

In this paper, a new adaptive fuzzy inference method using neural network based fuzzy reasoning is proposed to make a fuzzy logic control system more adaptive and more effective. In most cases, the design of a fuzzy inference system rely on the method in which an expert or a skilled human operator would operate in that special domain. However, if he has not expert knowledge for any nonlinear environment, it is difficult to control in order to optimize. Thus, using the proposed adaptive structure for the fuzzy reasoning system can controled more adaptive and more effective in nonlinear environment for changing input membership functions and output membership functions. The proposed fuzzy inference algorithm is called adaptive neuro-fuzzy control(ANFC). ANFC can adapt a proper membership function for nonlinear plant, based upon a minimum number of rules and an initial approximate membership function. Nonlinear function approximation and rotary inverted pendulum control system ar employed to demonstrate the viability of the proposed ANFC.

• Korean Journal of Cognitive Science
• /
• v.10 no.4
• /
• pp.71-83
• /
• 1999

Conventional symbolic inference systems lack flexibility because they do not well reflect flexible semantic structure of knowledge and use symbolic logic for their basic inference mechanism. For solving this problem. we have recently proposed the 'Connectionist Semantic Network(CSN)' as a model for flexible knowledge representation and inference based on neural networks. The CSN is capable of carrying out both approximate reasoning and commonsense reasoning based on similarity and association. However. we have difficulties in representing general and structured high-level knowledge and variable binding using the connectionist framework of the CSN. In this paper. we propose a hybrid system called SymCSN(Symbolic CSN) that combines a symbolic module for representing general and structured high-level knowledge and a connectionist module for representing and learning low-level semantic structure Simulation results show that the SymCSN is a plausible model for human-like flexible knowledge representation and inference.