• 제어로봇시스템학회:학술대회논문집
• /
• 2003.10a
• /
• pp.1309-1314
• /
• 2003

If an agent has a learning ability using previous knowledge, then it is expected that the agent can speed up learning by interacting with environment. In this paper, we present an improved reinforcement learning algorithm using domain knowledge which can be represented by problem-independent features and their classifiers. Here, neural networks are employed as knowledge classifiers. To show the validity of our proposed algorithm, computer simulations are illustrated, where navigation problem of a mobile robot and a micro aerial vehicle(MAV) are considered.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.09a
• /
• pp.216-219
• /
• 2003

Among the techniques to protect private information by adopting biometrics, speaker verification is expected to be widely used due to advantages in convenient usage and inexpensive implementation cost Speaker verification should achieve a high degree of the reliability in the verification nout the flexibility in speech text usage, and the efficiency in verification system complexity. Continuants have excellent speaker-discriminant power and the modest number of phonemes in the category, and multilayer perceptrons (MLPs) have superior recognition ability and fast operation speed. In consequence, the two provide viable ways for speaker verification system to obtain the above properties. This paper implements a system to which continuants and MLPs are applied, and evaluates the system using a Korean speech database. The results of the experiment prove that continuants and MLPs enable the system to acquire the three properties.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.09a
• /
• pp.492-495
• /
• 2003

Fuzzy Min-Max Neural Network(FMMNN) is a powerful classifier, It has, however, some problems. Learning result depends on the presentation order of input data and the training parameter that limits the size of hyperbox. The latter problem affects the result seriously. In this paper, the new approach to alleviate that without loss of on-line learning ability is proposed. The committee machine is used to achieve the multi-resolution FMMNN. Each expert is a FMMNN with fixed training parameter. The advantages of small and large training parameters are used at the same time. The parameters are selected by performance and independence measures. The Decision of each expert is guided by the gating network. Therefore the regional and parametric divide and conquer scheme are used. Simulation shows that the proposed method has better classification performance.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.09a
• /
• pp.179-182
• /
• 2003

This paper proposes an adaptive hybrid genetic algorithm (aHGA) for effectively solving the complex reliability optimization problems. The proposed aHGA uses a loca1 search technique and an adaptive scheme for respectively constructing hybrid algorithm and adaptive ability. For more various comparisons with the proposed adaptive algorithm, other aHGAs with conventional adaptive schemes are considered. These aHGAs are tested and analyzed using two complex reliability optimization problems. Numerical result shows that the proposed aHGA outperforms the other competing aHGAs.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.09a
• /
• pp.459-462
• /
• 2003

Selectivity estimation of queries not only provides useful information to the query processing optimization but also may give users with a preview of processing results. In this paper, we investigate the problem of selectivity estimation in the context of a spatial dataset. Although several techniques have been proposed in the literature to estimate spatial query result sizes, most of those techniques still have some drawback in the case that a large amount of memory is required to retain accurate selectivity. To eliminate the drawback of estimation techniques in previous works, we propose a new method called MW Histogram. Our method is based on two techniques: (a) MinSkew partitioning algorithm that processes skewed spatial datasets efficiently (b) Wavelet transformation which compression effect is proven. We evaluate our method via real datasets. With the experimental result, we prove that the MW Histogram has the ability of providing estimates with low relative error and retaining the similar estimates even if memory space is small.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.09a
• /
• pp.384-387
• /
• 2003

Most of the real world phenomena change over time. The ability to represent and to reason geographic data becomes crucial. A large amount of non-standard applications are dealing with data characterized by spatial, temporal and/or uncertainty features. Non-standard data like spatial and temporal data have an inner complex structure requiring sophisticated data representation, and their operations necessitate sophisticated and efficient algorithms. Current GIS technology is inefficient to model and to handle complex geographic phenomena, which involve space, time and uncertainty dimensions. This paper concentrates on developing a fuzzy spatio-temporal data model based on fuzzy set theory and relational data models. Fuzzy spatio-temporal operators are also provided to support dynamic query.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2000.11a
• /
• pp.243-251
• /
• 2000

In general, Rough Set theory is used for classification, inference, and decision analysis of incomplete data by using approximation space concepts in information system. Information system can include quantitative attribute values which have interval characteristics, or incomplete data such as multiple or unknown(missing) data. These incomplete data cause the inconsistency in information system and decrease the classification ability in system using Rough Sets. In this paper, we present various types of incomplete data which may occur in information system and propose INcomplete information Processing System(INiPS) which converts incomplete information system into complete information system in using Rough Sets.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2000.11a
• /
• pp.149-152
• /
• 2000

This paper addresses the wavelet fuzzy modeling using Virus-Evolutionary Genetic Algorithm (VEGA). We build a fuzzy system model which is equivalent to the wavelet transform after identifying the coefficients of wavelet transform. We can obtain an accurate system model with a small number of coefficients due to the energy compaction property of the wavelet transform. It thus means that we can construct a fuzzy system model with a small number of rules. In order to identify the wide-ranged coefficients of the wavelet transform, VEGA is adopted, which has prominent ability to avoid premature local convergence that is suitable to complex optimization problems. We demonstrate the superiority of our proposed fuzzy system modeling method over the previous results by modeling nonlinear function.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1998.06a
• /
• pp.143-147
• /
• 1998

It sis wisely stated that the most valuable knowledge that a person can acquire is the knowledge of how to learn. The human's learning is characterized by the ability to extract relationships between the different characters of a given situation . The ellipse is a first approach of comparison. We assimilate each character to a half axis of the ellipse and the result is a geometrical figure that varies according to values of the two characters. Thus, we take into account the two characters as an alone entity.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1998.06a
• /
• pp.153-157
• /
• 1998

It is wisely stated that the most valuable knowledge that a person cam acquire is the knowledge of how to learn. The human's learning is characterized by the ability to extract relationships between the different characters of a given situation. The ellipse is a first approach of comparison. We assimilate each character to a half axis of the ellipse and the result is a geometrical figure that varies according to values of the two characters. Thus, we take into account the two characters as an alone entity.