• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1995.10a
• /
• pp.124-140
• /
• 1995

The merit of fuzzy rule based systems stems from their capability of encoding qualitative knowledge of experts into quantitative rules. Recent advancement in automatic tuning or self-organization of fuzzy rules from experimental data further enhances their power, allowing the integration of the top-down encoding of knowledge with the bottom-up learning of rules. In this paper, methods of self-organizing fuzzy rules and of performing defuzzification and inference is presented based on a multi-resolution radial basis function network. The network learns an arbitrary input-output mapping from sample distribution as the union of hyper-ellipsoidal clusters of various locations, sizes and shapes. The hyper-ellipsoidal clusters, representing fuzzy rules, are self-organized based of global competition in such a way as to ensute uniform mapping errors. The cooperative interpolation among the multiple clusters associated with a mapping allows the network to perform a bidirectional many-to-many mapping, representing a particular from of defuzzification. Finally, an inference engine is constructed for the network to search for an optimal chain of rules or situation transitions under the constraint of transition feasibilities imposed by the learned mapping. Applications of the proposed network to skill acquisition are shown.

• 제어로봇시스템학회:학술대회논문집
• /
• 1996.10b
• /
• pp.936-939
• /
• 1996

In this paper, two specially designed associative mapping memories, called Associative Mapping Elements(AME) and Multiple-Digit Overlapping AME(MDO-AME), are presented for learning of nonlinear functions including kinematics and dynamics of robot manipulators. The proposed associative mapping memories consist of associative mapping rules(AMR) and weight update rules(WUR) which guarantee generalization and specialization of input-output relationship of learned nonlinear functions. Two simulation results, one for supervised learning and the other for unsupervised learning, are given to demonstrate the effectiveness of the proposed associative mapping memories.

• Journal of KIISE:Software and Applications
• /
• v.41 no.7
• /
• pp.501-517
• /
• 2014

Binary translation is a restructuring process in order to execute a program targeting a specific device on the other devices. In binary translation, it is very important to generate the translation rules between two devices. There are two methods for generating the translation rules, direct and indirect mapping. The direct mapping is the method for performance, while the indirect mapping is the method for retargetability. This paper suggests a binary translation method based on the direct mapping for the embedded systems. Because, however, the retargetability is also important requirement, we suggest the direct mapping based binary translation with considering the retargetability. In addition, we implement an automatic generation tool for translation rules to prove our concept. Through this method, we can generate the translation rules with considering the performance as well as the retargetability. Furthermore, we can reduce costs for the binary translation.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1993.06a
• /
• pp.1266-1269
• /
• 1993

The paper introduces a new method for fuzzy processing. The method allows handing a piece of information lost in the classic fuzzification process, and thus neglected by other methods. Processing the result after fuzzification is sustained by the interpretation that the input-output set mapping, specified by the IF-THEN rules, can be regarded as a direct mapping of their corresponding alpha-cuts. Processing involves just singletons as intermediary results, the final result being a combination of singletons obtained from fired rules.

• The Transactions of the Korea Information Processing Society
• /
• v.6 no.12
• /
• pp.3539-3547
• /
• 1999

Korean linguistic information for the generation modulo of KANT(Knowledge-based Accurate Natural language Translation) system was constructed. As KANT has a language-independent generation engine, the construction of Korean linguistic information means the development of the Korean generation module. Constructed information includes concept-based mapping rules, category-based mapping rules, syntactic lexicon, template rules, grammar rules based on the unification grammar, lexical rules and rewriting rules for Korean. With these information in sentences were successfully and completely generated from the interlingua functional structures among the 118 test set prepared by the developers of KANT system.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.7
• /
• pp.35-47
• /
• 2011

This paper proposes a set of rules to automatically generate OWL ontologies from relational databases. The purpose of the rules is to allow semantic access to existing RDB data without any database schema transformation and data migration process. In other words, the rules help a RDBMS play as a web ontology repository as well. However, the use of the mapping rules between RDB and OWL proposed by other studies for the objective causes troubles as follows. First, databases including the tables with a specific structure can't be translated into OWL. Second, the process for extracting an OWL individual unnecessarily lead to database join operations, or several SQL queries. On the other hand, our rules is designed to prevent these problems, can generate OWL classes and properties from database schemas and can generate OWL individuals from the database instances. In addition, an ontology generated by our rules is an OWL 2 DL ontology.

• Journal of information and communication convergence engineering
• /
• v.14 no.2
• /
• pp.97-105
• /
• 2016

With the increasing demand for interoperability among existing learning resource systems in order to enable the sharing of learning resources, such resources need to be annotated with ontologies that use different metadata standards. These different ontologies must be reconciled through ontology mediation, so as to cope with information heterogeneity problems, such as semantic and structural conflicts. In this paper, we propose an ontology-mapping technique using Semantic Web Rule Language (SWRL) to generate semantic mapping rules that integrate learning resources from different systems and that cope with semantic and structural conflicts. Reasoning rules are defined to support a semantic search for heterogeneous learning resources, which are deduced by rule-based inference. Experimental results demonstrate that the proposed approach enables the integration of learning resources originating from multiple sources and helps users to search across heterogeneous learning resource systems.

• The Journal of the Acoustical Society of Korea
• /
• v.40 no.2
• /
• pp.161-168
• /
• 2021

In this paper, a Silent Speech Interface (SSI) technology was proposed in which Doppler frequency shifts of the reflected signal were used to synthesize the speech signals when 40kHz ultrasonic signal was incident to speaker's mouth region. In SSI, the mapping rules from the features derived from non-speech signals to those from audible speech signals was constructed, the speech signals are synthesized from non-speech signals using the constructed mapping rules. The mapping rules were built by minimizing the overall errors between the estimated and true speech parameters in the conventional SSI methods. In the present study, the mapping rules were constructed so that the distribution of the estimated parameters is similar to that of the true parameters by using Generative Adversarial Networks (GAN). The experimental result using 60 Korean words showed that, both objectively and subjectively, the performance of the proposed method was superior to that of the conventional neural networks-based methods.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.35S no.4
• /
• pp.118-127
• /
• 1998

In this paper, we propose a voice personality transformation method which makes one person's voice sound like another person's voice. In order to transform the voice personality, vocal tract transfer function is used as a transformation parameter. Comparing with previous methods, the proposed method can obtain high-quality transformed speech with low computational complexity. Conversion between the vocal tract transfer functions is implemented by a linear mapping based on soft clustering. In this process, mean LPC cepstrum coefficients and mean removed LPC cepstrum modeled by the low dimensional vector are used as transformation parameters. To evaluate the performance of the proposed method, mapping rules are generated from 61 Korean words uttered by two male and one female speakers. These rules are then applied to 9 sentences uttered by the same persons, and objective evaluation and subjective listening tests for the transformed speech are performed.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2019.01a
• /
• pp.369-372
• /
• 2019

세계 주요 자동차 회사들은 효율적인 차량용 소프트웨어 개발을 위해 AUTOSAR 표준을 필수로 적용하고 있다. AUTOSAR 기반 소프트웨어의 기능은 러너블(runnable) 단위로 구현되며 이는 태스크에 매핑되어 동작하는데, 러너블-태스크 매핑은 시스템 오버헤드 발생과 러너블의 실제 수행 시점에 크게 영향을 미치므로 시스템 성능 측면에서 매우 중요한 작업이다. 본 논문에서는 자동차의 제어를 보조하는 타겟 응용에 대하여 최적의 성능을 보이는 러너블-태스크 매핑을 찾고자 기존 연구에서 제안된 6개의 매핑 규칙을 적용하며, 기존 규칙의 한계점을 개선한 매핑 규칙을 제안하여 추가로 적용한다. Infineon 사의 AURIX 보드와 ETAS 사의 AUTOSAR 플랫폼 상에 타겟 응용을 구현하여 실험한 결과, 기존 매핑 규칙에 비해 개선된 규칙을 적용하였을 때 종단 간 응답시간이 21.23% 단축되었다.