• 인지과학
• /
• 제26권4호
• /
• pp.377-392
• /
• 2015

한국어 의미역 결정(Semantic Role Labeling)은 주로 기계학습에 의해 이루어지며 많은 말뭉치 자원을 필요로 한다. 그러나 한국어 의미역 결정 시스템에 사용되는 Korean PropBank는 의미역이 부착된 용언과 용언 격틀이 PropBank에 비해 각각 1/5, 1/2 수준에 불과하다. 따라서 본 논문에서는 한국어 의미역 결정 시스템을 위해 의미역이 부착된 용언과 용언 격틀을 확장하여 Korean PropBank를 확장 시키고자 한다. 대부분의 의미역 결정 시스템은 학습 도메인에 의존적이기 때문에 적용 도메인 변경에 따른 성능 하락이 나타날 수 있다. 본 논문에서는 기존의 학습 말뭉치와 적은 양의 새로운 학습 말뭉치를 활용하여 새로운 도메인에 대해 의미역 결정 시스템의 성능 하락을 최소화 할 수 있는 도메인 적응 기술을 Structural SVM(S-SVM)과 Deep Neural Network(DNN) 기반 한국어 의미역 결정 시스템에 적용하여 그 실효성을 알아보고자 한다.

• 대한산업공학회지
• /
• 제27권3호
• /
• pp.239-249
• /
• 2001

Competitive coevolution models, often called host-parasite models, are searching models that imitate the biological coevolution that is a series of reciprocal changes in two competing species. The models are known to be an effective method of solving complex and dynamic problems such as game problems, neural network design problems and constraint satisfaction problems. However, previous models consider only ectoparasites that live on the outside of the host when designing the models, not considering endoparasites that live on the inside of the host. This has a limitation to exploiting some information. In this paper, we develop an artificial adaptation model simulating the process in which hosts coevolve with both ectoparasites and endoparasites. In the model, the endoparasites play important roles as follows. By means of them, we can keep the history on results of previous competition between hosts and parasites, and use endogeneous fitness, not exogeneous. Extensive experiments are carried out to show the coevolution phenomenon and to verify the performance of the proposed model. Nim game problems and neural network problems are used as test-bed problems. The results are reported in this paper.

• 전자공학회논문지CI
• /
• 제42권3호
• /
• pp.17-28
• /
• 2005

최근 컴퓨터 게임에서 지능 캐릭터는 게이머들의 흥미를 계속 유발시킬 수 있기 때문에 더욱 더 중요한 요소로 부각되고 있다. 지능 캐릭터를 구현하는 대표적인 방법으로 신경망을 사용하여 상대 캐릭터의 행동패턴과 게임 규칙을 학습하는 방법이 연구되었다. 그러나 게임의 규칙은 갑자기 변경될 수 있으며 온라인 게임과 같은 상황에서는 게이머에 따라서 행동 특성이 크게 다를 수 있다. 본 논문에서는 지능 캐릭터가 이러한 환경의 변화에 적응하는 방법으로서 개체 수준 적응 알고리즘과 개체군 수준 적응 알고리즘을 제안한다. 개체 수준 적응 알고리즘에서 각 지능 캐릭터는 자신의 게임 점수의 변화를 계속해서 관찰하면서, 최종적으로 획득한 점수들을 고려하여 환경의 변화를 판단하고, 만약 변화가 감지된 경우에는 다시 새로운 학습을 시작한다. 대용량 온라인 게임과 같이 다수의 사용자가 있는 게임에서는 지능 캐릭터들이 다양한 행동 패턴과 전략을 가지고 있는 여러 상대 캐릭터들과 대전한다. 개체군 수준 적응 알고리즘은 유전자 알고리즘을 이용하여 지능 캐릭터들의 출현을 조절하여 게임 월드내의 균형이 유지되도록 한다. 제안한 알고리즘의 성능을 평가하기 위하여 간단한 대전 액션 게임을 구현하고 그 환경 상에서 게임 규칙과 상대 캐릭터들의 행동 패턴을 변화시키면서 실험하였다. 실험 결과 지능 캐릭터는 제안한 기법을 이용하여 환경 변화에 적응할 수 있음을 보였다.

• 말소리와 음성과학
• /
• 제9권2호
• /
• pp.85-94
• /
• 2017

Conventional speaker verification systems using time delay neural network, identity vector and probabilistic linear discriminant analysis (TDNN-Ivector-PLDA) are known to be very effective for verifying long-duration speech utterances. However, when test utterances are of short duration, duration mismatch between enrollment and test utterances significantly degrades the performance of TDNN-Ivector-PLDA systems. To compensate for the I-vector mismatch between long and short utterances, this paper proposes to use probabilistic linear discriminant analysis (PLDA) model adaptation with augmented data. A PLDA model is trained on vast amount of speech data, most of which have long duration. Then, the PLDA model is adapted with the I-vectors obtained from short-utterance data which are augmented by using vocal tract length perturbation (VTLP). In computer experiments using the NIST SRE 2008 database, the proposed method is shown to achieve significantly better performance than the conventional TDNN-Ivector-PLDA systems when there exists duration mismatch between enrollment and test utterances.

• 전기학회논문지
• /
• 제64권9호
• /
• pp.1356-1362
• /
• 2015

Modular snake-like robots are robust for failure and have flexible locomotions for obstacle environment than of walking robot. This requires an adaptation capability which is obtained from a learning approach, but has not been analysed as well. In order to investigate the property of adaptation of locomotion for different terrains, NEAT controllers are trained for a flat terrain and tested for obstacle terrains. The input and output characteristics of the adaptation for the neural network controller are analyzed for different terrains in frequency domain.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2000년도 제15차 학술회의논문집
• /
• pp.362-362
• /
• 2000

The purpose of this thesis is to reduce force ripple of linear pulse motor(LPM) using neural network and to enhance precision. In order to this, we propose a new controller using a neural network to compensate disturbances. The structure includes adaptation block which learns the dynamics of the periodic disturbance and forces the interferences, caused by disturbances. The proposed controller compensates an unmodeled dynamics in the LPM. The neural network changes a current command to reduce position error and force ripple of the LPM. We compare proposed controller with PI controller. Simulation result shows that the proposed controller has better performance than a PI controller without neural network.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 1996년도 추계학술대회 학술발표 논문집
• /
• pp.27-34
• /
• 1996

GENIE is a learning-based engine for building intelligent systems. Learning in GENIE proceeds by incrementally modeling its human or technical environment using a neural network and a genetic algorithm. The neural network is used to represent the knowledge for solving a given task and has the ability to grow its structure. The genetic algorithm provides the neural network with training examples by actively exploring the example space of the problem. Integrated into the training examples by actively exploring the example space of the problem. Integrated into the GENIE system architecture, the genetic algorithm and the neural network build a virtually self-teaching autonomous learning system. This paper describes the structure of GENIE and its learning components. The performance is demonstrated on a robot learning problem. We also discuss the lessons learned from experiments with GENIE and point out further possibilities of effectively hybridizing genetic algorithms with neural networks and other softcomputing techniques.

• 한국경영과학회:학술대회논문집
• /
• 한국경영과학회 2005년도 추계학술대회 및 정기총회
• /
• pp.237-240
• /
• 2005

This paper describer a design expert system which determines the design values of shadow mask using Case-Based Reasoning. In Case-Based Reasoning, it is important to both retrieve similar cases and adapt the cases to meet the design specifications exactly. Especially, the difficulty in automating the adaptation process will prevent the designers from using the design expert systems efficiently and easily. This paper explains knowledge-based design support systems for shadow mask through neural network-based case adaptation. Specifically, we developed 1) representing design knowledge and 2) adaptive case-based reasoning method using RBFN (Radial Basis Function Network).

• 대한전기학회논문지:전기기기및에너지변환시스템부문B
• /
• 제53권9호
• /
• pp.569-575
• /
• 2004

This paper is proposed adaptive fuzzy-neural network(FNN) controller for high performance of induction motor drive. The design of this algorithm based on FNN controller that is implemented using fuzzy control and neural network. This controller uses fuzzy rule as training patterns of a neural network. Also, this controller uses the back-propagation method to adjust the weights between the neurons of neural network in order to minimize the error between the command output and actual output. A model reference adaptive scheme is proposed in which the adaptation mechanism is executed by fuzzy logic based on the error and change of error measured between the motor speed and output of a reference model. The control Performance of the adaptive FNN controller is evaluated by analysis for various operating conditions. The results of analysis prove that the proposed control system has strong high performance and robustness to parameter variation. and steady- state accuracy and transient response.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1993년도 한국자동제어학술회의논문집(국내학술편); Seoul National University, Seoul; 20-22 Oct. 1993
• /
• pp.14-19
• /
• 1993

In this paper, we proposed an indirect learning and direct adaptive control schemes using neural networks, i.e., composite adaptive neural control, for a class of continuous nonlinear systems. With the indirect learning method, the neural network learns the nonlinear basis of the system inverse dynamics by a modified backpropagation learning rule. The basis spans the local vector space of inverse dynamics with the direct adaptation method when the indirect learning result is within a prescribed error tolerance, as such this method is closely related to the adaptive control methods. Also hash addressing technique, similar to the CMAC functional architecture, is introduced for partitioning network hidden nodes according to the system states, so global neuro control properties can be organized by the local ones. For uniform stability, the sliding mode control is introduced when the neural network has not sufficiently learned the system dynamics. With proper assumptions on the controlled system, global stability and tracking error convergence proof can be given. The performance of the proposed control scheme is demonstrated with the simulation results of a nonlinear system.