• 한국콘텐츠학회논문지
• /
• 제4권4호
• /
• pp.133-139
• /
• 2004

본 논문은 기계 학습 기법 중에서 메모리 기반 학습을 사용하여 범용의 학습 가능한 한국어 문장 경계 인식기를 제안한다. 제안한 방법은 메모리 기반 학습 알고리즘 중 최근린 이웃(kNN) 알고리즘을 사용하였으며, 이웃들을 이용한 문장 경계 결정을 위한 스코어 값 계산을 위한 다양한 가중치 방법을 적용하여 이들을 비교 분석하였다 문장 경계 구분을 위한 자질로는 특정 언어나 장르에 제한적이지 않고 범용으로 적용될 수 있는 자질만을 사용하였다. 성능 실험을 위하여 ETRI 코퍼스와 KAIST 코퍼스를 사용하였으며, 성능 척도로는 정확도와 재현율이 사용되었다. 실험 결과 제안한 방법은 적은 학습 코퍼스만으로도 $98.82\%$의 문장 정확률과 $99.09\%$의 문장 재현율을 보였다.

• 전기전자학회논문지
• /
• 제18권1호
• /
• pp.1-7
• /
• 2014

본 연구에서는 기존의 Viola-Jones 검출 프레임워크를 개선하여 하나의 특징 당 더 높은 효율을 가지며 검출대상이 아닌 서브 윈도우들을 더 빠르게 제거하는 개선된 학습 알고리즘을 제안한다. 학습의 결과로 생성된 물체 검출기는 서브윈도우를 특정 임계값까지 빠르게 제거하기 때문에 서브윈도우당 계산수가 줄어든다. 기존의 Viola-Jones 물체 검출기와 동일한 프레임워크이므로 검출 성능에는 영향을 주지 않는다. MIT-CMU 테스트 집합에 대해서 서브윈도우당 특징 계산 횟수를 측정하였으며 기존 계산 횟수의 45.5%로 줄어들어 검출 속도가 약 58.5% 향상됨을 확인하였다.

• 한국공작기계학회논문집
• /
• 제10권2호
• /
• pp.38-47
• /
• 2001

In this paper, a new scheme of adaptive-neuro control system is presented to implement real-time control of robot manipulator using Digital Signal Processors. Digital signal processors, DSPs, are micro-processors that are particularly developed for fast numerical computations involving sums and products of measured variables, thus it can be programmed and executed through DSPs. In addition, DSPs are as fast in computation as most 32-bit micro-processors and yet at a fraction of therir prices. These features make DSPs a viable computational tool in digital implementation of sophisticated controllers. Unlike the well-established theory for the adaptive control of linear systems, there exists relatively little general theory for the adaptive control of nonlinear systems. Adaptive control technique is essential for providing a stable and robust perfor-mance for application of robot control. The proposed neuro control algorithm is one of learning a model based error back-propagation scheme using Lyapunov stability analysis method.The proposed adaptive-neuro control scheme is illustrated to be a efficient control scheme for the implementation of real-time control of robot system by the simulation and experi-ment.

• 인터넷정보학회논문지
• /
• 제21권2호
• /
• pp.121-130
• /
• 2020

이 논문은 Generative Adversarial Network (GAN) 을 이용하여 증진된 이미지 데이터를 평가방식인 Inception Score (IS) 와 Frechet Inception Distance (FID) 계산시 inceptionV3 모델을 활용 하는 방식을 응용하여, 군 폐쇄망 네트워크 데이터를 이미지 형태로 평가하는 방법을 제안한다. 기존 존재하는 이미지 분류 모델들에 레이어를 추가하여 IncetptionV3 모델을 대체하고, 네트워크 데이터를 이미지로 변환 및 학습 하는 방법에 변화를 주어 다양한 시뮬레이션을 진행하였다. 실험 결과, atan을 이용해 8 * 8 이미지로 변환한 데이터에 대해 1개의 덴스 레이어 (Dense Layer)를 추가한 Densenet121를 학습시킨 모델이 네트워크 데이터셋 평가 모델로서 가장 적합하다는 결과를 도출하였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권6호
• /
• pp.2591-2611
• /
• 2020

Botnets have become one of the most significant threats to Internet-connected smartphones. A botnet is a combination of infected devices communicating through a command server under the control of botmaster for malicious purposes. Nowadays, the number and variety of botnets attacks have increased drastically, especially on the Android platform. Severe network disruptions through massive coordinated attacks result in large financial and ethical losses. The increase in the number of botnet attacks brings the challenges for detection of harmful software. This study proposes a smart framework for mobile botnet detection using static analysis. This technique combines permissions, activities, broadcast receivers, background services, API and uses the machine-learning algorithm to detect mobile botnets applications. The prototype was implemented and used to validate the performance, accuracy, and scalability of the proposed framework by evaluating 3000 android applications. The obtained results show the proposed framework obtained 98.20% accuracy with a low 0.1140 false-positive rate.

• 디지털산업정보학회논문지
• /
• 제9권2호
• /
• pp.111-119
• /
• 2013

Face recognition has recently become one of the most popular research areas in the fields of computer vision, machine learning, and pattern recognition because it spans numerous applications, such as access control, surveillance, security, credit-card verification, and criminal identification. In this paper, we propose a simple descriptor called an ECSP(Extended Center-Symmetric Pattern) for illumination-robust face recognition. The ECSP operator encodes the texture information of a local face region by emphasizing diagonal components of a previous CS-LBP(Center-Symmetric Local Binary Pattern). Here, the diagonal components are emphasized because facial textures along the diagonal direction contain much more information than those of other directions. The facial texture information of the ECSP operator is then used as the input image of an image covariance-based feature extraction algorithm such as 2D-PCA(Two-Dimensional Principal Component Analysis). Performance evaluation of the proposed approach was carried out using various binary pattern operators and recognition algorithms on the Yale B database. The experimental results demonstrated that the proposed approach achieved better recognition accuracy than other approaches, and we confirmed that the proposed approach is effective against illumination variation.

• 전기학회논문지P
• /
• 제65권2호
• /
• pp.108-113
• /
• 2016

The real-world problems usually show nonlinear and multi-variate characteristics, so it is difficult to establish concrete mathematical models for them. Thus, it is common to practice data-driven modeling techniques in these cases. Among them, most widely adopted techniques are regression model and intelligent model such as neural networks. Regression model has drawback showing lower performance when much non-linearity exists between input and output data. Intelligent model has been shown its superiority to the linear model due to ability capable of effectively estimate desired output in cases of both linear and nonlinear problem. This paper proposes modeling method of daily photovoltaic power systems using ELM(Extreme Learning Machine) based modular networks. The proposed method uses sub-model by fuzzy clustering rather than using a single model. Each sub-model is implemented by ELM. To show the effectiveness of the proposed method, we performed various experiments by dataset acquired during 2014 in real-plant.

• 전기학회논문지
• /
• 제65권6호
• /
• pp.1070-1079
• /
• 2016

In this paper, the design of recursive radial basis function neural networks based on incremental fuzzy c-means is introduced for processing the big data. Radial basis function neural networks consist of condition, conclusion and inference phase. Gaussian function is generally used as the activation function of the condition phase, but in this study, incremental fuzzy clustering is considered for the activation function of radial basis function neural networks, which could effectively do big data processing. In the conclusion phase, the connection weights of networks are given as the linear function. And then the connection weights are calculated by recursive least square estimation. In the inference phase, a final output is obtained by fuzzy inference method. Machine Learning datasets are employed to demonstrate the superiority of the proposed classifier, and their results are described from the viewpoint of the algorithm complexity and performance index.

• 대한산업공학회지
• /
• 제41권6호
• /
• pp.540-550
• /
• 2015

There have been many attempts to find knowledge from data using conventional statistics, data mining, artificial intelligence, machine learning and pattern recognition. In those research areas, knowledge is approached in two ways. Firstly, researchers discover knowledge represented in general features for universal recognition, and secondly, they discover exceptional and distinctive features. In process mining, an instance is sequential information bounded by case ID, known as process instance. Here, an exceptional process instance can cause a problem in the analysis and discovery algorithm. Hence, in this paper we develop a method to detect the knowledge of exceptional and distinctive features when performing process mining. We propose a method for anomaly detection named Distance-based Anomaly Process Instance Detection (DAPID) which utilizes distance between process instances. DAPID contributes to a discovery of distinctive characteristic of process instance. For verifying the suggested methodology, we discovered characteristics of exceptional situations from log data. Additionally, we experiment on real data from a domestic port terminal to demonstrate our proposed methodology.

• 한국공작기계학회:학술대회논문집
• /
• 한국공작기계학회 1999년도 추계학술대회 논문집 - 한국공작기계학회
• /
• pp.162-169
• /
• 1999

In this paper, it is presented a new scheme of adaptive-neuro control system to implement real-time control of robot manipulator using digital Signal Processors. Digital signal processors DSPs. are micro-processors that are particularly developed for variables. Digital version of most advanced control algorithms can be defined as sums and products of measured variables, thus it can be programmed and executed through DSPs. In addition, DSPs are as fast in computation as most 32-bit micro-processors and yet at a fraction of their prices. These features make DSPs a biable computatinal tool in digital implementation of sophisticated controllers. Unlike the well-established theory for the adaptive control of linear systems, there exists relatively little general theory for the adaptive control of nonlinear systems. Adaptive control technique is essential for providing a stable and robust performance for application of robot control. The proposed neuro control algorithm is one of learning a model based error back-propagation scheme using Lyapunov stability analysis method. The proposed adaptive-neuro control scheme is illustrated to be a efficient control scheme for implementation of real-time control of robot system by the simulation and experiment.