Earticle

최근 은행, 공공기관 등에서 사용되는 고전적인 소프트웨어에서 PC 해킹 및 악성코드 유포가 이루어지는 등 보안위협이 제기되고 있다. C 스타일로 제작되어 프로그램에 자체적으로 내장된 취약점의 특성상 이로 인한 해킹의 연쇄효과는 막대하며 특정 서비스 이용을 위해 반드시 설치 해야하는 프로그램인 경우 그 파급효과가 매우 크다는 점에서 개인정보 유출과 금융 보안에 매 우 중요한 영향을 미친다. 이에 본 논문에서는 이러한 프로그램 자체에 내장되어있는 보안 취약 점으로 인한 시스템 해킹 공격기법들의 특성과 feature를 도출하여 AI의 학습에 활용될 수 있는 기반기술을 제안하고자한다. 이에 더 나아가 취약점에 따른 공격(Exploitation) 템플릿들을 자동 생성 할 수 있는 시스템을 제안하고자 한다.

Recently, security threats such as PC hacking and malicious code distribution are being raised in classic software used in banks and public institutions. Due to the characteristics of vulnerabilities built in the C style and built into the program itself, the chain effect of hacking caused by this is enormous, and in the case of programs that must be installed to use a specific service, the ripple effect is very large, which is very important for personal information leakage and cyber security. have a significant impact In this paper, we propose a basic technology that can be used for AI learning by deriving the characteristics and features of system hacking attack techniques due to security vulnerabilities built into the program itself. Furthermore, we propose a system that can automatically generate exploit templates according to vulnerabilities.

최근 기계학습에서 판별 기술이 다양한 문제 해결에 활용되고 있다. 하지만 우리가 실제 다루고 있는 데이터셋들은 클래스 불균형이 많이 등장하고 있어 샘플 바이어스 등 왜곡된 결과가 도출되는 것을 해결하기 위해 다양한 방법들이 제안되고 있다. 그러나 어떤 클래스 불균형의 비율 및 불균형 완화 방법 별로 어떤 판별 알고리즘이 우수한지를 규명한 연구는 거의 존재하지 않는다. 따라서 본 연구의 목적은 다양한 클래스 불균형 상황에서 어떤 불균형 해소 방법과 classification algorithm의 적합성이 판별 성능을 극대화하는지를 파악하는 것이다. 이를 위해 본 연구에서는 UCI Repository에서 제공하는 벤치마크 데이터셋을 대상으로 클래스 불균형 문제 해소에 사용되는 방법을 각 대표적인 classification algorithm 에 적용하여 그 성능을 비교하자 한다. 본 연구에서는 과대표집법(over sampling)에 집중하여 고찰해보고자 한다.

In order to address the issue of low accuracy rate of current orchid type classification methods due to their similarities in the characteristics of orchid types, an effective orchid type classification method using data enhancement is suggested in this work, whose contribution depends on the utilization of data enhancement technologies, which can efficiently enhance the orchid type classification accuracy rate by providing sufficient and balanced sample sets. Specifically, in our approach, firstly, an image set of 12 orchid types containing 12,227 images is established; secondly, the characteristics of the above orchid image dataset are analyzed and studied; thirdly, the reasons for the processing difficulties are identified based on the above orchid image set; at last, some data enhancement technologies are applied to improve the classification accuracy rate of orchid types, which can also enhance the whole performance of orchid type classification. The experimental results display that our suggested classification method using data enhancement in the article can achieve a classification accuracy of 92.65% compared with the one not using data enhancement under the condition of insufficient and unbalanced image datasets.

Gait-based classification has gained much interest as a possible authentication method because it incorporate an intrinsic personal signature that is difficult to mimic. The study investigates machine learning techniques to mitigate the natural variations in gait among different subjects. We incorporated several machine learning algorithms into this study using the data mining package called Waikato Environment for Knowledge Analysis (WEKA). WEKA's convenient interface enabled us to apply various sets of machine learning algorithms to understand whether each algorithm can capture certain distinctive gait features. First, we defined 24 gait features by analyzing three-axis acceleration data, and then selectively used them for distinguishing subjects 10 years of age or younger from those aged 20 to 40. We also applied a machine learning voting scheme to improve the accuracy of the classification. The classification accuracy of the proposed system was about 81% on average.

The present research article focused on the factual findings of the potential usage of the combinational Feed-Forward Back Propagation Neural Network as a judgment making for lung cancer. In this context, Support Vector Machine is integrated with Feed-Forward Back Propagation Neural Network to create a hybrid algorithm that further helps in reducing the computation complexity of the classification. A set of 500 images are utilized in which 75% data is used for the training purpose and the rest 25% is used to achieve the classification. In the view of forgoing, a three-block mechanism is proposed for the classification in which the first block preprocesses the dataset, the second block extracts the features via the SURF technique followed by the optimization using Genetic Algorithm and the terminal block is for the classification via FFBPNN. The hybrid classification algorithm is named as Kernel Attribute Selected Classifier and the overall classification accuracy of the proposed algorithm is 98.08%. Herein, the objective of the study is to enhance the classification accuracy by applying a hybrid classification algorithm.

In this paper, an improved one-against-one support vector machine algorithm is used to classify multiple power quality disturbances. To solve the problem of parameter selection, an improved particle swarm optimization algorithm is proposed to optimize the parameters of the support vector machine. By proposing a new inertia weight expression, the particle swarm optimization algorithm can effectively conduct a global search at the outset and effectively search locally later in a study, which improves the overall classification accuracy. The experimental results show that the improved particle swarm optimization method is more accurate than a grid search algorithm optimization and other improved particle swarm optimizations with regard to its classification of multiple power quality disturbances. Furthermore, the number of support vectors is reduced.

With the development of the information society, the significant increase in images in online environments poses challenges to media image management. To adapt to the development trend of the big data era and improve the classification effect of media images, this study introduces a dense connection refinement network (DRCN) in convolutional neural network image recognition, combined with attention mechanism, to fully utilize image features of different scales and improve the judgment accuracy of object detection based on increasing feature reuse times. The results indicate the consistent loss value of 0.08 for the DRCN-Attention model, while achieving a peak recall rate of 85% after 30 iterations, and a mean average precision exceeding 80%. The classification accuracy of ships reaches 82%, which is 8% higher than the support vector machine model. This indicates that the proposed media image classification method has high classification accuracy, provides a new technical reference for the field of computer vision, and has certain application value in the intelligent management of media images in the era of big data.

Classification is a substantial operation in data mining, and each element is distributed taking into account its feature values in the corresponding class. Metaheuristics have been widely used in attempts to solve satellite image classification problems. This article proposes a hybrid approach, the flower pigeons-inspired optimization algorithm (FPIO), and the local search method of the flower pollination algorithm is integrated into the pigeon-inspired algorithm. The efficiency and power of the proposed FPIO approach are displayed with a series of images, supported by computational results that demonstrate the cogency of the proposed classification method on satellite imagery. For this work, the Davies-Bouldin Index is used as an objective function. FPIO is applied to different types of images (synthetic, Alsat-2B, and Sentinel-2). Moreover, a comparative experiment between FPIO and the genetic algorithm genetic algorithm is conducted. Experimental results showed that GA outperformed FPIO in matters of time computing. However, FPIO provided better quality results with less confusion. The overall experimental results demonstrate that the proposed approach is an efficient method for satellite imagery classification.

Foraging is a biological process, where a bacterium moves to search for nutriments, and avoids harmful substances. This paper proposes a hybrid approach integrating the bacterial foraging optimization algorithm (BFOA) in a radial basis function neural network, applied to image classification, in order to improve the classification rate and the objective function value. At the beginning, the proposed approach is presented and described. Then its performance is studied with an accent on the variation of the number of bacteria in the population, the number of reproduction steps, the number of elimination-dispersal steps and the number of chemotactic steps of bacteria. By using various values of BFOA parameters, and after different tests, it is found that the proposed hybrid approach is very robust and efficient for several-image classification.

This study proposed a novel framework for plant leaf disease identification. The proposed model consists of four steps including pre-processing, segmentation, feature extraction, and classification. At first, the unwanted noise and overfitting are removed, and also image contrast level is enhanced. Secondly, the Fuzzy C-Means (FCM) based Chameleon Swarm Algorithm (CSA) named as (FCM-CSA) is used for plant leaf diseased part segmentation. In the third stage, the feature extraction is performed using a fast GLCM feature extraction model. Finally, the Progressive Neural Architecture Search (PNAS) is used for plant leaf disease identification. The experimental investigations are carried out using MATLAB software with the Mendeley database. From this dataset, we have used Apple Cedar Apple Rust (ACAR), Cherry Powdery Mildew (CPM), Corn Common Rust (CCCR), Apple Healthy (AH), Grape Black Rot (GBR), Pepper Bell Bacterial Spot (PBBS), Potato Late Blight (PLB) and Tomato Leaf Mold (TLM) disease images. Different measures such as precision, recall, sensitivity, specificity, and accuracy results are used to validate the performance of the proposed model.

Soil surface displacement occurs often following timber harvesting. This displacement is caused by canopy removal from the forest area which makes bare soil to be exposed directly to the rainfall. Thus, the displacement may differ in due to precipitation intensity, and this phenomenon results in the displacement prediction to be uncertain. In this study, we investigated on 1) calculating displacement on timber harvested area in research forest by photogrammetry method, and 2) assessing erosion risk of the area by using three different supervised machine learning classification models. Two time-series of point cloud data (PCD) were acquired by photogrammetry method from June 10, 2022 and July 9, 2022, respectively. These two PCD sets were registered by ground control point (GCP). For displacement estimation, DEM of difference (DoD) was calculated by subtracting DEMs from registered PCD sets. A total of nine terrain variables were generated for analyzing the spatial relationship between displacements and topographic features. Erosion risk was analyzed by three different classification algorithms: random forest (RF), extra gradient boost (XGB), logistic regression (LG). Best performed model was RF with overall accuracy (OA) of 72.9% followed by XGB model (72%), and LG model (71.7%). The results of this research indicate that machine learning based classification algorithms have better performances on erosion risk analysis than statistical algorithm through this research.

기후변화로 인한 강우 패턴의 변화로 단시간 많은 비가 내리는 국지성 호우 및 선형강수대와 같은 위험기상의 발생 빈도와 강도가 빠르게 증가하고 있다. 이러한 위험 기상을 조기에 구별하고, 발생 가능성과 강도를 예측하는 것은 수재해 대응에 매우 중요하다. 본 연구는 극한강우를 유발하는 위험 기상 현상인 집중호우 및 선형강수대를 레이더 강우자료와 각 위험기상의 형태학적 특징을 이용하여 위험강우(선형강수, 정체형 강수, 선형-정체형 강수)를 분리하는 알고리즘을 적용하였다. 위험강우의 형태학적 특징과 정체특성은 종횡비, 누적 강수량, 강수영역 중첩 비율, 최대 강우량을 기준으로 구분한다. 3 km, 5 km 공간해상도의 누적강우량을 입력자료로 사용하여 2021년~2024년 5월부터 10월까지 국내 강우 사례를 대상으로 선형강수대 및 집중호우를 분리하고, 발생 위치 및 특성을 분석하였다. 3 km 해상도가 대표적인 선형강수대 사례에 대해 선형-정체형 강수 형태를 5 km 해상도보다 잘 분류하였다. 낮은 공간 해상도에서는 국지성 강우의 세부구조가 단순화되어 강우시스템이 과대 해석되는 경우가 있음을 확인하였다. 또한, 주요 홍수피해 사례에 대해서는 모두 형태학적 분류 알고리즘에 의해 선형강수대의 형태인 선형-정체형 강수 특성이 있었음을 확인할 수 있었다. 본 연구를 통해서 형태학적 위험기상 분류 알고리즘이 한국의 집중호우 분류에 유용함을 확인하였다.

Due to changes in rainfall patterns driven by climate change, the frequency and intensity of hazardous weather events such as localized heavy rain and line-shaped rainband systems that produce heavy rainfall over short periods are rapidly increasing. Early identification and prediction of the occurrence and intensity of such hazardous weather events are crucial for flood disaster response. This study adapted an algorithm to classify hazardous rainfall types (linear, stationary, and linear-stationary, others) using radar rainfall data and morphological characteristics of heavy rainfall and linear rainfall systems associated with extreme precipitation. Morphological features and stationary characteristics were distinguished based on aspect ratio, accumulated rainfall, rainfall area overlap ratio, and maximum rainfall. Using radar-based accumulated rainfall data at 3 km and 5 km spatial resolution, domestic rainfall cases from May to October 2021-2024 were analyzed to identify linear rainfall systems and heavy rainfall systems, and their occurrence locations and characteristics were examined. The 3 km resolution data more accurately classified linear-stationary rainfall patterns for typical linear rainfall system cases compared to 5 km resolution, it was confirmed that lower spatial resolution sometimes led to oversimplification of detailed structures of localized rainfall, resulting in misinterpretation of rainfall systems. Furthermore, for three major flood disaster cases, the morphological classification algorithm identified that linear-stationary rainfall types. This study confirmed the usefulness of the morphological hazardous weather classification algorithm for classifying heavy rainfall in Korea.

본 연구의 목적은 이용자의 탐색 행태, 시스템의 정보 구축 행태를 기반으로 초기 질의어의 범주에 해당하는 연관 용어들(해당 용어의 지식구조와 관련된 연관 용어들)을 학습기능을 통해 자동으로 제시해 줄 수 있는 지능형 검색 시스템을 구현하는 것이다. 이를 위해 학습을 통해 전문가 수준의 색인어를 추출할 수 있는 지능형자동색인 알고리즘, 자동분류에 관련한 클러스터링 알고리즘과 문서 범주화 알고리즘 그리고 범주 표현 알고리즘에 대한 이론적 연구를 수행하였으며, 이들 이론적 연구를 근거로 비용과 시간적인 측면에서 그리고 재현율과 정도율이란 측면에서 우수한 성능을 발휘할 수 있는 지능형검색시스템을 구현하였다.

This is to develop Intelligent Retrieval System which can automatically present early query's category terms(association terms connected with knowledge structure of relevant terminology) through learning function and it changes searching form automatically and runs it with association terms. For the reason, this theoretical study of Intelligent Automatic Indexing System abstracts expert's index term through learning and clustering algorism about automatic classification, text mining(categorization), and document category representation. It also demonstrates a good capacity in the aspects of expense, time, recall ratio, and precision ratio.

산업의 발전으로 도시화로 인해 건물의 규모가 커지면서, 건물의 공기 정화 및 쾌적한 실내 환경을 유지의 필요 성 또한 증가하고 있다. 냉동 시스템의 모니터링 기술의 발전으로 건물 내에 발생하는 전력 소모량을 관리할 수 있게 되었다. 특히 상업용 건물에서 발생하는 전력 소모량 중 약 40%가 냉동 시스템에서 일어난다. 따라서 본 연구 냉동시스 템 고장진단 알고리즘을 개발하기 위해서 냉동시스템의 구조를 이해하고, 냉동 시스템의 운영과정에서 발생하는 데이터 를 수집 분석하여 다양한 유형과 심각도를 가지는 고장 상황을 조기에 신속하게 탐지 분류하고자 하였다. 특히 분류가 어려운 고장 유형들의 분류 정확도를 향상시키기 위하여 3단계 진단 및 분류 알고리즘을 개발하여 제안하였다. 다수의 실험과 초모수 (hyper parameter) 최적화 과정을 거쳐 각 단계에 적합한 분류 모형으로 SVM과 LGBM에 기반 한 모형을 제시하였다. 본 연구에서는 고장에 영향을 미치는 특성을 최대한 보존하면서, 선행연구에서 어려움을 겪었던 냉매 관련 고장을 포함한 모든 고장 유형을 우수한 결과로 도출하였다.

As the size of buildings increases due to urbanization due to the development of industry, the need to purify the air and maintain a comfortable indoor environment is also increasing. With the development of monitoring technology for refrigeration systems, it has become possible to manage the amount of electricity consumed in buildings. In particular, refrigeration systems account for about 40% of power consumption in commercial buildings. Therefore, in order to develop the refrigeration system failure diagnosis algorithm in this study, the purpose of this study was to understand the structure of the refrigeration system, collect and analyze data generated during the operation of the refrigeration system, and quickly detect and classify failure situations with various types and severity . In particular, in order to improve the classification accuracy of failure types that are difficult to classify, a three-step diagnosis and classification algorithm was developed and proposed. A model based on SVM and LGBM was presented as a classification model suitable for each stage after a number of experiments and hyper-parameter optimization process. In this study, the characteristics affecting failure were preserved as much as possible, and all failure types, including refrigerant-related failures, which had been difficult in previous studies, were derived with excellent results.

본 논문은 다층을 갖는 네트워크를 가장 효율적으로 학습하는 것은 결국 최적의 가중치 벡터의 집합을 찾아가는 과정이라는 간단한 접근 방법을 기본으로 하고 있다. 일반적인 학습 문제의 단점을 극복하기 위해 제안 모델에서는 EBP와 OVSSA의 특징들을 결합한 방법을 사용한다. 즉 EBP가 지역 최소치에 빠질 수 있는 성질을 보강하기 위해 OVSSA의 확률 이론으로 빠져나갈 수 있도록 제안 방법은 각각 알고리즘의 장점만을 취하여 하나의 모델을 구성한다. 제안 알고리즘에서는 EBP에서 오류를 줄이기 위한 방법들을 에너지함수로 사용하고, 이 에너지를 OVSSA로 최소화 하는 방법을 사용하였다. 두 가지의 상이한 성질을 가지는 알고리즘이 합쳐질 수 있음을 간단한 실험 결과를 통해 확인한다.

This paper is based on a simple approach that the most efficient learning of a multi-layered network is the process of finding the optimal set of weight vectors. To overcome the disadvantages of general learning problems, the proposed model uses a combination of features of EBP and OVSSA. In other words, the proposed method can construct a single model by taking advantage of each algorithm so that it can escape to the probability theory of OVSSA in order to reinforce the property that EBP falls into local minimum value. In the proposed algorithm, methods for reducing errors in EBP are used as energy functions and the energy is minimized to OVSSA. A simple experimental result confirms that two algorithms with different properties can be combined.

인터넷을 통해 사용자에게 클라우드 컴퓨팅 서비스를 효율적으로 제공하기 위해서는 데이터 센터에 가상 화와 분산 컴퓨팅 기술을 기반으로 하여 IT 자원을 구성해야 한다. 본 논문은 폭넓은 분야에서 새로운 훈련 데이터 가 언제라도 추가될 수 있고, 또한 언제라도 훈련 데이터에 새로운 속성이 추가될 수 있다는 문제에 특별히 초점을 맞춘다. 이러한 경우, 기존 속성 집합들을 가지는 훈련 데이터로 생성된 규칙은 쓸모없게 된다. 더구나 새롭게 추가 된 데이터나 속성을 가지는 새로운 데이터는 기존 규칙과 결합될 수 없다. 본 논문은 이와 같은 경우를 자연스럽게 처리할 수 있는 보다 진보된 새 추론 엔진을 제안한다. 이 방법에서 기존의 데이터로 부터 생성된 규칙은 개선된 규칙을 생성하기 위한 새로운 데이터 집합과 결합될 수 있다.

To efficiently provide cloud computing services to users over the Internet, IT resources must be configured in the data center based on virtualization and distributed computing technology. This paper focuses specifically on the problem that new training data can be added at any time in a wide range of fields, and new attributes can be added to training data at any time. In such a case, rule generated by the training data with the former attribute set can not be used. Moreover, the rule can not be combined with the new data set(with the newly added attributes). This paper proposes further development of the new inference engine that can handle the above case naturally. Rule generated from former data set can be combined with the new data set to form the refined rule.

본 논문에서는 레이저스캐너만으로 이루어진 감지 시스템을 이용하여 도로 위에 있는 객체의 위치를 추정하고 분류 하는 알고리즘을 제안한다. 각각의 레이저 스캐너에서 획득한 데이터는 그리드 맵을 사용하여 데이터를 융합하였으며, 팽창 연산과 레이블링 방법을 사용하여 측정 오차를 보정하였다. 추출한 객체의 정보(길이, 폭)를 입력으로 사용한 퍼지 방법을 통해 객체를 보행자, 자전거, 차량으로 분류하였으며, 이러한 방법은 레이저스캐너로만 이루어진 감지 시스템의 정확도를 증가시켰다. 또한 본 논문에서는 실제 도로 환경에서 몇 가지 시나리오를 설정하여 실험을 하였다. 실험을 통 해 감지 시스템이 객체를 정확히 분류하는지, GPS-RTK 장비를 사용하여 획득한 위치 정보와 비교하여 객체의 위치 정 보를 정확히 추정하는지 검증하였다.

This paper proposes the on-road object detection and classification algorithm by using a detection system consisting of only laser scanners. Each sensor data acquired by the laser scanner is fused with a grid map and the measurement error and spot spaces are corrected using a labeling method and dilation operation. Fuzzy method which uses the object information (length, width) as input parameters can classify the objects such as a pedestrian, bicycle and vehicle. In this way, the accuracy of the detection system is increased. Through experiments for some scenarios in the real road environment, the performance of the proposed detection and classification system for the actual objects is demonstrated through the comparison with the actual information acquired by GPS-RTK.