데이터와 인공신경망 능력 계산

이덕균,박지은 한국정보통신학회 2022 한국정보통신학회논문지 Vol.26 No.1

Recently, various uses of artificial intelligence have been made possible through the deep artificial neural network structure of machine learning, demonstrating human-like capabilities. Unfortunately, the deep structure of the artificial neural network has not yet been accurately interpreted. This part is acting as anxiety and rejection of artificial intelligence. Among these problems, we solve the capability part of artificial neural networks. Calculate the size of the artificial neural network structure and calculate the size of data that the artificial neural network can process. The calculation method uses the group method used in mathematics to calculate the size of data and artificial neural networks using an order that can know the structure and size of the group. Through this, it is possible to know the capabilities of artificial neural networks, and to relieve anxiety about artificial intelligence. The size of the data and the deep artificial neural network are calculated and verified through numerical experiments. 최근 인공지능의 다양한 활용은 기계학습의 딥 인공신경망 구조를 통해 가능해졌으며 인간과 같은 능력을 보여주고 있다. 불행하게도 딥 구조의 인공신경망은 아직 정확한 해석이 이루어지고 있지 못하고 있다. 이러한 부분은 인공지능에 대한 불안감과 거부감으로 작용하고 있다. 우리는 이러한 문제 중에서 인공신경망의 능력 부분을 해결한다. 인공신경망 구조의 크기를 계산하고, 그 인공신경망이 처리할 수 있는 데이터의 크기를 계산해 본다. 계산의 방법은 수학에서 쓰이는 군의 방법을 사용하여 데이터와 인공신경망의 크기를 군의 구조와 크기를 알 수 있는 Order를 이용하여 계산한다. 이를 통하여 인공신경망의 능력을 알 수 있으며, 인공지능에 대한 불안감을 해소할 수 있다. 수치적 실험을 통하여 데이터의 크기와 딥 인공신경망을 계산하고 이를 검증한다.

인공지능 딥러닝의 역사와 현황, 그리고 미래 방향

이원진 대한치과의사협회 2022 대한치과의사협회지 Vol.60 No.5

Deep learning is a subset of machine learning, and machine learning is also a subset of artificial intelligence (AI). The biggest difference between machine learning and deep learning is that in the learning of artificial intelligence models, machine learning basically requires a human feature extraction process before learning, but deep learning does not require this process and the original data is directly used as input. The development of deep learning coincides with the development of artificial neural networks (ANNs), and many people have contributed to the development of artificial neural networks for decades. The following five models are the representative architectures most widely used in deep learning. That is, Deep Feedforward Neural Network (DFFNN), Convolutional Neural Network (CNN), Deep Belief Network (DBN), Autoencoders (AE), and Long Short-Term Memory (LSTM) Network. A convolutional neural network (CNN) is a feedforward NN composed of a convolutional layer, a ReLU activation function, and a pooling layer. CNNs provide properties of weight sharing and local connectivity to process high-dimensional data. In dental and medical fields, an AI model that can be interpretable or explainable (XAI) is needed to increase patient persuasiveness. In the future, explainable AI (XAI) will become an indispensable and practical component in order to obtain an improved, transparent, secure, fair and unbiased AI learning model.

인공신경망을 이용한 KOMPSAT-3/3A/5 영상으로부터 자연림과 인공림의 분류

이용석,박숭환,정형섭,백원경 대한원격탐사학회 2018 大韓遠隔探査學會誌 Vol.34 No.6

Natural forests are un-manned forests where the artificial forces of people are not applied to the formation of forests. On the other hand, artificial forests are managed by people for their own purposes such as producing wood, preventing natural disasters, and protecting wind. The artificial forests enable us to enhance economical benefits of producing more wood per unit area because it is well-maintained with the purpose of the production of wood. The distinction surveys have been performed due to different management methods according to forests. The distinction survey between natural forests and artificial forests is traditionally performed via airborne remote sensing or in-situ surveys. In this study, we suggest a classification method of forest types using satellite imagery to reduce the time and cost of in-situ surveying. A classification map of natural forest and artificial forest were generated using KOMPSAT-3, 3A, 5 data by employing artificial neural network (ANN). And in order to validate the accuracy of classification, we utilized reference data from 1/5,000 stock map. As a result of the study on the classification of natural forest and plantation forest using artificial neural network, the overall accuracy of classification of learning result is 77.03% when compared with 1/5,000 stock map. It was confirmed that the acquisition time of the image and other factors such as needleleaf trees and broadleaf trees affect the distinction between artificial and natural forests using artificial neural networks. 자연림은 산림의 조성 과 보육 등에 인공적인 사람의 힘이 가해지지 않은 자연 상태의 산림이다. 반면 인공림은 사람이 조성 및 보육관리 하는 숲으로 목재생산, 자연재해 예방, 방풍 등의 목적을 가지는 산림이다. 인공림은 목재생산 등 인간이 목적을 가지고 관리하여 단위 면적당 더 많은 목재를 생산할 수 있는 경제적 장점도 가지고 있다. 자연림과 인공림의 구분은 산림 형태의 관리 방법과 목정이 상이하여 산림조사에서 기본적으로 조사하는 요소이며, 자연림과 인공림의 구분은 항공사진 판독과 현지조사 등의 절차를 통해 이루어진다. 본 연구에서는 자연림과 인공림의 분류에 KOMPSAT-3, 3A, 5 위성 영상데이터에 인공신경망(Artificial Neural Network: ANN)을 적용하여 자연림과 인공림의 분류도를 만들고, 산림청의 1/5,000임상도의 자연림과 인공림 분류도와 비교하여 평가하였다. 인공신경망을 이용한 산림의 자연림과 인공림 구분의 연구를 진행한 결과, 1/5,000 임상도와 비교했을 때, 학습결과 분류 전체 정확도는 77.03%이다. 영상의 획득 시기와 산림의 침엽수와 활엽수 등 기타요인이 인공신경망을 이용한 산림의 인공림과 자연림의 구분에 많은 영향을 미치는 것을 확인하였다.

신경회로망을 이용한 비선형모델기준 적응제어기의 설계에 관한 연구

金昌圭,卓漢浩 진주산업대학교 1997 산업과학기술연구소보 Vol.- No.4

Artificial neural networks show promise as elements in control systems. One feature of traditional control system design which has been largely lacking in the work with neural networks to date is the analysis of the closed loop stability of the controlled system. The principal aim of this work was to develop a control architecture for which such analyses can be made Theoretical results were developed for the stability of systems using controllers based on approximate controllers, such as artificial neural networks. These results are based on fairly strong assumptions concerning the ability of the artificial neural networks to learn a model of the forward dynamics of the plant. The stability result are based on an application of the concepts of Liapunov stability to systems which are stable in the sense of Lagrange: that is, the stability is with respect to a region in the state space, rather than a point. These stability results motivate the selection of a model reference adaptive controller architecture. The controller incorporates a performance model, which provides a desired output trajectory in response to system inputs (commands), and a convergence model, which determines the desired perturbation dynamics of the true output about the model output trajectory. The conotrol input is selected by on line minimization of a cost function, based on a Liapunov-like function derived from the convergence model. This minimization procedure uses the neural networks model of the system dynamics to predict the response of the system to a candidate control input. The controller described was applied to the adaptive control of the inverted pendulum problem. The addition of a dither signal to the calculated control input was shown to enhance the ability of error back propagation to improve the artificial neural network model on line.

모듈형 인공신경망을 이용한 연직배수공법에서의 압밀침하량 예측

민덕기,황광모,전형원 한국지반공학회 2000 한국지반공학회논문집 Vol.16 No.2

In this paper, consolidation settlements with time at vertical drain sites were predicted by artificial neural networks. Laboratory test results and field measurements of two vertical drain sites were used for training and testing neural networks. Predicted consolidation settlements by trained artificial neural networks were compared with measured settlements by field instrumentation. To improve the prediction accuracy, modular artificial neural networks were studied. From the results of applying artificial neural networks to the same situation, it was shown that modular artificial neural network model was more accurate for the prediction of the consolidation settlements than the general model.

신경망(Neural network) 기계학습에 기초한 인공지능 관련 발명의 법적 문제 - 발명의 성립성과 발명자 적격성을 중심으로 -

김관식 한국지식재산학회 2021 産業財産權 Vol.- No.67

인공지능은 인공신경망을 이용한 기계학습에 기초하고 있는 점이 특징으로, 인공지능 관련 발명은 인공지능의 구현에 특징이 있는 발명, 인공지능 기술의 응용분야에 특징이 있는 발명, 인공지능에 의하여 생 성된 발명 등으로 구분할 수 있다. 인공지능 자체의 구현에 관한 발명과 인공지능의 응용에 관한 발명은 기존의 컴퓨터프로그램의 성립성 판단 기준, 영업방법에 관한 발명의 성립성 판단 기준 등을 원용하여 적용이 가능할 것이다. AI에 의하여 ‘생성된 발명’의 경우에는 AI의 생성물이 데이터나 이미 지인 경우에는 저작물에 해당할 여지는 있으나 자연법칙을 이용한 기 술적 사상의 창작으로서의 발명으로 보기는 힘들다. 학습데이터에 기초 하여 생성된 학습모델에 새로운 데이터의 입력에 의하여 자동적으로 인공지능 생성물이 생성되는 경우에는, 자연법칙을 이용한 기술적 사상 의 ‘창작’으로 인정할 수 있는지 여부가 쟁점으로 될 것으로 예상되는데, 이러한 창작성이 인공지능에 의하여 발휘된 것으로 인정하기는 어려울 것이다. 발명자의 특정과 관련하여, AI 자체에 특징이 있는 발명, AI의 응용에 특징이 있는 발명의 경우에는 AI의 설계, 구축, 응용, 학습에 주도적으로 관여한 자연인이 발명자로 쉽게 인정될 것이다. 반면에 AI에 의하여 ‘생성된 발명’의 경우, 발명자를 어느 사람(것)으로 특정하여야 하는가 의 문제가 있고 이러한 문제는 AI가 발명자로 될 수 있는 자격이 있는지 의 문제와도 연관이 있다. AI에 대하여 발명자(출원인, 특허권자)의 법인격을 부여할 수 있는지 의 문제는, 현재 실현된 인공지능의 기술적 특징 및 그 본질에 비추어 볼 때, 당분간 자연인에 필적하는 법인격을 부여하기는 힘들 것으로 생각된다. 다만 강한 AI(또는 AGI)가 향후 등장한다면 법인격의 부여와 더불어 이러한 AI에 의하여 생성된 발명에 대해서는 자연법칙을 이용 한 기술적 사상의 ‘창작’으로서 발명의 성립성을 인정할 수 있는 여지는 더욱 증가할 것이다. Artificial intelligence is characterized by the fact that it is based on machine learning using artificial neural networks. Artificial intelligence-related inventions can be classified into inventions characterized by the implementation of artificial intelligence, inventions characterized by the application field of artificial intelligence technology, and inventions created by artificial intelligence. For inventions related to the realization of artificial intelligence itself and inventions related to the application of artificial intelligence, it is possible to apply the criteria for judging the patent eligibility of computer programs and inventions related to business methods. In the case of “invention created” by AI, it is difficult to see it as an invention when the product of AI is data or image. In the case where an artificial intelligence product is automatically generated by inputting new data to a learning model created based on the learning data, the issue will be whether it can be recognized as a “creation” of technical ideas. It is, however, difficult to admit that such creativity was exerted by artificial intelligence. With regard to the identification of the inventors for the 'inventions created' by AI, there is a question of which person (thing) should be specified as an inventor. These issues are also related to the question of whether AI is qualified to become an inventor. With regard to the question of whether a legal personality as an inventor (applicant, patentee) can granted to AI, it is not plausible for the time being to give a legal personality similar to a natural person, in light of the technical characteristics and nature of artificial intelligence currently realized. However, if strong AI (or AGI) emerges in the future, there will be more room to recognize the subject-matter patentability of the inventions created by AI along with granting legal personality to AI.

배경기계학습을 사용한 침입탐지시스템

김형훈(Hyung-Hoon Kim),조정란(Jeong-Ran Cho) 한국컴퓨터정보학회 2019 韓國컴퓨터情報學會論文誌 Vol.24 No.5

The existing subtract image based intrusion detection system for CCTV digital images has a problem that it can not distinguish intruders from moving backgrounds that exist in the natural environment. In this paper, we tried to solve the problems of existing system by designing real - time intrusion detection system for CCTV digital image by combining subtract image based intrusion detection method and background learning artificial neural network technology. Our proposed system consists of three steps: subtract image based intrusion detection, background artificial neural network learning stage, and background artificial neural network evaluation stage. The final intrusion detection result is a combination of result of the subtract image based intrusion detection and the final intrusion detection result of the background artificial neural network. The step of subtract image based intrusion detection is a step of determining the occurrence of intrusion by obtaining a difference image between the background cumulative average image and the current frame image. In the background artificial neural network learning, the background is learned in a situation in which no intrusion occurs, and it is learned by dividing into a detection window unit set by the user. In the background artificial neural network evaluation, the learned background artificial neural network is used to produce background recognition or intrusion detection in the detection window unit. The proposed background learning intrusion detection system is able to detect intrusion more precisely than existing subtract image based intrusion detection system and adaptively execute machine learning on the background so that it can be operated as highly practical intrusion detection system.

오피스 임대료 결정 모형에 관한 연구 - 회귀분석과 신경망 이론을 중심으로 -

김선주,이상엽 한국지역학회 2008 지역연구 Vol.24 No.2

The purpose of this paper is to study on forecasting of Seoul office rental price based on regression and artificial neural networks. Neural networks are artificial intelligence tools inspired by the biological brain. An artificial neural network is formed from numerous interconnected neurons known as processing elements. Each processing element receives inputs through connections with other elements. Just like the synaptic strength in the biological neuron, each input is multiplied by a weight, and the sum of all weighted inputs is modified by a transfer function to produce an output signal. A neural network is a combination of a number of processing elements organized in layers. The arrangement of processing elements in different meaningful configurations leads to different neural network models suited to different classes of problems. This model is estimated with the data of offices having more than ten stories at 2004 in SAMS. There are three key determinant factors of rental prices: physical factors, geographical factors. The result of this study is that the prediction exactness of Neural Network Model is higher than regression analysis model. Regression analysis model has 0.33890 RMSE, and Neural Network Model 0.00958 RMSE. As a final remark, it should be noted that a general criticism of neural network models is the black-box computations that decrease trust in results.

Adoption of Artificial Neural Network for Rest, Enhanced Postprocessing of Beats, and Initial Melody Processing for Automatic Composition System

김경환,정성훈 한국디지털콘텐츠학회 2016 한국디지털콘텐츠학회논문지 Vol.17 No.6

This paper proposes a new method to improve the three problems of existing automatic composition method using artificial neural networks. The first problem is that the existing beat post-processing to fit into music theories could not handle all the cases of occurring. The second one is that the pitch space generated by artificial neural networks is distorted because the rest is trained with the pitch on the same neural network with large values. The last problem is caused by the difference between the initial melody and beats given by user and those generated by an artificial neural network in the process of new composition. In order to treat these problems, we propose an enhanced post-processing of beats, initial melody processing, and adoption of artificial neural network for rest. It was found from experiments that the proposed methods totally resolved the three problems.

Predicting PM2.5 Concentrations Using Artificial Neural Networks and Markov Chain, a Case Study Karaj City

Gholamreza Asadollahfardi,Hossein Zangooei,Shiva Homayoun Aria 한국대기환경학회 2016 Asian Journal of Atmospheric Environment (AJAE) Vol.10 No.2

The forecasting of air pollution is an important and popular topic in environmental engineering. Due to health impacts caused by unacceptable particulate matter (PM) levels, it has become one of the greatest concerns in metropolitan cities like Karaj City in Iran. In this study, the concentration of PM2.5 was predicted by applying a multilayer percepteron (MLP) neural network, a radial basis function (RBF) neural network and a Markov chain model. Two months of hourly data including temperature, NO, NO2, NOx, CO, SO2 and PM10 were used as inputs to the artificial neural networks. From 1,488 data, 1,300 of data was used to train the models and the rest of the data were applied to test the models. The results of using artificial neural networks indicated that the models performed well in predicting PM2.5 concentrations. The application of a Markov chain described the probable occurrences of unhealthy hours. The MLP neural network with two hidden layers including 19 neurons in the first layer and 16 neurons in the second layer provided the best results. The coefficient of determination (R2), Index of Agreement (IA) and Efficiency (E) between the observed and the predicted data using an MLP neural network were 0.92, 0.93 and 0.981, respectively. In the MLP neural network, the MBE was 0.0546 which indicates the adequacy of the model. In the RBF neural network, increasing the number of neurons to 1,488 caused the RMSE to decline from 7.88 to 0.00 and caused R2 to reach 0.93. In the Markov chain model the absolute error was 0.014 which indicated an acceptable accuracy and precision. We concluded the probability of occurrence state duration and transition of PM2.5 pollution is predictable using a Markov chain method.