• 전자공학회논문지
• /
• 제54권2호
• /
• pp.76-82
• /
• 2017

랜섬웨어를 포함한 악성코드를 빠르게 탐지하여 빅데이터를 보호하기 위해 본 연구에서는 인공지능의 딥러닝으로 학습된 이미지 분석을 통한 악성코드 분석 기법을 제안한다. 우선 악성코드들에서 일반적으로 사용하는 2,400여개 이상의 데이터를 분석하여 인공신경망 Convolutional neural network 으로 학습하고 데이터를 이미지화 하였다. 추상화된 이미지 그래프로 변환하고 부분 그래프를 추출하여 악성코드가 나타내는 집합을 정리하였다. 제안한 논문에서 추출된 부분 집합들 간의 비교 분석을 통해 해당 악성코드들이 얼마나 유사한지를 실험으로 분석하였으며 학습을 통한 방법을 이용하여 빠르게 추출하였다. 실험결과로부터 인공지능의 딥러닝을 이용한 정확한 악성코드 탐지 가능성과 악성코드를 이미지화하여 분류함으로써 더욱 빠르고 정확한 탐지 가능성을 보였다.

• 한국콘텐츠학회논문지
• /
• 제20권1호
• /
• pp.448-459
• /
• 2020

최근 국민의 생명과 사회 안전에 대한 관심이 부각됨에 따라 국가 기관들의 정보활동에 대한 논의가 확대되고 있다. 특히 경찰 조직은 공공안녕에 대한 위험의 예방과 대응 관련 정보의 수집·작성 및 배포를 주요 활동 범위에 포함하고 있지만 수사첩보의 처리와 활용에 대한 정보경찰과 수사경찰 구성원 간의 인식 차이는 경찰 수사력 강화에 장애요인으로 작용하고 있다. 본 연구는 경찰조직의 범죄정보 수집 활동을 제고할 수 있는 방안을 모색하기 위하여 비공식적 사회연결망 분석을 통해 경찰 구성원들의 유기적 연계관계를 파악하고, 비공식적 상호작용에서 나타나는 정보활동의 맥락과 그 관계적 속성을 도출하였다. 분석 결과, 정보경찰과 수사경찰 구성원 간의 역할 수행이 상이한 구조적 특성을 가지고 있음을 확인할 수 있었다. 둘째, 발신 및 수신 중심성이 높은 특정 구성원들이 범죄정보 수집 활동에서 중심적 역할을 담당하고 있는 것으로 확인되었다. 셋째, 형사과와 정보과 간의 협조체제 구축을 위해서는 상대조직의 정보공유자를 발굴하여 정보협조의 공백을 최소화하고 범죄정보의 교류가 상호 의존적 관계를 형성하도록 하는 것이 필요하다고 판단되었다. 이러한 분석 결과를 기반으로 본 연구는 향후 경찰조직의 범죄정보 수집 활동을 개선하기 위한 효율적 협조체제 관계망을 구축하는데 기여할 수 있을 것으로 본다.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2023년도 추계학술발표대회
• /
• pp.540-542
• /
• 2023

Due to the serious issues posed by facial manipulation technologies, many researchers are becoming increasingly interested in the identification of face forgeries. The majority of existing face forgery detection methods leverage powerful data adaptation ability of neural network to derive distinguishing traits. These deep learning-based detection methods frequently treat the detection of fake faces as a binary classification problem and employ softmax loss to track CNN network training. However, acquired traits observed by softmax loss are insufficient for discriminating. To get over these limitations, in this study, we introduce a novel discriminative feature learning based on Vision Transformer architecture. Additionally, a separation-center loss is created to simply compress intra-class variation of original faces while enhancing inter-class differences in the embedding space.

• 전기학회논문지
• /
• 제67권12호
• /
• pp.1657-1664
• /
• 2018

This paper proposes a decision making scheme for choosing the best move at each state of game in order to implement an artificial intelligence othello game player. The proposed decision making scheme predicts the various possible states of the game when the game has progressed from the current state, evaluates the degree of possibility of winning or losing the game at the states, and searches the best move based on the evaluation. In this paper, we generate learning data by decomposing the records of professional players' real game into states, matching and accumulating winning points to the states, and using the Artificial Neural Network that learned them, we evaluated the value of each predicted state and applied the Minimax search to determine the best move. We implemented an artificial intelligence player of the Othello game by applying the proposed scheme and evaluated the performance of the game player through games with three different artificial intelligence players.

• International Journal of Computer Science & Network Security
• /
• 제22권3호
• /
• pp.107-112
• /
• 2022

The automotive industry is motivated to provide more and more amenities to its customers. The industry is taking advantage of artificial intelligence by increasing different sensors and gadgets in vehicles machoism is forward collision warning, at the same time road accidents are also increasing which is another concern to address. So there is an urgent need to provide an A.I based system to avoid such incidents which can be address by using artificial intelligence and global positioning system. Automotive/smart vehicles protection has become a major study of research for customers, government and also automotive industry engineers In this study a two layered novel hypothetical approach is proposed which include in-time vehicle/obstacle detection with auto warning mechanism for collision detection & avoidance and later in a case of an accident manifestation GPS & video camera based alerts system and interrupt generation to nearby ambulance or rescue-services units for in-time driver rescue.

• Nuclear Engineering and Technology
• /
• 제53권10호
• /
• pp.3275-3285
• /
• 2021

A Nuclear Power Plant (NPP) is a complex dynamic system-of-systems with highly nonlinear behaviors. In order to control the plant operation under both normal and abnormal conditions, the different systems in NPPs (e.g., the reactor core components, primary and secondary coolant systems) are usually monitored continuously, resulting in very large amounts of data. This situation makes it possible to integrate relevant qualitative and quantitative knowledge with artificial intelligence techniques to provide faster and more accurate behavior predictions, leading to more rapid decisions, based on actual NPP operation data. Data-driven models (DDM) rely on artificial intelligence to learn autonomously based on patterns in data, and they represent alternatives to physics-based models that typically require significant computational resources and might not fully represent the actual operation conditions of an NPP. In this study, a feed-forward backpropagation artificial neural network (ANN) model was trained to simulate the interaction between the reactor core and the primary and secondary coolant systems in a pressurized water reactor. The transients used for model training included perturbations in reactivity, steam valve coefficient, reactor core inlet temperature, and steam generator inlet temperature. Uncertainties of the plant physical parameters and operating conditions were also incorporated in these transients. Eight training functions were adopted during the training stage to develop the most efficient network. The developed ANN model predictions were subsequently tested successfully considering different new transients. Overall, through prompt prediction of NPP behavior under different transients, the study aims at demonstrating the potential of artificial intelligence to empower rapid emergency response planning and risk mitigation strategies.

• International Journal of Computer Science & Network Security
• /
• 제23권3호
• /
• pp.144-150
• /
• 2023

With progressive advancements in Man-made consciousness (computer based intelligence) and Profound Learning (DL), contributing altogether to Normal Language Handling (NLP), the precision and nature of Machine Interpretation (MT) has worked on complex. There is a discussion, but that its no time like the present the human interpretation became immaterial or excess. All things considered, human flaws are consistently dealt with by its own creations. With the utilization of brain networks in machine interpretation, its been as of late guaranteed that keen frameworks can now decipher at standard with human interpreters. In any case, simulated intelligence is as yet not without any trace of issues related with handling of a language, let be the intricacies and complexities common of interpretation. Then, at that point, comes the innate predispositions while planning smart frameworks. How we plan these frameworks relies upon what our identity is, subsequently setting in a one-sided perspective and social encounters. Given the variety of language designs and societies they address, their taking care of by keen machines, even with profound learning abilities, with human proficiency looks exceptionally far-fetched, at any rate, for the time being.

• Nuclear Engineering and Technology
• /
• 제54권12호
• /
• pp.4684-4692
• /
• 2022

Radionuclide identification is an important part of the nuclear material identification system. The development of artificial intelligence and machine learning has made nuclide identification rapid and automatic. However, many methods directly use existing deep learning models to analyze the gamma-ray spectrum, which lacks interpretability for researchers. This study proposes an explainable radionuclide identification algorithm based on the convolutional neural network and class activation mapping. This method shows the area of interest of the neural network on the gamma-ray spectrum by generating a class activation map. We analyzed the class activation map of the gamma-ray spectrum of different types, different gross counts, and different signal-to-noise ratios. The results show that the convolutional neural network attempted to learn the relationship between the input gamma-ray spectrum and the nuclide type, and could identify the nuclide based on the photoelectric peak and Compton edge. Furthermore, the results explain why the neural network could identify gamma-ray spectra with low counts and low signal-to-noise ratios. Thus, the findings improve researchers' confidence in the ability of neural networks to identify nuclides and promote the application of artificial intelligence methods in the field of nuclide identification.

• 문화기술의 융합
• /
• 제4권4호
• /
• pp.261-266
• /
• 2018

본 연구는 인공지능 기반 작곡 프로그램 현황을 살펴보고 실정을 고려한 제언을 제공하고자 한다. 인공지능 기반 작곡 프로그램은 기존의 '전문가 시스템' 방식의 알고리즘을 벗어나 심층신경망 이론의 발전 및 빅데이터 처리 기술 향상과 더불어 눈부신 성장을 보이고 있다. 이에 따라 클래식 음악과, 팝음악을 작곡하는데 있어 인공지능 기반 작곡 프로그램이 학계와 산업계에서 다양하게 제안되고 있으며, 최근 수년 사이 대중의 평가도 달라지고 있다. 다만 해당 기술 개발과 관련하여 여전한 한계점들이 분명히 존재하는 바, 대중의 인식 문제, 데이터베이스화되지 않은 가치 있는 사료들의 누락, 관련 법규의 미비, 음악적인 부분보다는 기술적 관점에서 해당 산업이 주도되는 점 등을 개선할 필요가 있겠다. 이 같은 점이 보완된다면, 인공지능 기반 기술은 국가 경쟁력 확보와 유지에 있어 중요한 역할을 해낼 것으로 보인다.

• ETRI Journal
• /
• 제46권2호
• /
• pp.277-289
• /
• 2024

In this paper, we describe a neural network-based application that recommends multiple items using dialog context input and simultaneously outputs a response sentence. Further, we describe a multi-item recommendation by specifying it as a set of clothing recommendations. For this, a multimodal fusion approach that can process both cloth-related text and images is required. We also examine achieving the requirements of downstream models using a pretrained language model. Moreover, we propose a gate-based multimodal fusion and multiprompt learning based on a pretrained language model. Specifically, we propose an automatic evaluation technique to solve the one-to-many mapping problem of multi-item recommendations. A fashion-domain multimodal dataset based on Koreans is constructed and tested. Various experimental environment settings are verified using an automatic evaluation method. The results show that our proposed method can be used to obtain confidence scores for multi-item recommendation results, which is different from traditional accuracy evaluation.