• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.11
• /
• pp.4246-4267
• /
• 2020

As the network goes deep into all aspects of people's lives, the number and the complexity of network traffic is increasing, and traffic classification becomes more and more important. How to classify them effectively is an important prerequisite for network management and planning, and ensuring network security. With the continuous development of deep learning, more and more traffic classification begins to use it as the main method, which achieves better results than traditional classification methods. In this paper, we provide a comprehensive review of network traffic classification based on deep learning. Firstly, we introduce the research background and progress of network traffic classification. Then, we summarize and compare traffic classification based on deep learning such as stack autoencoder, one-dimensional convolution neural network, two-dimensional convolution neural network, three-dimensional convolution neural network, long short-term memory network and Deep Belief Networks. In addition, we compare traffic classification based on deep learning with other methods such as based on port number, deep packets detection and machine learning. Finally, the future research directions of network traffic classification based on deep learning are prospected.

• Journal of Korea Game Society
• /
• v.17 no.4
• /
• pp.101-108
• /
• 2017

This study explores how Artificial Intelligence (AI), which is tremendously developed these days, applies to the game and advances. It analyzes the reality of AI and provides reasonable suggestion in Poker, one of the most popular games. Specifically, this study focuses on Texas Hold'em, the most favored kind in the world among various kinds of Poker games and deals with two AIs, Libratus and DeepStack that have applied to the game. Several news media report the growth of AI, but this study will multi-dimensionally discusses how and why AI works in Poker, the real problems of AI, and suggestions for advancement.

• 전기의세계
• /
• v.31 no.4
• /
• pp.281-287
• /
• 1982

PASCAL has a recursive nested structure and uses deep binding of identifiers. This paper studies the problems and techniques in storage management for PASCAL on the IBM 370 system, and presents run-time storage administration algorithms which use stack scheme and heap efficiently on the view of storage. The stack-scheme was used to implement the feature of recursive nested structure and the heap was used to implement the feature of the dynamic allocation procedure and pointer variable, allowing an additional dynamic storage recovery procedure.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.58 no.1
• /
• pp.93-99
• /
• 2009

In this paper, transient voltage response of Polmer Electrolyte Membrane Fuel Cell (PEMFC) stack is analyzed and voltage dynamic characteristic is modeled for optimal design of power conditioning system (PCS). According that the load is changed, the corresponding operating voltage of fuel cell stack is also varied with a certain deep and rising time due to the chemical and mechanical responses. This transient behavior can affect on the operation with respect of PI gain in controller, duty ratio, capacitor of capacitor and so on. So in this paper the detailed theoretical analysis of transient voltage dynamics is explained and the methodology of dynamic modeling is introduced. In addition, the validity and feasibility of the proposed dynamic model is verified by experimental results under various load conditions.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2015.10a
• /
• pp.250-252
• /
• 2015

다양한 분야에서 활발히 연구되는 빅 데이터와 최근 이슈가 되고 있는 딥러닝(Deep-learning) 등은 컴퓨터공학 분야뿐만 아니라 다양한 분야와 접목하여 이에 대한 관심이 증가하고 있다. 대규모 클러스터를 통하여 빅데이터와 딥러닝 같은 계산 집약적인(computational-intensive) 작업을 빠르게 처리할 수 있다. 하지만 대규모 클러스터의 잦은 유휴상태는 클러스터의 활용률은 매우 낮아지게 한다. 본 논문에서는 작업 실행 시간 개선과 클러스터 활용 효율성을 향상시키는 HTCondor 클러스터를 위한 동적 자원 스케줄링 기법을 제안한다. 동적으로 자원 할당을 위해 가상머신으로 HTCondor 클러스터 환경을 구성하였으며, 가상머신의 관리를 위해 OpenStack을 사용하였다. OpenStack기반 HTCondor 클러스터 환경에서 HTCondor Python API와 OpenStack Python API를 사용하여 우리가 제안하는 동적 자원 스케줄링 기법을 구현하였으며, 실험을 통해 제안하는 기법의 성능 및 실현 가능성을 확인하였다.

• Transactions of the Korean hydrogen and new energy society
• /
• v.31 no.5
• /
• pp.436-443
• /
• 2020

The performance prediction model of a solid oxide fuel cell stack has been developed using deep neural network technique, one of the machine learning methods. The machine learning has been received much interest in various fields, including energy system mo- deling. Using machine learning technique can save time and cost requried in developing an energy system model being compared to the conventional method, that is a combination of a mathematical modeling and an experimental validation. Results reveal that the mean average percent error, root mean square error, and coefficient of determination (R2) range 1.7515, 0.1342, 0.8597, repectively, in maximum. To improve the predictability of the model, the pre-processing is effective and interpolative machine learning and application is more accurate than the extrapolative cases.

• JSTS:Journal of Semiconductor Technology and Science
• /
• v.13 no.6
• /
• pp.647-654
• /
• 2013

The design and analysis of analog circuit application on CMOS technology are a challenge in deep sub-micrometer process. This paper is a study on the performance value of Double Gate (DG) Metal Oxide Semiconductor Field Effect Transistor (MOSFET) with Gate Stack and the channel engineering Single Halo (SH), Double Halo (DH). Four different structures have been analysed keeping channel length constant. The short channel parameters and different sub-threshold analog figures of merit (FOMs) are analysed. This work extensively provides the device structures which may be applicable for high speed switching and low power consumption application.

• Journal of Korea Society of Industrial Information Systems
• /
• v.19 no.1
• /
• pp.53-60
• /
• 2014

IEEE 802.15.4 is one of the most widely adopted physical layer standards in the area of LR-WPAN(Low-Rate Wireless Personal Area Network). Numerous previous researches have contributed to deep insights on energy efficiency, transmission throughput, and reliability that IEEE 802.15.4 delivers to the LR-WPAN. As a research that is orthogonal and complementary to previous researches, we explore the implementation and practical performance evaluation of IEEE 802.15.4 MAC software. We implement the MAC software from the perspective of the networking stack, exploring the issues raised when the MAC software serves as a functional component in a complete networking stack consisting of MAC, network as well as well as application support layers. The performance is evaluated on a realistic experimental software environment integrated with operating system, networking stack, and applications.

• Annual Conference on Human and Language Technology
• /
• 2018.10a
• /
• pp.689-691
• /
• 2018

의존 구문 분석은 자연어 이해 영역의 대표적인 과제 중 하나이다. 본 논문에서는 한국어 의존 구분 분석의 성능 향상을 위해 Deep Bi-affine Network 와 스택 포인터 네트워크의 앙상블 모델을 제안한다. Bi-affine 모델은 그래프 기반 방식, 스택 포인터 네트워크의 경우 그래프 기반과 전이 기반의 장점을 모두 사용하는 모델로 서로 다른 모델의 앙상블을 통해 성능 향상을 기대할 수 있다. 두 모델 모두 한국어 어절의 특성을 고려한 자질을 사용하였으며 세종 의존 구문 분석 데이터에 대해 UAS 90.60 / LAS 88.26(Deep Bi-affine Network), UAS 92.17 / LAS 90.08(스택 포인터 네트워크) 성능을 얻었다. 두 모델에 대한 앙상블 기법 적용시 추가적인 성능 향상을 얻을 수 있었다.

• International journal of advanced smart convergence
• /
• v.12 no.4
• /
• pp.142-146
• /
• 2023

We present a method for generating 3D structures and rendering objects by combining VAE (Variational Autoencoder) and GAN (Generative Adversarial Network). This approach focuses on generating and rendering 3D models with improved quality using residual learning as the learning method for the encoder. We deep stack the encoder layers to accurately reflect the features of the image and apply residual blocks to solve the problems of deep layers to improve the encoder performance. This solves the problems of gradient vanishing and exploding, which are problems when constructing a deep neural network, and creates a 3D model of improved quality. To accurately extract image features, we construct deep layers of the encoder model and apply the residual function to learning to model with more detailed information. The generated model has more detailed voxels for more accurate representation, is rendered by adding materials and lighting, and is finally converted into a mesh model. 3D models have excellent visual quality and accuracy, making them useful in various fields such as virtual reality, game development, and metaverse.