• 반도체디스플레이기술학회지
• /
• 제21권3호
• /
• pp.1-6
• /
• 2022

Modern computer systems usually have special hardware for operations used in deep learning workload even edge computing environment. Non-volatile memories (NVMs) have been considered for alternative memory storage because they consume little static energy and occupy small area. However, there is a problem for NVMs to be directly adopted. An NVM cell has limited write endurance, so that the lifetime of NVM-based memory system is much shorter than that of conventional memory system. To overcome this problem for the deep learning system, this paper proposes a novel method to extend the lifetime based on the analysis of the deep learning workloads. If an incoming block has more than a predefined number of frequently used values, the cacheline is defined as write friendly block. During the victim selection, the cacheline has lower possibility to be chosen as victim. The experimental results show that the lifetime is increased by about 50% and energy consumption is decreased by 3% with a little performance hurt.

• International Journal of Internet, Broadcasting and Communication
• /
• 제14권3호
• /
• pp.161-170
• /
• 2022

Recently, NAND-type flash memory has been regarded to be new promising storage media for large-scale database systems. For flash memory to be employed for that purpose, we need to reduce its expensive update cost caused by the inablity of in-place updates. To remedy such a drawback in flash memory, we propose a new flash-aware buffering scheme that enables virtual flushing of dirty pages. To this end, we slightly alter the tradional algorithms used for the logging scheme and buffer management scheme. By using the mechanism of virtual flushing, our proposed buffering scheme can efficiently prevent the frequenct occureces of page updates in flash storage. Besides the advantage of reduced page updates, the proposed viurtual flushing mechanism works favorably for shorneing a recocery time in the presense of failure. This is because it can reduce the time for redo actions during a recovry process. Owing to those two benefits, we can say that our scheme couble be very profitable when it is incorporated into cutting-edge flash-based database systems.

• International Journal of Computer Science & Network Security
• /
• 제23권4호
• /
• pp.166-171
• /
• 2023

The Stock price analysis is an increasing concern in a financial time series. The purpose of the study is to analyze the price parameters of date, high, low, and news feed about the stock exchange price. Long short term memory (LSTM) is a cutting-edge technology used for predicting the data based on time series. LSTM performs well in executing large sequence of data. This paper presents the Long Short Term Memory Model has used to analyze the stock price ranges of 10 days and 20 days by exponential moving average. The proposed approach gives better performance using technical indicators of stock price with an accuracy of 82.6% and cross entropy of 71%.

• 한국정보기술학회 영문논문지
• /
• 제9권1호
• /
• pp.115-125
• /
• 2019

In this paper, the image is received from the camera and the lane is sensed. There are various ways to detect lanes. Generally, the method of detecting edges uses a lot of the Sobel edge detection and the Canny edge detection. The minimum use of multiplication and division is used when designing for the hardware configuration. The images are tested using a black box image mounted on the vehicle. Because the top of the image of the used the black box is mostly background, the calculation process is excluded. Also, to speed up, YCbCr is calculated from the image and only the data for the desired color, white and yellow lane, is obtained to detect the lane. The median filter is used to remove noise from images. Intermediate filters excel at noise rejection, but they generally take a long time to compare all values. In this paper, by using addition, the time can be shortened by obtaining and using the result value of the median filter. In case of the Sobel edge detection, the speed is faster and noise sensitive compared to the Canny edge detection. These shortcomings are constructed using complementary algorithms. It also organizes and processes data into parallel processing pipelines. To reduce the size of memory, the system does not use memory to store all data at each step, but stores it using four line buffers. Three line buffers perform mask operations, and one line buffer stores new data at the same time as the operation. Through this work, memory can use six times faster the processing speed and about 33% greater quantity than other methods presented in this paper. The target operating frequency is designed so that the system operates at 50MHz. It is possible to use 2157fps for the images of 640by360 size based on the target operating frequency, 540fps for the HD images and 240fps for the Full HD images, which can be used for most images with 30fps as well as 60fps for the images with 60fps. The maximum operating frequency can be used for larger amounts of the frame processing.

• 정보보호학회논문지
• /
• 제32권3호
• /
• pp.487-499
• /
• 2022

딥러닝과 같은 기계학습 기술은 최근에 광범위하게 활용되고 있다. 이러한 딥러닝은 최근 낮은 컴퓨팅 성능을 가지는 임베디드 기기 및 엣지 디바이스에서 보안성 향상을 위해 ARM TrustZone과 같은 신뢰 수행 환경에서 수행되는데, 이와 같은 실행 환경에서는 제한된 컴퓨팅 자원으로 인해 정상적인 수행에 방해를 받는다. 이를 극복하기 위해 DNN 모델 partitioning을 통해 TEE의 제한된 memory를 효율적으로 사용하며 DNN 모델을 보호하는 TPMP를 제안한다. TPMP는 최적화된 memory 스케줄링을 통해 기존의 memory 스케줄링 방법으로 수행할 수 없었던 모델들을 TEE 내에서 수행하여 시스템 자원 소모를 거의 증가시키지 않으면서 DNN의 높은 기밀성을 달성한다.

• JSTS:Journal of Semiconductor Technology and Science
• /
• 제1권2호
• /
• pp.125-131
• /
• 2001

When the program voltage is applied to a word line, a part of the boosted channel charge in inhibited bit lines is lost due to the coupling between the string select line (SSL) and the adjacent word line. This phenomenon causes the program disturbance in the cells connected to the inhibited bit lines. This program disturbance becomes more serious, as the word line pitch is decreased. To reduce the word line coupling, the rising edge of the word-line voltage waveform was changed from a pulse step into a ramp waveform with a controlled slope. The word-line ramping circuit was composed of a timer, a decoder, a 8 b D/A converter, a comparator, and a high voltage switch pump (HVSP). The ramping voltage was generated by using a stepping waveform. The rising time and the stepping number of the word-line voltage for programming were set to $\mutextrm{m}-$ and 8, respectively,. The ramping circuit was used in a 512Mb NAND flash memory fabricated with a $0.15-\mutextrm{m}$ CMOS technology, reducing the SSL coupling voltage from 1.4V into a value below 0.4V.

• 문화기술의 융합
• /
• 제7권1호
• /
• pp.615-619
• /
• 2021

이 연구는 사물인터넷, 클라우드 컴퓨팅 그리고 에지 컴퓨팅 등 많은 임베디드 시스템에서 성능 및 에너지 효율을 높이고자 최적화하는 메모리 시스템에 초점을 맞추어 그 성능 개선 기법을 제안한다. 제안하는 기법은 최근 많이 이용되고 있는 머신 러닝 알고리즘을 기반으로 메모리 시스템 성능을 도모한다. 머신 러닝 기법은 학습을 통하여 다양한 응용에 사용될 수 있는데, 메모리 시스템 성능 개선에서 사용되는 데이터의 분류 태스크에 적용될 수 있다. 정확도 높은 머신 러닝 기법 기반 데이터 분류는 데이터의 사용 패턴에 따라 데이터를 적절하게 배치할 수 있게 하여 전체 시스템 성능 개선을 도모할 수 있게 한다.

• Journal of Information Processing Systems
• /
• 제20권3호
• /
• pp.375-390
• /
• 2024

Edge computing architecture has effectively alleviated the computing pressure on cloud platforms, reduced network bandwidth consumption, and improved the quality of service for user experience; however, it has also introduced new security issues. Existing anomaly detection methods in big data scenarios with cloud-edge computing collaboration face several challenges, such as sample imbalance, difficulty in dealing with complex network traffic attacks, and difficulty in effectively training large-scale data or overly complex deep-learning network models. A lightweight deep-learning model was proposed to address these challenges. First, normalization on the user side was used to preprocess the traffic data. On the edge side, a trained Wasserstein generative adversarial network (WGAN) was used to supplement the data samples, which effectively alleviates the imbalance issue of a few types of samples while occupying a small amount of edge-computing resources. Finally, a trained lightweight deep learning network model is deployed on the edge side, and the preprocessed and expanded local data are used to fine-tune the trained model. This ensures that the data of each edge node are more consistent with the local characteristics, effectively improving the system's detection ability. In the designed lightweight deep learning network model, two sets of convolutional pooling layers of convolutional neural networks (CNN) were used to extract spatial features. The bidirectional long short-term memory network (BiLSTM) was used to collect time sequence features, and the weight of traffic features was adjusted through the attention mechanism, improving the model's ability to identify abnormal traffic features. The proposed model was experimentally demonstrated using the NSL-KDD, UNSW-NB15, and CIC-ISD2018 datasets. The accuracies of the proposed model on the three datasets were as high as 0.974, 0.925, and 0.953, respectively, showing superior accuracy to other comparative models. The proposed lightweight deep learning network model has good application prospects for anomaly traffic detection in cloud-edge collaborative computing architectures.

• 전자공학회논문지B
• /
• 제33B권10호
• /
• pp.120-129
• /
• 1996

This paper presents an efficient block segmentation and classification using the edge information of the document image. We extract four prominent features form the edge gradient and orientaton, all of which, and thereby the block clssifications, are insensitive to the background noise and the brightness variation of of the image. Using these four features, we can efficiently classify a document image into the seven categrories of blocks of small-size letters, large-size letters, tables, equations, flow-charts, graphs, and photographs, the first five of which are text blocks which are character-recognizable, and the last two are non-character blocks. By introducing the clumn interval and text line intervals of the document in the determination of th erun length of CRLA (constrained run length algorithm), we can obtain an efficient block segmentation with reduced memory size. The simulation results show that the proposed algorithm can rigidly segment and classify the blocks of the documents into the above mentioned seven categories and classification performance is high enough for all the categories except for the graphs with too much variations.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2000년도 추계학술대회 논문집 학회본부 B
• /
• pp.262-264
• /
• 2000

A numerical method for the analysis of 3D eddy current in conductors due to applied time varying field is suggested using the finite element method. In the approximation of the field quantifies, the edge element is used, because it reduce the required computer memory and the computing time compared with the nodal elements. With edge elements, furthermore, the field governing equations become simple because the electric scalar potential ${\phi}$ can be set to zero. The modified magnetic vector potential($A^*$) is used as a state variable. The analysed results are compared with the experimentally measured ones for the TEAM workshop problem3.