• Journal of KIISE:Computer Systems and Theory
• /
• v.31 no.5_6
• /
• pp.268-278
• /
• 2004

In this paper, we describe our experience in the design and implementation of the SMART file system to handle multimedia workload. Our work has three design objectives: (ⅰ) efficient support for sequential workload, (ⅱ) avoiding disk fragmentation, (ⅲ) logical unit based file access. To achieve these three objectives, we develop a file system where a file consists of linked list of Data Unit Group. Instead of tree like structure of the legacy Unix file system, we use single level file structure. Our file system can also access the file based upon the logical unit which can be video frame or audio samples. Data Unit Group is a group of logical data units which is allocated continuous disk blocks. At the beginning of each Data Unit Group, there exists an index array. Each index points to the beginning of logical data units, e.g. frames in the Data Unit Group. This index array enables the random access and sequencial access of semantic data units. SMART file system is elaborately tailored to effectively support multimedia workload. We perform physical experiments and compare the performance of SMART file system with EXT2 file system and SGI XFS file system. In this experiment, SMART file system exhibits superior performance under streaming workload.

• Journal of KIISE:Software and Applications
• /
• v.26 no.11
• /
• pp.1261-1269
• /
• 1999

다층 퍼셉트론은 다양한 응용 분야에 성공적으로 적용되고 있는 대표적인 신경회로망 모델이다. 그러나 다층 퍼셉트론의 학습에서 나타나는 플라토에 기인한 느린 학습 속도와 지역 극소는 실제 응용문제에 적용함에 있어서 가장 큰 문제로 지적되어왔다. 이 문제를 해결하기 위해 여러 가지 다양한 학습알고리즘들이 개발되어 왔으나, 계산의 비효율성으로 인해 실제 문제에는 적용하기 힘든 예가 많은 등, 현재까지 만족할 만한 해결책은 제시되지 못하고 있다. 본 논문에서는 다층퍼셉트론의 베이시스 함수로 사용되는 시그모이드 함수를 보다 일반화된 형태로 정의하여 사용함으로써 학습에 있어서의 플라토를 완화하고, 지역극소에 빠지는 것을 줄이는 접근방법을 소개한다. 본 방법은 기존의 변형된 가중치 수정식을 사용한 학습 속도 향상의 방법들과는 다른 접근 방법을 택함으로써 기존의 방법들과 함께 사용하는 것이 가능하다는 특징을 갖고 있다. 제안하는 방법의 성능을 확인하기 위하여 간단한 패턴 인식 문제들에의 적용 실험 및 기존의 학습 속도 향상 방법을 함께 사용하여 시계열 예측 문제에 적용한 실험을 수행하였고, 그 결과로부터 제안안 방법의 효율성을 확인할 수 있었다. Abstract A multilayer perceptron is the most well-known neural network model which has been successfully applied to various fields of application. Its slow learning caused by plateau and local minima of gradient descent learning, however, have been pointed as the biggest problems in its practical use. To solve such a problem, a number of researches on learning algorithms have been conducted, but it can be said that none of satisfying solutions have been presented so far because the problems such as computational inefficiency have still been existed in these algorithms. In this paper, we propose a new learning approach to minimize the effect of plateau and reduce the possibility of getting trapped in local minima by generalizing the sigmoidal function which is used as the basis function of a multilayer perceptron. Adapting a new approach that differs from the conventional methods with revised updating equation, the proposed method can be used together with the existing methods to improve the learning performance. We conducted some experiments to test the proposed method on simple problems of pattern recognition and a problem of time series prediction, compared our results with the results of the existing methods, and confirmed that the proposed method is efficient enough to apply to the real problems.

• KIPS Transactions on Computer and Communication Systems
• /
• v.8 no.9
• /
• pp.217-224
• /
• 2019

Recently, sensor networks are built and used on many kinds of fields such as home, traffic, medical treatment and power grid. Sensing data manipulation on these fields could be a serious threat on property and safety. Thus, a proper way to block sensing data manipulation is necessary. In this paper, we propose IoT(Internet of Things) sensing data validation mechanism based on data obfuscation and variance analysis to remove manipulated sensing data effectively. IoT sensor device modulates sensing data with obfuscation function and sends it to a user. The user demodulates received data to use it. Fake data which are not modulated with proper obfuscation function show different variance aspect with valid data. Our proposed mechanism thus can detect fake data by analyzing data variance. Finally, we measured data validation time for performance analysis. As a result, block rate for false data was improved by up to 1.45 times compared with the existing technique and false alarm rate was 0.1~2.0%. In addition, the validation time on the low-power, low-performance IoT sensor device was measured. Compared to the RSA encryption method, which increased to 2.5969 seconds according to the increase of the data amount, the proposed method showed high validation efficiency as 0.0003 seconds.

• Journal of Convergence for Information Technology
• /
• v.9 no.12
• /
• pp.141-146
• /
• 2019

Economic analysis predicts that the drone market will grow, and the growth of the toy and hobby drone market is expected to gradually expand. Drone expectations are rising due to the net economic function of drone market growth, but accidents due to improper management and operations are also increasing. The difference in toy drone performance is incomparably small compared to industrial drone performance, but the ordinary buyer can not know whether the difference can cause an accident during use. The toy drones used in this study were obtained from KC and CE certification, and 20 kinds of drones were used. The flight time ranged from a minimum of 3 minutes to a maximum of 12 minutes, and the control distance ranged from a minimum of 20m to a maximum of 380m. Therefore, it is necessary to secure product safety through sampling inspection of the radio wave output of toy drones, and it is also necessary to mount an algorithm that automatically lowers the altitude or hover when exceeding the limit flight distance. For future research, we will build data to establish toy drone safety standards through a altitude testing and impact testing of toy drone.

• The Journal of the Acoustical Society of Korea
• /
• v.40 no.5
• /
• pp.530-536
• /
• 2021

Recently, due to the development of deep learning, end-to-end speech recognition, which directly maps graphemes to speech signals, shows good performance. Especially, among the end-to-end models, conformer shows the best performance. However end-to-end models only focuses on the probability of which grapheme will appear at the time. The decoding process uses a greedy search or beam search. This decoding method is easily affected by the final probability output by the model. In addition, the end-to-end models cannot use external pronunciation and language information due to structual problem. Therefore, in this paper conformer with lexicon transducer is proposed. We compare phoneme-based model with lexicon transducer and grapheme-based model with beam search. Test set is consist of words that do not appear in training data. The grapheme-based conformer with beam search shows 3.8 % of CER. The phoneme-based conformer with lexicon transducer shows 3.4 % of CER.

• Korean Journal of Remote Sensing
• /
• v.37 no.6_1
• /
• pp.1709-1718
• /
• 2021

The waveform data of the Airborne Bathymetric LiDAR (ABL; LiDAR: Light Detection And Ranging) system provides data with improved accuracy, resolution, and reliability compared to the discrete-return data, and increases the user's control over data processing. Furthermore, we are able to extract additional information about the return signal. Waveform decomposition is a technique that separates each echo from the received waveform with a mixture of water surface and seabed reflections, waterbody backscattering, and various noises. In this study, a new waveform decomposition technique based on a Gaussian model was developed to improve the point extraction performance from the ABL waveform data. In the existing waveform decomposition techniques, the number of decomposed echoes and decomposition performance depend on the peak detection results because they use waveform peaks as initial values. However, in the study, we improved the approximation accuracy of the decomposition model by adding the estimated potential peak candidates to the initial peaks. As a result of an experiment using waveform data obtained from the East Coast from the Seahawk system, the precision of the decomposition model was improved by about 37% based on evaluating RMSE compared to the Gaussian decomposition method.

• The Journal of Korean society of community based occupational therapy
• /
• v.8 no.3
• /
• pp.1-12
• /
• 2018

Objective : This study conducted a preliminary study to develop a occupational therapy model focused on improving living functions within the community care system. Methods : From June to July, 2018, the literature on community care was researched, focusing on cases of Japan's Management Tool for Daily Life Performance (MTDLP), Sweden, United Kingdom, Germany and domestic S Elderly Care Centers and I Health Centers. Based on this information, a group of experts developed a occupational therapy model within the community care system. Results : Assessment tools such as occupation-based health promotional table, interest checklist, occupational goals for improving living functions, sheet for evaluation of living functions, survey of daily life time (weekday and weekend), and sheet for transition of living functions were developed to conduct evaluation for occupational therapy. The improving living functions program, analysis of activities based on ICF model, lifestyle redesign program, cognitive exercise therapy, the Lee Silverman Voice Treatment (LSVT), hospice, and home modification were also organized interventions already in place by occupational therapists. Conclusion : This study showed specific measures and models for the implementation of occupational therapy within community care systems. Occupational therapy is positioned as a specialized area that is essential to the client, and we look forward to the use of this model.

• Tunnel and Underground Space
• /
• v.31 no.6
• /
• pp.494-507
• /
• 2021

With the increasing use of TBM, research has recently been conducted in Korea to analyze TBM data with machine learning techniques to predict the ground in front of TBM, predict the exchange cycle of disk cutters, and predict the advance rate of TBM. In this study, classification prediction of rock characteristics of slurry shield TBM sites was made by combining traditional rock classification techniques and machine learning techniques widely used in various fields with machine data during TBM excavation. The items of rock characteristic classification criteria were set as RQD, uniaxial compression strength, and elastic wave speed, and the rock conditions for each item were classified into three classes: class 0 (good), 1 (normal), and 2 (poor), and machine learning was performed on six class algorithms. As a result, the ensemble model showed good performance, and the LigthtGBM model, which showed excellent results in learning speed as well as learning performance, was found to be optimal in the target site ground. Using the classification model for the three rock characteristics set in this study, it is believed that it will be possible to provide rock conditions for sections where ground information is not provided, which will help during excavation work.

• Tunnel and Underground Space
• /
• v.32 no.6
• /
• pp.502-517
• /
• 2022

As the use of TBM increases, research has recently increased to to analyze TBM data with machine learning techniques to predict the exchange cycle of disc cutters, and predict the advance rate of TBM. In this study, a regression prediction of disc cutte wear of slurry shield TBM site was made by combining machine learning based on the machine data and the geotechnical data obtained during the excavation. The data were divided into 7:3 for training and testing the prediction of disc cutter wear, and the hyper-parameters are optimized by cross-validated grid-search over a parameter grid. As a result, gradient boosting based on the ensemble model showed good performance with a determination coefficient of 0.852 and a root-mean-square-error of 3.111 and especially excellent results in fit times along with learning performance. Based on the results, it is judged that the suitability of the prediction model using data including mechanical data and geotechnical information is high. In addition, research is needed to increase the diversity of ground conditions and the amount of disc cutter data.

• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.8
• /
• pp.371-380
• /
• 2023

Relation extraction is to extract relationships between named entities from text. Traditionally, relation extraction methods only extract relations between predetermined subject and object entities. However, in end-to-end relation extraction, all possible relations must be extracted by considering the positions of the subject and object for each pair of entities, and so this method uses time and resources inefficiently. To alleviate this problem, this paper proposes a method that sets directions based on the positions of the subject and object, and extracts relations according to the directions. The proposed method utilizes existing relation extraction data to generate direction labels indicating the direction in which the subject points to the object in the sentence, adds entity position tokens and entity type to sentences to predict the directions using a pre-trained language model (KLUE-RoBERTa-base, RoBERTa-base), and generates representations of subject and object entities through probabilistic crossover operation. Then, we make use of these representations to extract relations. Experimental results show that the proposed model performs about 3 ~ 4%p better than a method for predicting integrated labels. In addition, when learning Korean and English data using the proposed model, the performance was 1.7%p higher in English than in Korean due to the number of data and language disorder and the values of the parameters that produce the best performance were different. By excluding the number of directional cases, the proposed model can reduce the waste of resources in end-to-end relation extraction.