• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.7
• /
• pp.1749-1773
• /
• 2024

Recent advances in deep neural networks (DNNs) have greatly improved the accuracy and universality of various intelligent applications, at the expense of increasing model size and computational demand. Since the resources of end devices are often too limited to deploy a complete DNN model, offloading DNN inference tasks to cloud servers is a common approach to meet this gap. However, due to the limited bandwidth of WAN and the long distance between end devices and cloud servers, this approach may lead to significant data transmission latency. Therefore, device-edge collaborative inference has emerged as a promising paradigm to accelerate the execution of DNN inference tasks where DNN models are partitioned to be sequentially executed in both end devices and edge servers. Nevertheless, collaborative inference in heterogeneous edge environments with multiple edge servers, end devices and DNN tasks has been overlooked in previous research. To fill this gap, we investigate the optimization problem of collaborative inference in a heterogeneous system and propose a scheme CIS, i.e., collaborative inference scheme, which jointly combines DNN partition, task offloading and scheduling to reduce the average weighted inference latency. CIS decomposes the problem into three parts to achieve the optimal average weighted inference latency. In addition, we build a prototype that implements CIS and conducts extensive experiments to demonstrate the scheme's effectiveness and efficiency. Experiments show that CIS reduces 29% to 71% on the average weighted inference latency compared to the other four existing schemes.

• Journal of Intelligence and Information Systems
• /
• v.19 no.2
• /
• pp.39-54
• /
• 2013

Recent explosive increase of electronic commerce provides many advantageous purchase opportunities to customers. In this situation, customers who do not have enough knowledge about their purchases, may accept product recommendations. Product recommender systems automatically reflect user's preference and provide recommendation list to the users. Thus, product recommender system in online shopping store has been known as one of the most popular tools for one-to-one marketing. However, recommender systems which do not properly reflect user's preference cause user's disappointment and waste of time. In this study, we propose a novel recommender system which uses data mining and multi-model ensemble techniques to enhance the recommendation performance through reflecting the precise user's preference. The research data is collected from the real-world online shopping store, which deals products from famous art galleries and museums in Korea. The data initially contain 5759 transaction data, but finally remain 3167 transaction data after deletion of null data. In this study, we transform the categorical variables into dummy variables and exclude outlier data. The proposed model consists of two steps. The first step predicts customers who have high likelihood to purchase products in the online shopping store. In this step, we first use logistic regression, decision trees, and artificial neural networks to predict customers who have high likelihood to purchase products in each product group. We perform above data mining techniques using SAS E-Miner software. In this study, we partition datasets into two sets as modeling and validation sets for the logistic regression and decision trees. We also partition datasets into three sets as training, test, and validation sets for the artificial neural network model. The validation dataset is equal for the all experiments. Then we composite the results of each predictor using the multi-model ensemble techniques such as bagging and bumping. Bagging is the abbreviation of "Bootstrap Aggregation" and it composite outputs from several machine learning techniques for raising the performance and stability of prediction or classification. This technique is special form of the averaging method. Bumping is the abbreviation of "Bootstrap Umbrella of Model Parameter," and it only considers the model which has the lowest error value. The results show that bumping outperforms bagging and the other predictors except for "Poster" product group. For the "Poster" product group, artificial neural network model performs better than the other models. In the second step, we use the market basket analysis to extract association rules for co-purchased products. We can extract thirty one association rules according to values of Lift, Support, and Confidence measure. We set the minimum transaction frequency to support associations as 5%, maximum number of items in an association as 4, and minimum confidence for rule generation as 10%. This study also excludes the extracted association rules below 1 of lift value. We finally get fifteen association rules by excluding duplicate rules. Among the fifteen association rules, eleven rules contain association between products in "Office Supplies" product group, one rules include the association between "Office Supplies" and "Fashion" product groups, and other three rules contain association between "Office Supplies" and "Home Decoration" product groups. Finally, the proposed product recommender systems provides list of recommendations to the proper customers. We test the usability of the proposed system by using prototype and real-world transaction and profile data. For this end, we construct the prototype system by using the ASP, Java Script and Microsoft Access. In addition, we survey about user satisfaction for the recommended product list from the proposed system and the randomly selected product lists. The participants for the survey are 173 persons who use MSN Messenger, Daum Caf$\acute{e}$, and P2P services. We evaluate the user satisfaction using five-scale Likert measure. This study also performs "Paired Sample T-test" for the results of the survey. The results show that the proposed model outperforms the random selection model with 1% statistical significance level. It means that the users satisfied the recommended product list significantly. The results also show that the proposed system may be useful in real-world online shopping store.

• Environmental Engineering Research
• /
• v.14 no.2
• /
• pp.111-119
• /
• 2009

A prototype GIS-based decision support system (DSS) was developed by using a database management system (DBMS), a model management system (MMS), a knowledge-based system (KBS), a graphical user interface (GUI), and a geographical information system (GIS). The method of selecting a dispersion model or a modeling scheme, originally devised by Park and Seok, was developed using our GIS-based DSS. The performances of candidate models or modeling schemes were evaluated by using a single index(statistical score) derived by applying fuzzy inference to statistical measures between the measured and predicted concentrations. The fumigation dispersion model performed better than the models such as industrial source complex short term model(ISCST) and atmospheric dispersion model system(ADMS) for the prediction of the ground level $SO_2$ (1 hr) concentration in a coastal area. However, its coincidence level between actual and calculated values was poor. The neural network models were found to improve the accuracy of predicted ground level $SO_2$ concentration significantly, compared to the fumigation models. The GIS-based DSS may serve as a useful tool for selecting the best prediction model, even for complex terrains.

• Proceedings of the Korean Society for Agricultural Machinery Conference
• /
• 2000.11c
• /
• pp.513-520
• /
• 2000

An integrated grader which measures soluble solid content, color and weight of fresh apples was developed by NAMRI. The prototype grader consists of the near infrared spectroscopy and machine vision system. Image processing system and an algorithm to evaluate color were developed to speed up the color evaluation of apples. To avoid the light glare and specular reflection, an half-spherical illumination chamber was designed and fabricated to detect the color images of spherical-shaped apples more precisely. A color revision model based on neural network was developed. Near-infrared(NIR) spectroscopy system using NIR reflectance method developed by Lee et al(1998) of NAMRI was used to evaluate soluble solid content. In order to observe the performance of the grader, tests were conducted on conditions that there are 3 classes in weight sorting, 4 classes in combination of color and soluble solid content, and thus 12 classes in combined sorting. The average accuracy in weight, color and soluble solid content is more than about 90 % with the capacity of 3 fruits per second.

• The Transactions of the Korean Institute of Power Electronics
• /
• v.4 no.4
• /
• pp.357-367
• /
• 1999

When traction system of 2-motor driven electric vehicle(EV) is consisted of two motors (IPMSM) . two inverters. and one traction controller, control performances of IPMSM for an electric vehicle is affected by parameter variation b because of large current magnitude and wide current phase angle. To solve this problem, new parameter estimator for L Ld and Lq is constructed by neu때 network technique. And new vector control algorithm with parameter estimator by n neural network is proposed for IPMSM.And also. an advanced traction control algorithm is proposed using fuzzy c controller in order to enhance the driveability oftwo-wheel drive EVs with fitted with a traction control system Performances of the proposed algorithm are examined by simulations and the experimental resul않 with respect to t the prototype IPMSM and EV.

• Journal of Electrical Engineering and Technology
• /
• v.11 no.6
• /
• pp.1714-1728
• /
• 2016

This paper presents improved harmonics extraction based on Adaptive Linear Neuron (ADALINE) algorithm for single phase photovoltaic (PV) shunt active power filter (SAPF). The proposed algorithm, named later as Improved ADALINE, contributes to better performance by removing cosine factor and sum of element that are considered as unnecessary features inside the existing algorithm, known as Modified Widrow-Hoff (W-H) ADALINE. A new updating technique, named as Fundamental Active Current, is introduced to replace the role of the weight factor inside the previous updating technique. For evaluation and comparison purposes, both proposed and existing algorithms have been developed. The PV SAPF with both algorithms was simulated in MATLAB-Simulink respectively, with and without operation or connection of PV. For hardware implementation, laboratory prototype has been developed and the proposed algorithm was programmed in TMS320F28335 DSP board. Steady state operation and three critical dynamic operations, which involve change of nonlinear loads, off-on operation between PV and SAPF, and change of irradiances, were carried out for performance evaluation. From the results and analysis, the Improved ADALINE algorithm shows the best performances with low total harmonic distortion, fast response time and high source power reduction. It performs well in both steady state and dynamic operations as compared to the Modified W-H ADALINE algorithm.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.17 no.6
• /
• pp.1497-1508
• /
• 1993

Real-time mobile robot controllers usually have been designed focused on control theory without paying attention to the importance of system integration. This paper demonstrates that autonomous mobile robots require a real-time controller with a wide range of capabilities in addition to control theory. An architectural frame work supporting these capabilities has been designed in actual hardware environments. Individual modules such as a path planner, a path tracking controller, position estimators, wheel controllers and other cruical elements have been successfully integrated into the control system using this frame work. The overall performance of the system was investigated via a series of tracking experiments with a prototype mobile robot named LCAR deveoped in the laboratory. The context of the research involves the architecture, its implementation and experimental results.

• KIPS Transactions on Software and Data Engineering
• /
• v.7 no.11
• /
• pp.411-418
• /
• 2018

With the proliferation of artificial intelligence technology, it is becoming important to obtain, store, and utilize scientific data in research and science sectors. A number of methods for extracting meaningful objects such as graphs and tables from research articles have been proposed to eventually obtain scientific data. Existing extraction methods using heuristic approaches are hardly applicable to electronic documents having heterogeneous manuscript formats because they are designed to work properly for some targeted manuscripts. This paper proposes a prototype of an object extraction system which exploits a recent deep-learning technology so as to overcome the inflexibility of the heuristic approaches. We implemented our trained model, based on the Faster R-CNN algorithm, using the Google TensorFlow Object Detection API and also composed an annotated data set from 100 research articles for training and evaluation. Finally, a performance evaluation shows that the proposed system outperforms a comparator adopting heuristic approaches by 5.2%.

• Journal of IKEEE
• /
• v.26 no.3
• /
• pp.488-493
• /
• 2022

In manufacturing sites, bearing fault in eletrically driven motors cause the entire system to shut down. Stopping the operation of this environment causes huge losses in time and money. The reason of this bearing defects can be various factors such as wear due to continuous contact of rotating elements, excessive load addition, and operating environment. In this paper, a motor driving environment is created which is similar to the domestic manufacturing sites. In addition, based on the established environment, we propose a dataset for bearing fault detection by collecting changes in vibration characteristics that vary depending on normal and defective conditions. The sensor used to collect the vibration characteristics is Microphone G.R.A.S. 40PH-10. We used various machine learning models to build a prototype bearing fault detection system trained on the proposed dataset. As the result, based on the deep neural network model, it shows high accuracy performance of 92.3% in the time domain and 98.3% in the frequency domain.

• Journal of Biosystems Engineering
• /
• v.21 no.4
• /
• pp.414-421
• /
• 1996

한국과 일본의 경우 건표고를 외관의 품질상태 에 따라 12등급에서 16등급으로 구분하고 있다. 그리고 등급판정 작업은 임의로 추출한 샘플을 대상으로 전문 감정가에 의해 수작업으로 수행되고 있다. 건표고의 품질을 결정짓는 외관의 품질인자들은 갓과 내피에 고루 분포하고 있다. 본 논문에서는 컴퓨터 영상처리 시스템에 의거하여 개발한 건표고 자동 등급판정 및 선별 시작시스템의 구조와 기능 그리고 성능에 대하여 설명하였다. 개발한 시작시스템은 표고의 이송과 취급자동화를 위한 진동이송기, 반전장치, 컨베이어 이송장치와 두 세트의 컴퓨터 영상처리 시스템, 그리고 시스템 통괄제어를 위한 IBM PC AT호환 컴퓨터, 디지털 입출력 보드, 전공압실린더 구동제어를 위한 PLC등으로 구성하였다. 등급판정의 효율성 및 실시간 작업시스템을 고려하여 건표고의 등급판정은 두 세트의 컴퓨터 영상처리 시스템을 이용하여 이송되는 건표고의 갓 또는 내피 중 어디가 위를 향하는 지에 따라 두 단계에 걸쳐 독립적으로 판정을 수행하도록 하였다. 첫 번째 영상처리부에서는 갓표면 영상으로부터 4등급의 고품질 표고를 분류하며 두 번째 영상처리부에서는 내피표면 영상으로부터 중간 및 저품질 표고를 8개의 등급으로 분류한다. 실시간 영상정보처리를 목적으로 기존에 개발한 신경회로망을 이용한 등급판정 알고리즘을 시작시스템에 적용하였다. 개발한 시작기는 88% 이상의 등급판정 정확도를 보여 주었으며, 전공압시스템의 구동제약으로 인하여 표고 1개당 약0.7초의 선별시간이 소요되었다. 일조 선별라인의 경우 본 연구에서 제안한 시작기의 선별능력은 표고가 일차 처리부로 갓이 위로 올라와 있는 상태로 계속 공급된다면 시간당 대략 5,000여 개의 표고를 처리할 수 있을 것으로 기대된다.보강하여 가능하면 B-Pillar의 Middle이 Bending type collapse을 방지하여 Pelvis와 Door가 먼저 접촉하는 방법 등이 적용가능하다. 제작하기 이전에 설계된 부품에 대한 스프링 상수 및 내구특성을 체계적으로 규명하여 제품 시험의 횟수를 줄이고, 보다 정밀한 제품을 제작할 수 있도록 하기 위한 것이다.세포수는 초기 배반포기배에서 팽윤 배반포기배로 진행됨에 따라 두배에서 세배 정도 증가되었음을 알 수 있었다. 또한, differential labelling과 bisbenzimide기법에서 얻어진 각각의 총세포수를 비교하였을 때 총세포수는 발달의 진행 정도에 따라 증가되며 그와 동시에 동일한 군 간의 세포수도 거의 유사함을 알 수 있었다. 따라서, ICM과 TE를 differential labelling하는 기법은 수정란의 quality를 평가하는데 매우 유용한 기법으로서 착상전 embryo 발달을 연구하는데 효과적으로 이용될 수 있다는 것을 시사한다. 고도의 유의차를 나타낸 반면 비수구, 초생수로구 및 Bromegrass 목초구 간에는 아무런 유의차가 인정되지 않았다. 7. 농지보전 처리구인 배수구와 초생수로구는 비처리구에 비해 낮은 침두 유출량과 낮은 토양유실량을 나타내었다.구보다 14% 절감되는 것으로 나타났다.작용하는 것으로 사료된다.된다.정량 분석한 결과이다. 시편의 조성은 33.6 at% U, 66.4 at% O의 결과를 얻었다. 산화물 핵연료의 표면 관찰 및 정량 분석 시험시 시편 표면을 전도성 물질로 증착시키지 않고, Silver Paint 에 시편을 접착하는 방법으로도 만족한 시험 결과를 얻을 수 있었다.째, 회복기 중에 일어나는 입자들의 유입은 자기폭풍의 지속시간을 연장시키는 경향을 보이며 큰