• 제목/요약/키워드: ICT Training

검색결과 226건 처리시간 0.029초

자율주행 영상데이터의 신뢰도 향상을 위한 AI모델 기반 데이터 자동 정제 (AI Model-Based Automated Data Cleaning for Reliable Autonomous Driving Image Datasets)

  • 김가나;김학일
    • 방송공학회논문지
    • /
    • 제28권3호
    • /
    • pp.302-313
    • /
    • 2023
  • 본 연구는 과학기술정보통신부가 2017년부터 1조원 이상을 투자한 'AI Hub 댐' 사업에서 구축된 인공지능 모델 학습데이터의 품질관리를 자동화할 수 있는 프레임워크의 개발을 목표로 한다. 자율주행 개발에 사용되는 AI 모델 학습에는 다량의 고품질의 데이터가 필요하며, 가공된 데이터를 검수자가 데이터 자체의 이상을 검수하고 유효함을 증명하는 데는 여전히 어려움이 있으며 오류가 있는 데이터로 학습된 모델은 실제 상황에서 큰 문제를 야기할 수 있다. 본 논문에서는 이상 데이터를 제거하는 신뢰할 수 있는 데이터셋 정제 프레임워크를 통해 모델의 인식 성능을 향상시키는 전략을 소개한다. 제안하는 방법은 인공지능 학습용 데이터 품질관리 가이드라인의 지표를 기반으로 설계되었다. 한국정보화진흥원의 AI Hub을 통해 공개된 자율주행 데이터셋에 대한 실험을 통해 프레임워크의 유효성을 증명하였고, 이상 데이터가 제거된 신뢰할 수 있는 데이터셋으로 재구축될 수 있음을 확인하였다.

교육 수요 조사를 통한 스마트 수산 양식 분야의 융합형 인재 양성 교육 프로그램 개발 방향 설정을 위한 탐색적 연구 (An exploratory study on establishment of a development direction on education training program for cultivating convergence human resources in smart aquaculture through a demand survey)

  • 권인영;김태호
    • 수산해양기술연구
    • /
    • 제56권3호
    • /
    • pp.265-276
    • /
    • 2020
  • The objective of this study is to develop education programs for cultivating smart aquaculture experts through a education demand survey of industries, high school students, university (graduate) students and field workers. The industry demand analysis was conducted as an in-depth interview on representives from seven companies. Education demand surveys were conducted on 96 students and field workers in the Jeonnam region. Results on the demand survey were analyzed using frequency analysis and cross-analysis. The company representatives responded that they want to participated in internship and retraining programs to proactively secure manpowers with convergence capabilities about smart aquaculture. Seven companies preferred manpowers with basic competencies on ICT (Information and Communications Technologies) or aquaculture. The most respondents in the demand survey want to participate in the education program for experience on advanced technology, self-development and enhancement of work capability. On the other hand, some respondents said that the education is time-consuming and that the education program does not fit their level. Thus, the education program should be developed in a way to minimize the spatial and temporal limitations of education targets and to improve understanding of non-majors by reflecting the demands of human resources in the industrial field.

국제중재의 산업별 유의성 연구 (A Study of Industrial Significance on International Arbitration)

  • 손승표;김기홍
    • 한국중재학회지:중재연구
    • /
    • 제27권4호
    • /
    • pp.115-131
    • /
    • 2017
  • The purpose of this paper is to consider key issues and to categorize international arbitrations on industrial significance. Thus this study focused on synthesizing industrial significance, retaining international competitiveness, and securing national implications of international arbitration as a legal service trade system. To obtain these goals, the liberalization of international arbitration proxy service, invitation of international arbitration board regional offices such as PCA to AAA-ICDR, and drawing of international mediation institutions to invigorate the legal service market should be exerted. To revitalize the international arbitration, it would be renounced the control rights of economy-related disputes with internationalism while policies should be improved to enhance the execution of arbitration agreements and awards. It is emphasized the necessities of the process through the institutional mediation agency to resolve more court mediation and international dispute cases, and to handle international arbitration cases after training professional legal experience through cooperation with international arbitration boards such as LCIA, ICC, AAA, and PCA. Ultimately, to revitalize the international arbitration, the industrial significance of arbitration should be analyzed and the critical industrial influence of arbitration such as in the semiconductor, petrochemistry, and ICT sectors should be expanded to gain the competitiveness of the global legal service market with the assistance of institutional complementary measures.

도시재생을 위한 건물공간관리 플랫폼 사용자 역량 교육에 관한 연구 (A Study on User Competency Training for Building Space Management Platform for Urban Regeneration)

  • 강현주;김치용
    • 한국멀티미디어학회논문지
    • /
    • 제23권3호
    • /
    • pp.499-507
    • /
    • 2020
  • This paper is one of the types of 'sharing economy', a social economic model that borrows and shares things, spaces, and services from the development of the 4th Industrial Revolution. Through sharing of empty space and time in the city, I would like to suggest a way to reduce the closure of small business owners in order to create jobs, which is one of urban problems in the community. We also build a platform that utilizes the free time and space of buildings through space sharing, one of the types of sharing economy, and provides education programs for start-up education, promotion, marketing, and consulting by matching small business owners with building owners. Therefore, in this paper, by sharing the space and time, the landlord and the small business can share the profits of the small business by reducing the business owner's closure and the job creation plan. Coaching urban regeneration was proposed.

다차원 데이터에 대한 심층 군집 네트워크의 성능향상 방법 (Performance Improvement of Deep Clustering Networks for Multi Dimensional Data)

  • 이현진
    • 한국멀티미디어학회논문지
    • /
    • 제21권8호
    • /
    • pp.952-959
    • /
    • 2018
  • Clustering is one of the most fundamental algorithms in machine learning. The performance of clustering is affected by the distribution of data, and when there are more data or more dimensions, the performance is degraded. For this reason, we use a stacked auto encoder, one of the deep learning algorithms, to reduce the dimension of data which generate a feature vector that best represents the input data. We use k-means, which is a famous algorithm, as a clustering. Sine the feature vector which reduced dimensions are also multi dimensional, we use the Euclidean distance as well as the cosine similarity to increase the performance which calculating the similarity between the center of the cluster and the data as a vector. A deep clustering networks combining a stacked auto encoder and k-means re-trains the networks when the k-means result changes. When re-training the networks, the loss function of the stacked auto encoder and the loss function of the k-means are combined to improve the performance and the stability of the network. Experiments of benchmark image ad document dataset empirically validated the power of the proposed algorithm.

디지털 매체를 활용한 일러스트레이션 교육 방안에 관한 연구 (A Study on Instructional Methods for Illustration Using Digital Media)

  • 윤은재
    • 복식
    • /
    • 제64권6호
    • /
    • pp.120-130
    • /
    • 2014
  • The present study aimed to find out the most effective instructional method for illustration to aid the $21^{st}$-century post digital generation learners in their education. Although the literature discusses government-led enactment and revision of educational policy and curriculums and research reports by the Ministry of Education and other government agencies as well as many other research articles deal with futuristic instructional methods, the instructional methods in practice have not fully changed yet. Illustration is a means of visual communication and a subject that best represent 'expressionism' and 'speed' of the post digital generation. The present study exemplifies instructional methods to establish the need for lesson plans that best suit individuality of students and for re-training instructors. Also, the feedback process with students proved that the post digital generation learned Photoshop quickly and improved creativity via image synthesis even in short-term instructions. The present findings can be possibly used as reference data for instructional methods for illustration using digital media that best suits learners of the post digital generation. This being said, further studies should delve into specific instructional methods.

Utilization of Open Access Repositories for Visibility of Academic Publications by Lecturers in South-East, Nigeria

  • Orsu, Nkem Emilia
    • International Journal of Knowledge Content Development & Technology
    • /
    • 제9권4호
    • /
    • pp.47-68
    • /
    • 2019
  • The study investigates the utilization of open access repositories for visibility of academic publications by lecturers in South-East, Nigeria. The specific objectives were to investigate the types of academic publications available, platforms used in self-archiving, level of satisfaction derived by lecturers with the contribution of open access repositories, challenges and strategies. It adopted descriptive survey design. Proportionate stratified sampling technique was used to draw a sample size of 418 from the total population of 4,187 lecturers. Questionnaire and observation check list were used for data collection. Frequency count, percentage, and mean score were used to analyze the data. The study revealed that there are enough academic publications available such as books, journals, projects, and thesis with the exception of handouts. It also revealed that Google Scholar, Research Gate and Open Access Journal were the main platforms lecturers use in self-archiving; while other platforms like Institutional Repositories, among others were minimally employed. Lecturers are less satisfied with the contribution of open access repositories. All the suggested challenges such as inadequate power supply, poor network, lack of awareness and inadequate research grants were all accepted. Similarly, all the suggested strategies for improving them were accepted. Recommendations were made based on the findings such as more awareness creation on importance of open access repositories; re-training of lecturers and provision of adequate ICT infrastructures that will improve the utilization of open access repositories by lecturers which could enhance the global visibility of academic publications from the Nigerian Universities.

A Study on the National Spatial Data Infrastructure of U.S.A

  • Koh, June-Hwan
    • 한국측량학회지
    • /
    • 제25권6_1호
    • /
    • pp.485-497
    • /
    • 2007
  • By the rapid development of Information Communication Technology (ICT) and Geo-spatial Technology (GT) and the increased usage of spatial data for planning and infrastructure management, the National Geographic Information System (NGIS) for more efficient and effective utilization of spatial information has been developed by the central government in Korea since 1995. NGIS is the base of Spatial Data Infrastructure (SDI). SDI is developed as one of National Information Infrastructures (NII). Among the hierarchy of SDI, National Spatial Data Infrastructure (NSDI) has very important role in the success of SDI development. Many research articles show that the USA's NSDI initiatives, development strategy have been strongly influenced all over the world. In these viewpoints, to propose the future directions of Korean NGIS, the development of NSDI strategy of USA is reviewed by literature through published book and internet resources. The conclusions of this study are as follow: 1) top-down and bottom-up approach are needed for integrated data sharing and standardization. 2) the creative and evolutionary vision and strategy has to be suggested. 3) the training program and lecture material has to be developed and diffused to the users and providers of spatial data. 4) governance system has to be built for NSDI evaluation. 5) the formation of geo-spatial forum to discuss the spatial-related problems and make research agenda, etc.

스마트폰의 3축 가속도 센서를 이용한 농구 자세 인식 (Basket ball motion recognition using a 3-axis accelerometer sensor of smart phone)

  • 호종갑;이상준;왕창원;정화영;나예지;민세동
    • 한국정보처리학회:학술대회논문집
    • /
    • 한국정보처리학회 2015년도 추계학술발표대회
    • /
    • pp.1372-1374
    • /
    • 2015
  • 본 논문에서는 농구 경기에서의 대표적 자세 중 Standing shoot, Jump shoot, Pass, Dribble, Lay-up shoot, 총 5가지 자세를 인식하기 위해 각 자세와 3축 가속도 값과의 상관관계를 보여주고 있다. 스마트폰에 내장되어 있는 가속도 센서로부터 데이터를 생성해주는 어플리케이션인 Sensor log를 활용하여 얻은 3축 가속도 값으로 수직, 수평축과 3축 가속도의 크기를 구해 Instance로 사용하였다. 위 데이터는 대표적인 데이터 마이닝 도구인 Weka tool을 이용하여 각 모션과 데이터 값의 상관관계를 확인하였고, 실험 결과 10-fold에서 평균 59.8%를 보였으나 Training set과 Test set의 결과 80.8%를 보였다.

발전소 안전교육 가상현실 콘텐츠 개발 방법 연구 - 밀폐공간 및 용접화재를 중심으로 - (Study on Development Method of VR Training Contents for Power Plant Safety Education - Focused on sealed space and welding fire -)

  • 민설희;박봉성;김유식;이성기;이동준;강기헌
    • 한국콘텐츠학회:학술대회논문집
    • /
    • 한국콘텐츠학회 2019년도 춘계종합학술대회
    • /
    • pp.367-368
    • /
    • 2019
  • 여러 가지 안전사고 항목 중에 밀폐공간 작업과 관련된 사고는 다른 사고에 비해 사망으로 이어질 가능성이 매우 높으며, 사고 발생 시 2명 이상이 동시에 사망하거나 부상당할 가능성이 높은 사고다. 안전보건공단 자료에 의하면, 사고자 2명 중 1명이 밀폐공간 질식사고자이다. 또한 용접작업은 발전소 작업 현장에서는 보일러 튜브 교체 및 수리 등 다양한 부분에서 많이 사용되고 있다. 특히 발전소는 위험물질을 다량 보유하고 있는 곳으로, 고열, 불티에 의한 화재폭발은 대형 인명사고로 이어질 가능성이 매우 높다. 그러나 밀폐공간 작업 및 용접작업에 대한 안전교육은 직접 체험이 어렵기 때문에 효과적인 교육 방법에 대한 연구가 필요하다. 그러므로 본 연구에서는 가상현실 기술을 적용하여 발전소 작업 현장 인원을 대상으로 안전교육이 가능한 밀폐공간 작업 및 용접작업 시 일어날 수 있는 일을 체험할 수 있도록 한 가상현실 안전교육 콘텐츠를 개발한 내용에 대해 살펴보고자 한다.

  • PDF