• 한국정보처리학회논문지
• /
• 제3권1호
• /
• pp.9-32
• /
• 1996

본고는 분산 멀티미디어 시스템 환경에서 여러가지 응용 서비스를 지원하기 위한 범용 멀티미디어 데이타 처리 모델을 제시하고 그것을 구현하기 위한 객체지 향, 클 라이언트 서버 구조에 대하여 기술하였다. 이 모델은 분산 멀티미디어 시스템에서 필 요한 원격 데이타의 이용을 위한 통신망 투명성과 실시간 멀티미디어 입출력, 그리고 미디어의 통합이나 동기화 등의 멀티미디어 처리를 지원하는 범용 모델이다. 이 모델 은, 스트립 계층, 멀티미디어 프리젠테이션 계층, 하이퍼 프리젠테이션 계층등으로 구 성된다. 본 고는 이 모델의 각 계층에 해당되는 추상적인 데이타개념을 정의하고 이것 이 각 계층의 서비스와 실재로 객체지향 기법으로 제공되는 AP(Application Programming Interface)가 어떻게 연결되는지 기술하였다. 그리고 실제 구현하기 위한 구체적인 문제 를 예를 들면서 다루었고 마지막에 추후 연구해야 할 방향을 제시하며 결론을 맺었다.

• 한국인터넷방송통신학회논문지
• /
• 제23권4호
• /
• pp.115-119
• /
• 2023

현재의 인공지능에서 사용되는 자연어 처리 모델은 거대하여 실시간으로 데이터를 처리하고 분석하는 것은 여러가지 어려움들을 야기하고 있다. 이런 어려움을 해결하기 위한 방법으로 메모리를 적게 사용해 처리의 효율성을 개선하는 방법을 제안하고 제안된 모델의 성능을 확인하였다. 본 논문에서 제안한 모델의 성능평가를 위해 적용한 기법은 BERT[1] 모델의 어텐션 헤드 개수와 임베딩 크기를 작게 조절해 큰 말뭉치를 나눠서 분할 처리 후 출력값의 평균을 통해 결과를 산출하였다. 이 과정에서 입력 데이터의 다양성을 주기위해 매 에폭마다 임의의 오프셋을 문장에 부여하였다. 그리고 모델을 분류가 가능하도록 미세 조정하였다. 말뭉치를 분할 처리한 모델은 그렇지 않은 모델 대비 정확도가 12% 정도 낮았으나, 모델의 파라미터 개수는 56% 정도 절감되는 것을 확인하였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권2호
• /
• pp.807-825
• /
• 2020

In recent years, business environments have become more complex; therefore, enterprises must be capable of responding flexibly and agilely. For these purposes, effective enterprise systems service monitoring and early decision making based on the same, emerge as core competency of the enterprise. In addition, enterprise system techniques that filter meaningful data are needed to event processing. However, the existing study related with this is nothing but discovering of service faults by monitoring depending upon API of BPEL engine or middleware, or is nothing but processing of simple events based on low-level events. Accordingly, there would be limitations to provide useful business information. In this study, we present an extended event processing model that enables delivery of more valuable and useful business information through situation detection. Primarily, the event processing architecture in an enterprise system is proposed as a definite approach, and then define an event meta-model suitable for the proposed architecture. Based on the defined model, we propose the syntax and semantics of the elements that make up the event processing language include various and progressive event operators, the rules, complex event pattern, etc. In addition, an event context mechanism is proposed to analyze more delicate events. Finally, the effectiveness and applicability of proposed approach is presented through a case study.

• 한국IT서비스학회지
• /
• 제19권5호
• /
• pp.65-81
• /
• 2020

This study aims to examine the effects of users' perceptions of COVID-19 risk on their seeking and processing of relevant information as COVID-19 emerges and spreads worldwide in 2019. We apply the risk information seeking and processing model (RISP Model) to verify whether users' COVID-19 related information seeking and processing behaviors have a positive effect on their preventive and information sharing behaviors. To achieve this research goal, an online survey was conducted with about 400 of social media users. The users' perceptions of risk for COVID-19 increased their perceived insufficiency of COVID-19 information. In addition, the perceived insufficiency of users' information formed a positive relationship with seeking and searching of information behaviors. The processing of COVID-19 related information has increased related preventive behaviors and sharing of information through social media. While searching for information related to COVID-19 prompted personal information sharing behaviors, it did not significantly affect preventive behaviors. Accordingly, in order to promote COVID-19 preventive behaviors as well as overall user health-related behaviors it can be inferred that additional measures are needed in addition to pursuing relevant information.

• 한국공작기계학회:학술대회논문집
• /
• 한국공작기계학회 2001년도 춘계학술대회 논문집(한국공작기계학회)
• /
• pp.192-196
• /
• 2001

This work describes the modeling of cutting process for feedback control based on signal processing in end-milling. Here, cutting force is used to design control model by a variety of schemes which are moving average, ensemble average, peak value, root mean square and analog low-pass filtering. It is expected that each model offers its own peculiar advantage in following cutting force control.

• 한국소성가공학회:학술대회논문집
• /
• 한국소성가공학회 1994년도 춘계학술대회 논문집
• /
• pp.159-166
• /
• 1994

A constitutive model was proposed to analyze creep densification and grain growth of alumina powder compacts during high temperature processing. Theoretical results from the constitutive model were compared with various experimental data of alumina powder compacts in the literature including pressureless sintering, sinter forging and hot pressing. The proposed constitutive equations were implemented into finite element analysis program (ABAQUS) to simulate densification for more complicated geometry and loading conditions. High temperature forming processing of alumina compact with complicated shape was simulated. Processing of Alumina Powder Compacts

• 전자공학회논문지B
• /
• 제29B권1호
• /
• pp.11-17
• /
• 1992

This paper is to propose an analytical queuing model for parallel processing of sequential program with nested loops. The analytical results are compared with the results from the implemented multiprocessor system composed of four intel 8088 microprocessor, eight 2KB shared common memories, and a hardware token ring. At results, this study shows that the processed results are almost similar in proposed analytical model and real system. Proposed analytical model can be applied to evaluate parallel processing of sequential program with nested loops.

• 한국콘텐츠학회논문지
• /
• 제18권1호
• /
• pp.465-475
• /
• 2018

최근 소셜 미디어, IoT 등에 대한 활용이 증가됨에 따라 대용량의 그래프 스트림이 생성되고 있으며 그래프 스트림을 실시간으로 처리하기 위한 많은 연구들이 진행되고 있다. 본 논문에서는 그래프가 지속적으로 변경될 때 이전 결과 데이터를 재사용하는 점진적인 그래프 스트림 처리 기법을 제안한다. 또한, 점진적 처리와 정적인 처리를 선택적으로 수행하기 위한 비용 모델을 제안한다. 제안하는 비용 모델은 실제 처리된 이력을 바탕으로 재계산 영역의 탐색 비용 및 처리 비용의 예측 값을 계산하여 점진적 처리가 정적인 처리보다 이득인 경우 점진적 처리를 수행한다. 제안하는 점진적 처리는 그래프 갱신이 발생하면 변경되는 부분만을 처리하여 효율성을 증가시킨다. 또한, 변경되는 부분의 이전 결과 데이터만을 수집하여 점진적인 처리를 수행함으로써 디스크 I/O 비용을 감소시킨다. 다양한 성능평가를 통해 제안하는 기법이 기존 기법에 비해 성능이 우수함을 보인다.

• Journal of Information Processing Systems
• /
• 제15권6호
• /
• pp.1378-1391
• /
• 2019

In recent years, automatic document summarization has been widely studied in the field of natural language processing thanks to the remarkable developments made using deep learning models. To decode a word, existing models for abstractive summarization usually represent the context of a document using the weighted hidden states of each input word when they decode it. Because the weights change at each decoding step, these weights reflect only the local context of a document. Therefore, it is difficult to generate a summary that reflects the overall context of a document. To solve this problem, we introduce the notion of a general context and propose a model for summarization based on it. The general context reflects overall context of the document that is independent of each decoding step. Experimental results using the CNN/Daily Mail dataset show that the proposed model outperforms existing models.

• 산업경영시스템학회지
• /
• 제18권33호
• /
• pp.29-37
• /
• 1995

An Automated Storage/Retrieval Systems(AS/RS) has been gradually emphasized because of the change of production and distribution environment. This paper develops algorithm and Shortest Distance Model that can reduce the traveling time of a stacker crane for efficient operation of AS/RS. In order to reduce the traveling time of a stacker crane, we determine the order processing and then the sequencing of storage/retrieval for each item. Order processing is determined based on the SPT(Shortest Processing Time) concept considering a criterion of retrieval coordinate. The sequencing of storage/retrieval is determined based on the Shortest Distance Model by using a modified SPP(Shortest Path Problem) of network problem. A numerical example is provided to illustrate the developed algorithm and Shortest Distance Model.