• 한국정보과학회논문지:시스템및이론
• /
• 제30권5_6호
• /
• pp.213-222
• /
• 2003

일반적으로 하이퍼볼릭 공간상에서 대칭 패턴을 생성하는 알고리즘은 벡터표현 방식에 기반한 포워드 매핑 알고리즘을 사용한다. 기존의 알고리즘에서는 복사한 대칭 패턴을 표현하는 레이어의 증가에 따라 메모리의 사용이 기하급수적으로 증가한다 이러한 문제점으로 인해 전체 패턴의 정밀한 표현이 불가능하다. 또한 기본 패턴으로 비트맵 이미지를 사용하기 어렵고 벡터형태의 결과를 이미지로 변환하는 추가의 처리를 필요로 한다. 본 논문에서는 하이퍼볼릭 공간에서 대칭 패턴을 생성하는데 있어 정밀하고도 효율적인 계산 방법인 백워드 매핑 알고리즘을 제안한다.

• 감성과학
• /
• 제5권1호
• /
• pp.79-90
• /
• 2002

기억은 상황인식을 개발하고 유지하는 데 중요한 역할을 한다. 본 연구에서는 비행상황 인식의 민감도에 영향을 주는 기억과정들이 조종사의 전문성 정도에 따라 어떻게 다른지를 조사하였다. 실험 1에서는 조종사가 상황정보를 회상하기 위해 사용하는 여러 기억과정들을 비교하였다. 즉, 조종실 계기판에 제시되는 상이한 형태(공간적 및 언어적)의 상황정보에 관한 조종사의 기억을 즉각회상과 지연 회상법으로 평가하였다. 실험 2에서는 조종사의 상황인식 민감도를 측정하였고, 그 민감도와 실험 1에서 수집한 여러 회상기억 측정치들과의 상관관계를 조사하였다. 두 실험결과를 요약하면, 전문성 효과가 상황정보의 활동기억보다는 장기기억에서 나타났고, 상황인식 민감도를 높이는 데 요구되는 상황정보의 기억표상이 전문성 정도에 따라 달랐다. 실험결과의 이론적 설명이 심리학의 전문성 이론을 중심으로 논의되었다.

• International Journal of Advanced Culture Technology
• /
• 제5권1호
• /
• pp.83-89
• /
• 2017

In this paper, we propose an extended method of one-hot Long Short-Term Memory (LSTM) and evaluate the performance on spam filtering task. Most of traditional methods proposed for spam filtering task use word occurrences to represent spam or non-spam messages and all syntactic and semantic information are ignored. Major issue appears when both spam and non-spam messages share many common words and noise words. Therefore, it becomes challenging to the system to filter correct labels between spam and non-spam. Unlike previous studies on information filtering task, instead of using only word occurrence and word context as in probabilistic models, we apply a neural network-based approach to train the system filter for a better performance. In addition to one-hot representation, using term weight with attention mechanism allows classifier to focus on potential words which most likely appear in spam and non-spam collection. As a result, we obtained some improvement over the performances of the previous methods. We find out using region embedding and pooling features on the top of LSTM along with attention mechanism allows system to explore a better document representation for filtering task in general.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제6권6호
• /
• pp.1546-1564
• /
• 2012

Pattern-based query processing has not attracted much attention in wireless sensor network though its counterpart has been studied extensively in data stream. The methods used for data stream usually consume large memory and much energy. This conflicts with the fact that wireless sensor networks are heavily constrained by their hardware resources. In this paper, we use piece wise representation to represent sensor nodes' collected data to save sensor nodes' memory and to reduce the energy consumption for query. After getting data stream's and patterns' approximated line segments, we record each line's slope. We do similar matching on slope sequences. We compute the dynamic time warping distance between slope sequences. If the distance is less than user defined threshold, we say that the subsequence is similar to the pattern. We do experiments on STM32W108 processor to evaluate our strategy's performance compared with naive method. The results show that our strategy's matching precision is less than that of naive method, but our method's energy consumption is much better than that of naive approach. The strategy proposed in this paper can be used in wireless sensor network to process pattern-based queries.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제36권11호
• /
• pp.947-954
• /
• 2009

본 논문은 입력받은 다수 이미지의 Visual Hull을 이용하여 Surfel 기반의 물체를 복원하는 새로운 방법을 제안한다. Sufel은 포인트 집합면을 효과적으로 근사하여 나타내는 포인트 표현 단위이다. LDC(Layered Depth Cube) Surfel 샘플링 방식에 입력 이미지로부터 근사되는 형상을 대표하는 Visual Hull의 개념을 연동하여 이미지로부터 물체의 Surfel 기하 정보를 생성할 수 있다. Surfel 표현 방식은 기존의 폴리곤 기반 방식에 비해 상대적으로 적은 메모리를 소모하고 LDC 해상도를 자유롭게 바꿀 수 있기 때문에, 대상 물체를 복원하는 품질을 조절 가능하며, 주어진 메모리 자원에 대해서 최대의 품질을 가진 결과를 얻을 수 있다.

• 한국문헌정보학회지
• /
• 제44권4호
• /
• pp.115-130
• /
• 2010

In the archival field, the last decade has witnessed much discussion on archives' broad responsibilities for social memory. Considering that the social role of archives has stemmed from postmodern thinking suggests a paradigm shift from viewing archives as static recorded objects to viewing them as dynamic evidence of human memory. The modern archives and archivists are products of nineteenth-century positivism, limiting their function to archiving written documents within stable organizations. The new thoughts on the social role of archives provide a chance to realize that traditional archival practices have preserved only a sliver of organizational memory, thus ignoring fluid records of human activities and memory. Archival description is the primary method for users to access materials in archives. Thus, it can determine how archival materials will be used (or not used). The traditional archival description works as the representation of archival materials and is directly projected from the hierarchy of organizational documents. This paper argues that archivists will need to redefine archival description to be more sensitive to atypical types of archival materials from various cultural contexts. This paper surveys the postmodern approaches to archival concepts in relation to descriptive practices. It also examines some issues related to representing historically marginalized groups in archival description who were previously neglected in traditional archival practices.

• 기록학연구
• /
• 제69호
• /
• pp.153-187
• /
• 2021

인간 삶은 '서사'로 재현되기도 하고, 삶의 과정에서 인간 스스로 '서사'를 만들기도 한다. 따라서 '서사'는 삶과 시간을 이해하는 중요한 수단이다. 과거, 현재, 미래는 단절된 시간들이 아니다. 과거는 미래에 개방되어 있고, 미래는 과거의 의미를 규정하며, 과거는 현재의 삶의 지평으로 환원된다. '서사'는 이러한 열린 시간성을 전제로 형성된다. 기록은 그 자체로서 하나의 '서사'이며, 이 '서사'에 의해 과거의 시간이 재현된다. 또한 기록을 읽어내는 행위에 의해 기록의 '서사'가 아닌 새로운 '서사'가 만들어지기도 한다. 실재와 기록된 '서사'는 순환적이고 변증법적인 구조로 서로를 보완한다. 이글은 실재와 기록, 그리고 기억이 맺는 관계에 대해 현상학적 방법을 동원해 탐구한다. 탐구 도구로 폴 리쾨르의 해석학을 채택하여, 기록이 어떻게 실재와 관계 지워지고 해석되는지, 그리고 아카이브 정의에 복무하는 기억이 어떤 방식으로 형성되어야 하는지에 대해 살펴본다.

• Wind and Structures
• /
• 제9권3호
• /
• pp.231-243
• /
• 2006

An improvement to the spectral representation algorithm for the simulation of wind velocity fields on large scale structures is proposed in this paper. The method proposed by Deodatis (1996) serves as the basis of the improved algorithm. Firstly, an interpolation approximation is introduced to simplify the computation of the lower triangular matrix with the Cholesky decomposition of the cross-spectral density (CSD) matrix, since each element of the triangular matrix varies continuously with the wind spectra frequency. Fast Fourier Transform (FFT) technique is used to further enhance the efficiency of computation. Secondly, as an alternative spectral representation, the vectors of the triangular matrix in the Deodatis formula are replaced using an appropriate number of eigenvectors with the spectral decomposition of the CSD matrix. Lastly, a turbulent wind velocity field through a vertical plane on a long-span bridge (span-wise) is simulated to illustrate the proposed schemes. It is noted that the proposed schemes require less computer memory and are more efficiently simulated than that obtained using the existing traditional method. Furthermore, the reliability of the interpolation approximation in the simulation of wind velocity field is confirmed.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1989년도 하계종합학술대회 논문집
• /
• pp.233-236
• /
• 1989

This paper deals with the expert system using network configuration and input information composed of protective relays and tripped circuit breakers. This system has knowlegebase independent on network dimension because network representation consists of the type of the matrix. Therefore, the knowlege of network representation is simplified, the space of knowlege is reduced, the addition of facts to the knowlege is easy and the expansion of facts is possible. In this paper, the network representation is defined to system matrix. This expert system based on the system matrix diagnoses normal, abnormal operations of protective devices as well as possible fault sections. The brach and bound search technique is used: breadth first technique mixed with depth first technique of primitive PROLOG search technique. This system will be used for real time operations. This expert system obtaines the solution using the pattern matching in working memory without no listing approach for rule control. This paper is written in PROLOG, the A.I. language.

• 한국컴퓨터정보학회논문지
• /
• 제24권2호
• /
• pp.1-7
• /
• 2019

This paper presents CTP (Compact Translation Page), a compact representation of translation pages, for page mapping-based flash translation layers to improve RAM utilization and reduce the response time of solid state drives. CTP can store translation information twice in a translation page and the total number of translation pages stored in flash is reduced to half. Therefore, CTP halves the RAM size of the directory of translation pages and uses the saved RAM space for translation cache. CTP shows the best response time when compared to existing page mapping-based flash translation layers.