• Annual Conference on Human and Language Technology
• /
• 2018.10a
• /
• pp.532-536
• /
• 2018

심층인공신경망을 이용한 대화 모델링 연구가 활발하게 진행되고 있다. 본 논문에서는 대화에서 발화의 감정과 화행을 분류하기 위해 멀티태스크(multitask) 학습을 이용한 End-to-End 시스템을 제안한다. 우리는 감정과 화행을 동시에 분류하는 시스템을 개발하기 위해 멀티태스크 학습을 수행한다. 또한 불균형 범주 분류를 위해 계단식분류(cascaded classification) 구조를 사용하였다. 일상대화 데이터셋을 사용하여 실험을 수행하였고 macro average precision으로 성능을 측정하여 감정 분류 60.43%, 화행 분류 74.29%를 각각 달성하였다. 이는 baseline 모델 대비 각각 29.00%, 1.54% 향상된 성능이다. 본 논문에서는 제안하는 구조를 이용하여, 발화의 감정 및 화행 분류가 End-to-End 방식으로 모델링 가능함을 보였다. 그리고, 두 분류 문제를 하나의 구조로 적절히 학습하기 위한 방법과 분류 문제에서의 범주 불균형 문제를 해결하기 위한 분류 방법을 제시하였다.

• The Transactions of the Korea Information Processing Society
• /
• v.6 no.2
• /
• pp.261-274
• /
• 1999

Recently, automated business processing systems, so called workflow systems, have been studied. have been studied. even though each task has noticible duration inherently, most of previous models regard it as a simple finite automation where state transitions occur instantly. In this paper, we propose a modeling method for workflow each of whose component task has minimum processing time and due-date. Firstly we provide a priority function to choose the proper alternatives when tasks fail before their due-date. The priority function considers the time-relevant factor and the semantic compatibility level to decide the best alternative. secondly, we identify six different types of execution requirements for tasks with noticible durations. They are serial, parallel, simultaneous start simultaneous termination, interleaved and nested dependency. Note that previous models deal with only serial execution and parallel execution. Lastly, we also propose graph-based methods to detect logical conflicts in time constraint-based workflow specifications.

• The KIPS Transactions:PartD
• /
• v.12D no.7 s.103
• /
• pp.1049-1064
• /
• 2005

The demands of increasingly complicated software have led to the proliferation of object-oriented design methodologies in embedded systems. To execute a system designed with objects in target hardware, a task set should be derived from the objects, representing how many tasks reside in the system and which task processes which event arriving at an object. The derived task set greatly influences the responsiveness of the system. Nevertheless, it is very difficult to derive an optimal task set due to the discrepancy between objects and tasks. Therefore, the common method currently used by developers is to repetitively try various task sets. This paper proposes Scenario-based Implementation Synthesis Architecture (SISA) to solve this problem. SISA encompasses a method for deriving a task set from a system designed with objects as well as its supporting development tools and run-time system architecture. A system designed with SISA not only consists of the smallest possible number of tasks, but also guarantees that the response time for each event in the system is minimized. We have fully implemented SISA by extending the ResoRT development tool and applied it to an existing industrial PBX system. The experimental results show that maximum response times were reduced $30.3\%$ on average compared to when the task set was derived by the best known existing methods.

• Proceedings of the Korea Multimedia Society Conference
• /
• 1998.10a
• /
• pp.9-14
• /
• 1998

멀티미디어를 실시간으로 서비스하는 비디오 서버의 설계 시, 네크워크 가용공간, 버퍼 공간 및 디스크 전송률 등이 고려되어야 하며, 특히 버퍼 공간과 디스크 전송률을 동시게 고려하는 기법이 제시되었지만, VOD 서버에서는 보통 MPEG데이터가 VBR 형식으로 되어 있기 때문에 CBR로 모델링하면 시스템의 가용성이 떨어지게 된다. 또한 수용제어를 위한 버퍼 예약기법에서는 VBR 형태로 데이터를 모델링하는 기법들이 제시되었지만, 이 경우에도 디스크 전송률을 계산할 때는 CBR형식으로 가정하거나 아예 고려를 하지 않기 때문에 역시 시스템 자원을 효율적으로 사용할 수 없다. 본 논문에서는 VBR 비디오 서버의 설계를 위하여 서버에서 처리되는 각 스트림들을 하드 실시간 태스크 모델로 모델링하고 실시간 태스크들을 효율적으로 처리할 수 있는 실시간 스케쥴링 알고리즘을 제안한다. 이 실시간 스케쥴링 알고리즘을 디스크 액세스의 승인제어, 디스크 스케쥴링, 버퍼 관리에 이용하므로서 연속성을 보장할 수 있는 효율적인 VBR 연속 미디어 서버를 설계할 수 있다. 비디오 서버의 성능은 시뮬레이션을 이용하여 분석하였다.

• Journal of KIISE:Computer Systems and Theory
• /
• v.27 no.6
• /
• pp.549-557
• /
• 2000

Distributed Memory Machine(DMM) is necessary for the effective computation of the data which is complicated and very large. Task scheduling is a method that reduces the communication time among tasks to reduce the total execution time of application program and is very important for the improvement of DMM. Task Duplicated based Scheduling(TDS) method improves execution time by reducing communication time of tasks. It uses clustering method which schedules tasks of the large communication time on the same processor. But there is a problem that cannot optimize communication time between task sending data and task receiving data. Hence, this paper proposes a new method which solves the above problem in TDS. Modified Task Duplicated based Scheduling(MTDS) method which can approximately optimize the communication time between task sending data and task receiving data by checking the optimal condition, resulted in the minimization of task execution time by reducing the communication time among tasks. Also system modeling shows that task execution time of MTDS is about 70% faster than that of TDS in the best case and the same as the result of TDS in the worst case. It proves that MTDS method is better than TDS method.

• Proceedings of the Korean Information Science Society Conference
• /
• 2000.10c
• /
• pp.108-110
• /
• 2000

무선 네트웍 환경에서의 이동컴퓨팅에서는 특성상 잦은 끊김과 높은 에러율 때문에 비연결시에도 수행을 계속할 수 있는 기능이 필요하다. 이러한 이동컴퓨팅 환경에서 비연결 수행기능을 제공하기 위해 많은 개념과 이론들이 제안되고 있다. 본 논문에서 마코프(Marcov) 모델링 기법을 이용하여 모빌 컴퓨팅 환경에서 비연결 수행상태를 포함한 평균 태스크의 수행시간을 분석하였다. 모빌 컴퓨팅 환경은 데이터 호딩(Data Hoarding), 비연결 수행(Disconnected operation), 정지(Stop)의 3가지 상태로 구성할 수 있다. 이러한 3가지 상태에서 여러 가지 입력 파라메터들(에러율 Error rate), 재연결율(Recovery rate), 태스크 수행 정지 확률(Stop), 호딩 오버헤드(Hoarding overhead), 재연결 오버헤드(Reintegration overhead)들이 태스크 수행시간에 미치는 영향을 분석하였다. 이러한 분석을 통해서 통신망 단절을 고려한 이동컴퓨팅에서 보다 효과적인 태스크 수행기법을 선택할 수 있다.

• The KIPS Transactions:PartD
• /
• v.12D no.7 s.103
• /
• pp.957-964
• /
• 2005

According to rapid growth of web environment, user interface design needs to support the complex interactions between human and computer. In the paper we suggest the object modeling method for Qualify Improvement of User Interface. We propose the 4 business event's object modeling phases such as business event object modeling, task object modeling, transaction object modeling, and form object modeling to enhance visual cohesion of UI. As a result, this 4 phases in this paper allows us to enhance visual cohesion of User Interface prototype. We have found that the visual cohesion of business events become strong and unskilled designer can develope the qualified user interface prototype. And it also improves understanding of business task and reduces prototype system development iteration.

• Journal of KIISE:Computer Systems and Theory
• /
• v.27 no.5
• /
• pp.506-517
• /
• 2000

In distributed real-time system, a task activated by the completion of its preceding task can be modeled as a periodic task with activation jitter. An activation jitter of a task is defined asthe difference between the worst case and the best case response time of its preceding task. Becausethe existing approaches assume that the best case response time is much smaller than the actual one,the activation jitter and the worst case response time of lower priority tasks are overestimated. Thispaper proposes a new analysis technique to calculate the best case response time more precisely andto reduce the activation jitter bounds. The proposed technique obtains the best case response time byconsidering the relative phase between tasks. The precise analysis of the activation jitters can reducethe worst case response time of other tasks and increase the schedulability. The simulation resultsshow that the proposed analysis technique improves the accuracy of the best case and the worst caseresponse time up to 40% and 6%, respectively.

• The Transactions of the Korea Information Processing Society
• /
• v.5 no.3
• /
• pp.662-670
• /
• 1998

In order to overcome the problems with first generation expert systems at the symbol level, recently various knowledge level development techniques of second generation expert systems have been proposed. But, these techniques are conceptual modelling techniques. This paper modifies and complements these conceptual modelling techniques and proposes a Task Object Modelling (TOM) technique as a practical knowledge level expert system development technique. This paper defines a Task Object(TO) as a knowledge unit consisted of a goal, execution conditions, behaviour knowledge, and so on. And, we define a Task Object Diagram(TOD) to depict structual, dynamic, and functional aspects of TO easily. We also define Inference Types as basic units to describe behaviour knowledge of TOs. In order to utilize the proposed TOM technique as not a simple conceptual modelling technique but a practical second generalion expert system development technique, we implement a TOD editor, a TO editor, and TO processing algorithm based on the state of TOs. Also we implement a Inference Types Library, in which each inference type is corresponded to an IRE(Jntelligent Rules Element) method, to transform the defined inference types into IRE methods automatically.

• Proceedings of the Korean Information Science Society Conference
• /
• 2001.10c
• /
• pp.853-855
• /
• 2001

네트워크 기술의 발전으로 저비용으로 고성능을 얻고자 하는 클러스터 시스템에 대한 연구가 많아지고 있다. SPMD(Single Program Multiple Data) 형태의 병렬 프로그램을 사용한 클러스터 시스템의 주된 성능 장애는 부하 불균등 현상이다 본 논문에서는 이러한 문제를 해결하기 위해 마스터 노드가 정보를 모으는 횟수와 주기를 시뮬레이션을 통해 최적의 값으로 결정하고 그 주기 동안에 각 노드의 태스크 당 평균 수행시간을 계산한다. 통신비용의 오버헤드를 고려한 시스템의 실행시간을 평균 수행시간으로 예측하여 각 노드가 이동할 태스크의 수를 결정하는 동적 부하 균등 알고리즘을 제안한다 제안한 알고리즘의 클러스터 시스템을 모델링하고 성능 분석을 위한 시뮬레이션을 한다.