• Annual Conference on Human and Language Technology
• /
• 2018.10a
• /
• pp.605-607
• /
• 2018

본 논문에서는 음절 임베딩과 양방향 LSTM-CRF 모델을 이용한 한국어 문장 자동 띄어쓰기 시스템을 제안한다. 문장에 대한 자질 벡터 표현을 위해 문장을 구성하는 음절을 Unigram 및 Bigram으로 나누어 각 음절을 연속적인 벡터 공간에 표현하고, 양방향 LSTM을 이용하여 현재 자질에 양방향 자질들과 의존성을 부여한 새로운 자질 벡터를 생성한다. 이 새로운 자질 벡터는 전방향 신경망과 선형체인(Linear-Chain) CRF를 이용하여 최적의 띄어쓰기 태그 열을 예측하고, 생성된 띄어쓰기 태그를 기반으로 문장 자동 띄어쓰기를 수행하였다. 문장 13,500개와 277,718개 어절로 이루어진 학습 데이터 집합과 문장 1,500개와 31,107개 어절로 이루어진 테스트 집합의 학습 및 평가 결과는 97.337%의 음절 띄어쓰기 태그 분류 정확도를 보였다.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.4
• /
• pp.115-124
• /
• 2021

Since function block diagram (FBD) programs are widely used to implement safety-critical systems, effective testing for FBD programs has become important. Mutation testing, a fault-based testing, is highly effective in fault detection but computationally expensive. To support testers for FBD programs, we propose an automated mutant generator for FBD programs. We designed the MuGenFBD tool with the cost and equivalent mutant issues in consideration. We conducted experiments on real industrial examples to present the performance of MuGenFBD. The results show that MuGenFBD can generate mutants for FBD programs automatically with low probability of equivalent mutants and low cost. This tool can effectively support mutation analysis and mutation-adequate test generation for FBD programs.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.12 no.6
• /
• pp.531-536
• /
• 2002

In this paper, we have applied Genetic Algorithms(GAs) to Intrusion Detection System(TDS), and then proposed and simulated the misuse detection model firstly. We have implemented with the KBD contest data, and tried to simulated in the same environment. In the experiment, the set of record is regarded as a chromosome, and GAs are used to produce the intrusion patterns. That is, the intrusion rules are generated. We have concentrated on the simulation and analysis of classification among the Data Mining techniques and then the intrusion patterns are produced. The generated rules are represented by intrusion data and classified between abnormal and normal users. The different rules are generated separately from three models "Time Based Traffic Model", "Host Based Traffic Model", and "Content Model". The proposed system has generated the update and adaptive rules automatically and continuously on the misuse detection method which is difficult to update the rule generation. The generated rules are experimented on 430M test data and almost 94.3% of detection rate is shown.3% of detection rate is shown.

• Proceedings of the Korean Information Science Society Conference
• /
• 2007.10b
• /
• pp.232-237
• /
• 2007

컴퓨터 시스템의 설계는 알고리즘 수준의 모델링에서부터 시제품 수준까지 시스템을 구체화해 나가는 일련의 과정이다. 시스템 구현의 구체화 과정에는 단순하고 반복적인 구현이 많이 포함되며, 이 과정에서 많은 오류가 발생한다. 이러한 오류는 개발자가 알고리즘 수준에서는 드러나지 않는 복잡하고 아키텍처 의존적인 하드웨어-소프트웨어 동기화 메커니즘의 개발과 같은 시스템 구현의 구체화 과정을 모두 떠안고 있기 때문에 발생하는 것이다. 이 논문에서는, 이러한 문제를 극복하기 위하여, 알고리즘을 데이터 플로우로 모델링하면 이로부터 합성 가능한 하드웨어 플랫폼과 동기화 로직, 그리고 동기화를 위한 드라이버 소프트웨어 일제를 자동 생성하는 설계 과정을 제시하고자 한다. 제시된 설계 과정은 자체 개발한 통합 설계 도구 상에 구현되었으며, 이를 통해서 개발된 H.263 디코더 예제를 상용의 RTL 통합 시뮬레이션 도구인 Seamless CVE와, SoC 프로토타이핑 환경인 Altera Excalibur 시스템 상에서 테스트하여 그 완성도를 검증하였다.

• KIPS Transactions on Software and Data Engineering
• /
• v.9 no.8
• /
• pp.229-234
• /
• 2020

This paper presents the design and the implementation of CRESTIVE-DX, a concolic testing tool that distribute the concolic testing process over the embedded target system and the host system for efficient test generation of a target embedded program. CRESTIVE-DX conducts the execution of a target program on the target embedded system to consider possible machine-dependent behaviors of a target program execution, and conducts machine-independent parts, such as search-strategy heuristics, constraint solving, on host systems with high-speed computation unit, and coordinates their concurrent executions. CRESTIVE-DX is implemented by extending an existing concolic testing tool for C programs CREST. We conducted experiments with a test bed that consists of an embedded target system in the Arm Cortex A54 architecture and host systems in the x86-64 architecture. The results of experiments with Unix utility programs Grep, Busybox Awk, and Busybox Ed show that test input generation of CRESTIVE-DX is 1.59 to 2.64 times faster than that of CREST.

• KIPS Transactions on Software and Data Engineering
• /
• v.5 no.4
• /
• pp.165-170
• /
• 2016

Using the intent of the Android application, an application can execute other application's components. However, if interaction between these components are not processed normally, such problems as incorrect component execution and unhandled system broadcast may be occurred. In this paper, to generate test scenarios for inter application interaction, a testing approach is proposed using a merged intent list and a single merged diagram. The proposed method can effectively be carried out to check the abnormal interaction among the applications which was not considered in existing testing approaches.

• KIPS Transactions on Software and Data Engineering
• /
• v.2 no.10
• /
• pp.679-690
• /
• 2013

Smartphones have features that users feel free to install/delete the program they want. Their emergence makes many developers rush into the Smartphone application development market. Thus, developing good applications quickly is becoming even more intense competition in the market. Because, however, the application development and deployment procedures are simple in the Android environments and anyone can participate in the development easily, applications not validated thoroughly are likely to be deployed. Therefore, a systematic approach that can verify Android-based applications with fewer burdens is required. In this paper, we propose a method that generates automatically GUI-based testing scenarios for the Android applications. The automated test scenario generation can reduce the time which the developer spends on testing, thus it can improve the productivity of the development in the testing phase.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2023.07a
• /
• pp.541-544
• /
• 2023

LLM(Largescale Language Model)의 성능 향상을 위한 비용 효율적인 방법으로 ChatGPT, GPT-4와 같은 초거대 모델의 output에 대해 SLM(Small Language Model)을 finetune하는 방법이 주목받고 있다. 그러나, 이러한 접근법은 주로 범용적인 지시사항 모델을 위한 학습 방법으로 사용되며, 제한된 특정 도메인에서는 추가적인 성능 개선의 여지가 있다. 본 연구는 특정 도메인(Writing Assistant)에서의 성능 향상을 위한 새로운 방법인 Self-Guided Approach를 제안한다. Self-Guided Approach는 (1) LLM을 활용해 시드 데이터에 대해 도메인 특화된 metric(유용성, 관련성, 정확성, 세부사항의 수준별) 점수를 매기고, (2) 점수가 매겨진 데이터와 점수가 매겨지지 않은 데이터를 모두 활용하여 supervised 방식으로 SLM을 미세 조정한다. Vicuna에서 제안된 평가 방법인, GPT-4를 활용한 자동평가 프레임워크를 사용하여 Self-Guided Approach로 학습된 SLM의 성능을 평가하였다. 평가 결과 Self-Guided Approach가 Self-instruct, alpaca와 같이, 생성된 instruction 데이터에 튜닝하는 기존의 훈련 방법에 비해 성능이 향상됨을 확인했다. 다양한 스케일의 한국어 오픈 소스 LLM(Polyglot1.3B, PolyGlot3.8B, PolyGlot5.8B)에 대해서 Self-Guided Approach를 활용한 성능 개선을 확인했다. 평가는 GPT-4를 활용한 자동 평가를 진행했으며, Korean Novel Generation 도메인의 경우, 테스트 셋에서 4.547점에서 6.286점의 성능 향상이 발생했으며, Korean scenario Genration 도메인의 경우, 테스트 셋에서 4.038점에서 5.795 점의 성능 향상이 발생했으며, 다른 유사 도메인들에서도 비슷한 점수 향상을 확인했다. Self-Guided Approach의 활용을 통해 특정 도메인(Writing Assistant)에서의 SLM의 성능 개선 가능성을 확인했으며 이는 LLM에 비용부담을 크게 줄이면서도 제한된 도메인에서 성능을 유지하며, LLM을 활용한 응용 서비스에 있어 실질적인 도움을 제공할 수 있을 것으로 기대된다.

• KIPS Transactions on Software and Data Engineering
• /
• v.3 no.12
• /
• pp.537-546
• /
• 2014

The purpose of this study is to develope an algorithm which generates trip plans based on rating points of travel app users and travel experts to help potential travellers experience their desired destinations in advance. This algorithm uses the above rating points and the gradually created hierarchy to generate the most preferred and efficient trip courses. Users can go through video clips or panoramic VR videos of the actual destinations from their trip plans generated by the algorithm which may add excitement to their actual trips. With our heuristic methods, the more users input their ratings, the better trip plans can be generated. This algorithm has been tested on android OS and proven efficient in generating trip plans. This research introduces a way to experience travel destinations with panoramic VR video and proposes the algorithm which generates trip plans based on users' ratings. It is expected to be useful for travellers' trip planning and to contribute growth in the travel market.

• Journal of KIISE
• /
• v.43 no.4
• /
• pp.480-488
• /
• 2016

Search-Based Software Testing (SBST) is an effective technique for test data generation on large domain size. Although the performance of SBST seems to be affected by the structural characteristics of Software Under Test (SUT), studies for the comparison of SBST techniques considering structural characteristics are rare. In addition to the comparison study for SBST, we analyzed the best algorithm with different structural characteristics of SUT. For the generalization of experimental results, we automatically generated 19,800 SUTs by combining four metrics, which are expected to affect the performance of SBST. According to the experiment results, Genetic algorithm showed the best performance for SUTs with high complexity and test data evaluation with count ${\leq}20,000$. On the other hand, the genetic simulated annealing and the simulated annealing showed relatively better performance for SUTs with high complexity and test data evaluation with count ${\geq}50,000$. Genetic simulated annealing, simulated annealing and hill climbing showed better performance for SUTs with low complexity.