• Annual Conference on Human and Language Technology
• /
• 2022.10a
• /
• pp.205-209
• /
• 2022

페르소나 대화 시스템이 상대방의 개인화된 정보에 일관된 응답을 생성하는 것은 상당히 중요하며, 이를 해결하기 위해 최근에 많은 연구들이 활발히 이루어지고 있다. 그 중, PersonaChat 데이터셋에 대해 수반/중립/모순 관계를 라벨링한 DialoguNLI 데이터셋이 제안되었으며, 일관성 측정, 페르소나 속성 추론 태스크 등 여러 분야에 활용되고 있다. 그러나, 공개적으로 이용가능한 한국어로 된 대화 추론 데이터셋은 없다. 본 연구에서는 한국어로 번역된 페르소나 대화 데이터셋과 한국어 자연어 추론 데이터셋에 학습된 모델을 이용하여 한국어 대화 추론 데이터셋(KorDialogueNLI)를 구축한다. 또한, 사전학습된 언어모델을 학습하여 한국어 대화 추론 모델 베이스라인도 구축한다. 실험을 통해 정확도 및 F1 점수 평가 지표에서 KLUE-RoBERTa 모델을 미세조정(fine-tuning)시킨 모델이 가장 높은 성능을 달성하였다. 코드 및 데이터셋은 https://github.com/passing2961/KorDialogueNLI에 공개한다.

• Annual Conference on Human and Language Technology
• /
• 2022.10a
• /
• pp.613-618
• /
• 2022

효과적인 자연어 처리를 위해 제안된 Transformer 구조의 등장 이후, 이를 활용한 대규모 언어 모델이자 사전학습 모델인 BERT, GPT, OPT 등이 공개되었고, 이들을 한국어에 보다 특화한 KoBERT, KoGPT 등의 사전학습 모델이 공개되었다. 자연어 처리 모델의 확보를 위한 학습 자원이 늘어나고 있지만, 사전학습 모델을 각종 응용작업에 적용하기 위해서는 데이터 준비, 코드 작성, 파인 튜닝 및 저장과 같은 복잡한 절차를 수행해야 하며, 이는 다수의 응용 사용자에게 여전히 도전적인 과정으로, 올바른 결과를 도출하는 것은 쉽지 않다. 이러한 어려움을 완화시키고, 다양한 기계 학습 모델을 사용자 데이터에 보다 쉽게 적용할 수 있도록 AutoML으로 통칭되는 자동 하이퍼파라미터 탐색, 모델 구조 탐색 등의 기법이 고안되고 있다. 본 연구에서는 한국어 사전학습 모델과 한국어 텍스트 데이터를 사용한 자연어 처리 모델 산출 과정을 정형화 및 절차화하여, 궁극적으로 목표로 하는 예측 모델을 자동으로 산출하는 시스템의 설계를 소개한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2009.04a
• /
• pp.313-316
• /
• 2009

관계형 데이터베이스를 사용하는 응용시스템들은 어플리케이션 소스 코드에 내재된 SQL을 통하여 데이터를 액세스하게 되며, SQL을 통한 데이터베이스 액세스 성능은 응용시스템의 성능에 결정적인 영향을 미치기 때문에, 관계형 데이터베이스시스템의 튜닝에 대해서는 많은 연구가 이루어져 왔다. 그러나 데이터베이스시스템 튜닝을 통해 얻은 성능 수준은 응용시스템의 변경이나 데이터 저장량 증가 등 여러 요인에 의해 다시 저하될 수 있으며, 이와 같은 성능 수준의 변화를 정량적으로 관리하면서 최소 비용으로 성능 수준을 유지할 수 있도록 튜닝 시점을 결정하는 방법에 대한 연구는 현재까지 연구된 바가 없다. 이 논문에서 제시하는 DB 성능지수와 이를 이용한 데이터베이스시스템의 성능관리 모델은 특정 벤더에 종속되지 않고 성능수준을 정량화함으로써 이해관계자들에게 성능 수준의 변화를 가시적으로 보이고 데이터베이스 관리자의 성능관리 및 튜닝 시점 판단을 지원하여 일정 수준의 성능을 유지할 수 있는 방법으로 활용될 수 있다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.10a
• /
• pp.333-336
• /
• 2022

With the development of IT technology, cybercrime is becoming sophisticated and intelligent. In particular, in the case of BackDoor, which is used in the APT attack (intelligent continuous attack), it is very important to detect malicious behavior and respond to infringement because it is often unaware that it has been damaged by an attacker. This paper aims to build an EDR platform that can monitor, analyze, and respond to malicious behavior in real time by collecting, storing, analyzing, and visualizing logs in an endpoint environment in real time using open source-based analysis solutions ELK Stack and Sysmon.

• Transactions on Semiconductor Engineering
• /
• v.1 no.1
• /
• pp.1-8
• /
• 2023

This paper presents the design and implementation of Crystals-Kyber, a next-generation postquantum cryptography, as a hardware accelerator on an FPGA using High-Level Synthesis (HLS). We optimized the Crystals-Kyber algorithm using various directives provided by Vitis HLS, configured the AXI interface, and designed a hardware accelerator that can be implemented on an FPGA. Then, we used Vivado tool to design the IP block and implement it on the ZYNQ ZCU106 FPGA. Finally, the video was recorded and H.264 compressed with Python code in the PYNQ framework, and the video encryption and decryption were accelerated using Crystals-Kyber hardware accelerator implemented on the FPGA.

• Communications of Mathematical Education
• /
• v.37 no.3
• /
• pp.467-482
• /
• 2023

In this paper, we present our new teaching and learning materials on gradient descent method, which is widely used in artificial intelligence, available for college mathematics. These materials provide a good explanation of gradient descent method at the level of college calculus, and the presented SageMath code can help students to solve minimization problems easily. And we introduce how to solve least squares problem using gradient descent method. This study can be helpful to instructors who teach various college-level mathematics subjects such as calculus, engineering mathematics, numerical analysis, and applied mathematics.

• The Journal of the Korea Contents Association
• /
• v.12 no.8
• /
• pp.355-365
• /
• 2012

The purpose of this research was to review the current situation of Ssireum and to investigate a basic direction for globalization of Ssireum. Examining disposition, situation, and image of Ssireum through literatures analysis, survey, in-depth interview was conducted. In detail, it's recognition and image perceived by customers and general individuals were investigated, and it's perception and related image of foreigners were analyzed. We examined the situation and the various problems of Ssireum on the base of site and customers. We made a conclusion that Korea has to make a great deal of effort for globalization of Ssireum both as a sport through institutionalization of alternatives for the globalization and as a heritage which has a strong identity. As detail alternatives for the globalization of Ssireum, education through physical education in school, bringing up foreign athletes, maximization of event values, specialization of Ssireum organization, international instructor, glocalization, culture coding, enhancement of ritual, connection with festival and so on were suggested.

• KIPS Transactions on Computer and Communication Systems
• /
• v.9 no.11
• /
• pp.247-249
• /
• 2020

In the era of the 4th industrial revolution, the demand for convergence between ICT technologies is increasing in various fields. Accordingly, a new term that combines data, network, and artificial intelligence technology, DNA (Data, Network, AI) is in use. and has recently become a hot topic. DNA has various potential technology to be able to develop intelligent application in the real world. Therefore, this paper introduces the reviewed papers on the service image placement mechanism based on the logical fog network, the mobility support scheme based on machine learning for Industrial wireless sensor network, the prediction of the following BCI performance by means of spectral EEG characteristics, the warning classification method based on artificial neural network using topics of source code and natural language processing model for data visualization interaction with chatbot, related on DNA technology.

• Journal of The Korean Association of Information Education
• /
• v.15 no.3
• /
• pp.375-385
• /
• 2011

Computer education is moving its focus from skill oriented education to improving students' creativity and problem solving ability. Thus, the importance of programming education is being strengthened. However, programming education was biased to grammar oriented language that has been limits of students' interest. Robot programming is problem solving itself, and by allowing students to directly see the robot which is the output of programming, can help interest and motivate to the students. In fact, it is still observed that the students are facing difficulties due to various kinds of errors during the programming education. Therefore, this study categorizes and analyzes the errors students are facing during robot programming, and based on that, a support tool to help treat errors developed. The developed supporting system for error solving reduces the frequency of errors and provides the set of coding instruction, NXC language and error message in Korean, examples and detailed information for each stage of education, function removing major coding errors, and code sorting and indication of row number. This study also confirmed that the supporting tool is helpful in reducing and solving errors after input.

• The Journal of the Acoustical Society of Korea
• /
• v.18 no.7
• /
• pp.33-38
• /
• 1999

In this paper, we combine a generalized analysis-by-synthesis (AbS) structure and an algebraic excitation scheme to propose a new 4kb/s speech codec. This codec partly uses the structure of G.729. We design a line spectrum pair (LSP) quantizer, an adaptive codebook, and an excitation codebook to fit the 4 kb/s bit rate. The codec has a 25㎳ algorithmic delay, which corresponds to a 20㎳ frame size and a 5㎳ lookahead. At the bit rates below 4kb/s, most CELP speech codecs using the AbS principle have a drawback that results a rapid degradation of speech quality. To overcome this drawback we use the generalized AbS structure which is efficient for the low bit rate speech codec. LP coefficients are converted to LSP and quantized using a predictive 2-stage VQ. A low complexity algebraic codebook which uses shifting method is used for the fixed codebook excitation, and gains of the adaptive codebook and the fixed codebook are quantized using the VQ. To evaluate the performance of the proposed codec A-B preference tests are done with the fixed rate 8kb/s QCELP. As the result of the test, the performance of the codec is similar to that of the fixed rate 8kb/s QCELP.