• The Bulletin of The Korean Astronomical Society
• /
• v.32 no.2
• /
• pp.70.2-70.2
• /
• 2007

• Proceedings of the Korea Air Pollution Research Association Conference
• /
• 2001.11a
• /
• pp.441-442
• /
• 2001

TOMS (Total Ozone Mapping Spectrometer)는 Nimbus-7과 Meteor-3에 탑재되어 자외선 영역에서 오존의 양을 측정하는 임무를 가지고 1978년부터 1994년 12월까지 일별 오존자료를 제공하였다. 1996년 8월 17일에 발사된 일본의 ADEOS에 탑재된 TOMS는 1997년 6월 29일까지 자료를 생산했으며 1996년7월 2일 발사된 Earth Probe의 TOMS는 현재까지 실시간 자료를 보내오고 있다. TOMS는 지구의 대기에 의해 반사된 태양 광선중의 자외선 자료를 통해 오존의 양을 간접적으로 측정하는 것이 원리이다. (중략)

• Bulletin of the Korean Space Science Society
• /
• 2007.04a
• /
• pp.63-63
• /
• 2007

• Journal of Korea Multimedia Society
• /
• v.22 no.11
• /
• pp.1223-1232
• /
• 2019

The injection is the input method of the image feature vector from the encoder to the decoder. Since the image feature vector contains object details such as color and texture, it is essential to generate image captions. However, the bidirectional decoder model using the existing injection method only inputs the image feature vector in the first step, so image feature vectors of the backward sequence are vanishing. This problem makes it difficult to describe the context in detail. Therefore, in this paper, we propose the parallel injection method to improve the description performance of image captions. The proposed Injection method fuses all embeddings and image vectors to preserve the context. Also, We optimize our image caption model with Bidirectional Gated Recurrent Unit (Bi-GRU) to reduce the amount of computation of the decoder. To validate the proposed model, experiments were conducted with a certified image caption dataset, demonstrating excellence in comparison with the latest models using BLEU and METEOR scores. The proposed model improved the BLEU score up to 20.2 points and the METEOR score up to 3.65 points compared to the existing caption model.

• Journal of Advanced Navigation Technology
• /
• v.21 no.6
• /
• pp.674-679
• /
• 2017

As the importance of the 4th industrial revolution and ICT technology increased, it became a software centered society. Existing software training was limited to the composition of the learning environment, and a lot of costs were incurred early. In order to solve these problems, a learning method using a web compiler was developed. The web compiler supports various software languages and shows compilation results to the user via the web. However, Web compilers that support artificial intelligence technology are missing. In this paper, we designed and implemented a tensor flow based web compiler, Google's artificial intelligence library. We implemented a system for learning artificial intelligence by building a meteorJS based web server, implementing tensor flow and tensor flow serving, Python Jupyter on a nodeJS based server. It is expected that it can be utilized as a tool for learning artificial intelligence in software centered society.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.5 no.5
• /
• pp.892-901
• /
• 2001

Despite of many advantages over conventional radio paths, application of the Meteor Burst Path in commercial communication system is so far limited to a few extent because of its low duty rate, which is, less than several percent at best. In order to get through maximum number of data bits during the short interval of each burst, which normally lives a fractions of a second, a sophisticated communication protocol is called for. In this paper, author introduces a communication protocol which can effectively locate and fetch burst signal by sending out periodic detection signal from master station and, with minimal delay, switch over to transmission states so as to put as many data bits through the detected burst as the burst length permits. By analyzing time period required for each transaction steps, the author derives optimal packet length for various bursts which assures to get a message string through in minimum delay. According to the analysis, the author proposes a protocol including a routine which automatically accesses average length of bursts at each environment and automatically adapt length of data packet so as to maximize data throughput under Riven burst conditions.

• Journal of The Korean Astronomical Society
• /
• v.56 no.1
• /
• pp.75-89
• /
• 2023

We investigate the astronomical phenomenon records of Sukjong's Chunbang-Ilgi made by Sigangwon (Royal Educational Office of the Crown Prince) at which King Sukjong was the crown prince (i.e., 1667 January 24-1674 September 22). From the daily records of 2,799 days, we extract the astronomical records of 1,443 days and classify them into 14 categories. Then, we group the records of each category into five phenomena (Atmosphere, Eclipse, Daylight Appearance, Apparition, and Appulse) and compare them with the results of modern astronomical computations wherever possible. Except for Atmosphere group comprising records of meteorological events, such as solar halo, lunar halo, and unusual clouds, the significant findings in every other group are as follows: In Eclipse group, the solar eclipse that occurred on 1673 August 12 was unobservable in Korea, which is in contrast to the record of Joseonwangjo-Sillok (Annals of the Joseon Dynasty), which states that the sun was in eclipse around sunset time, as observed at Nam mountain. From the lunar eclipse records, we verify that the Joseon court did not change the date of the events observed after midnight. In Daylight Appearance group, we confirm that this phenomenon was observed during the daytime and not during twilight. We further suggest that if observation conditions are met, a celestial body brighter than -2.3 mag could be seen during the daytime with the naked-eye. In Apparition group, we find the possibilities that the Orionid meteor shower had influence on the meteor records and the seasonality on the aurora records. We also find that the Korean records in which the coma of comet C/1668 E1 was located below the horizon were overlooked in previous studies. Finally, we find that the records of Appulse group generally agree with the results of modern calculations. The records of Beom (trespass in literal) and Sik (eating in literal) events show average angular separations of 1.2° and 1.0°, respectively. In conclusion, we believe this work helps study the astronomical records of other logs of Sigangwon, such as Sukjong's Chunbang-Ilgi.

• The Bulletin of The Korean Astronomical Society
• /
• v.40 no.2
• /
• pp.60.3-60.3
• /
• 2015

FM 라디오 방송을 이용하여 별똥 개수의 시간 변화를 측정하는 시스템을 만들고 관측을 하였다. 시스템은 간단한 야기 안테나와 FM-라디오 수신기 및 PC 사운드 카드와 분석 소프트웨어 등으로 구성되어 있다. 관측 결과를 해석하고 별똥이 아닌 인위적 또는 자연적 신호를 골라내기 위한 몇 가지 실험을 수행하였다. 첫 4개월 간의 예비관측 결과를 발표한다.

• Publications of The Korean Astronomical Society
• /
• v.30 no.2
• /
• pp.715-718
• /
• 2015

Astronomy is a popular topic for the public in term of astronomical phenomenon such as occultations, solar and lunar eclipses or meteor showers. In term of education, astronomy also is popular as one of the world Science Olympiads. Social media, as the new trend in communicating and connecting people, plays a significant role in increasing the size of the astronomy community. Beyond IYA 2009, more and more astronomy activities have been done in many places in Indonesia. New astronomy communities have been formed in several cities and public engagement is also high in social media especially on Facebook and Twitter. In this paper, we will discuss the lesson learned from astronomy outreach achievements in Indonesia and the need for citizen science projects as a distance learning tool for the public as part of astronomy development in Indonesia. We argue and propose that this project will be also important up to a regional scope.

• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.275-280
• /
• 2023

문서 그라운딩된 대화 시스템의 응답 성능 개선을 위한 방법론을 제안한다. 사전 학습된 거대 언어 모델 LLM(Large Language Model)인 Llama2 모델에 Zero-Shot In-Context learning을 적용하여 대화 마지막 유저 질문에 대한 응답을 생성하는 태스크를 수행하였다. 본 연구에서 제안한 응답 생성은 검색된 top-1 문서와 대화 기록을 참조해 초기 응답을 생성하고, 생성된 초기 응답을 기반으로 검색된 문서를 대상으로 재순위화를 수행한다. 이 후, 특정 순위의 상위 문서들을 이용해 최종 응답을 생성하는 과정으로 이루어진다. 검색된 상위 문서를 이용하는 응답 생성 방식을 Baseline으로 하여 본 연구에서 제안한 방식과 비교하였다. 그 결과, 본 연구에서 제안한 방식이 검색된 결과에 기반한 실험에서 Baseline 보다 F1, Bleu, Rouge, Meteor Score가 향상한 것을 확인 하였다.