A Preliminary Study on Evaluation of Time- Dependent Radionuclide Removal Performance Using Artificial Intelligence for Biological Adsorbents

Janghee Lee,Seungsoo Jang,Min-Jae Lee,Woo-Sung Cho,Joo Yeon Kim,Sangsoo Han,신성균,Sun Young Lee,Dae Hyuk Jang,Miyong Yun,Song Hyun Kim 대한방사선방어학회 2023 방사선방어학회지 Vol.48 No.4

Background: Recently, biological adsorbents have been developed for removing radionuclides from radioactive liquid waste due to their high selectivity, eco-friendliness, and renewability. However, since they can be damaged by radiation in radioactive waste, a method for estimating the bio-adsorbent performance as a time should consider the radiation damages in terms of their renewability. This paper aims to develop a simulation method that applies a deep learning technique to rapidly and accurately estimate the adsorption performance of bio-adsorbents when inserted into liquid radioactive waste. Materials and Methods: A model that describes various interactions between a bio-adsorbent and liquid has been constructed using numerical methods to estimate the adsorption capacity of the bio-adsorbent. To generate datasets for machine learning, Monte Carlo N-Particle (MCNP) simulations were conducted while considering radioactive concentrations in the adsorbent column. Results and Discussion: Compared with the result of the conventional method, the proposed method indicates that the accuracy is in good agreement, within 0.99% and 0.06% for the R2 score and mean absolute percentage error, respectively. Furthermore, the estimation speed is improved by over 30 times. Conclusion: Note that an artificial neural network can rapidly and accurately estimate the survival rate of a bio-adsorbent from radiation ionization compared with the MCNP simulation and can determine if the bio-adsorbents are reusable.

流通近代化에 관한 一硏究

韓樟熙 全南大學校企業經營硏究所 1983 産業經濟硏究 Vol.10 No.1

In this study, the author purposed to develop a "knowledge-system" on the problem of the modernization of macromarketing system. For this purpose, the modernization of macromarketing system was defined as a planned change process to achieve and maintain the normative function mix which can maximize the efficiency of the system, that is, can minimize the total marketing costs, on the functionalists' viewpoint. And the existence of normative function mix was proved on the based of the facts that there are some substitutibilities among marketing functions and these substitutions can change the relationships between total marketing costs and marketing service output levels which determine the efficiency of the system. Next, under the assumption that the amount of marketing objects is consistant, factors which affect the normative function mix itself and its acquisition and maintenance were found out and the relationships among these factors were examined. In addition, on the grounds of these findings, some changes which were thought to be necessary for the modernization of the system were presented. And the ways which permit more effective accomplishment of these changes were investigated by using the Sheth and Frazier's "Extended Model of Strategy Mix Choice for Planned Social Change." Further, because these changes can be introduced only when any change agent(s) appear(s) and because the most effective change agent varies with attempted changes, some consideration on the schemata of change agent classification were made. In this study, the Zif's "Macromarketing Management Classification Schema" was recommended for its classification. But any study on the effects of the environment of macromarketing system on the modernization processes and any study on the processes which permit more rapid and economic completion of changes were not made. Besides, any inductive research for justifying the findings of this study was not carried out in it.

Development of a nondestructive assay method using Raman spectroscopy in the pharmaceutical production process of a freeze-dried injection with gemcitabine as active pharmaceutical ingredient

Jaejin Kim,Janghee Han,Young-Ah Woo 대한화학회 2021 Bulletin of the Korean Chemical Society Vol.42 No.12

Freeze-dried injectable drugs based on gemcitabine are mainly used as anticancer drugs for pancreatic cancer. High-performance liquid chromatography (HPLC) analysis is mainly used to analyze the active pharmaceutical ingredient (API) content in the pharmaceuticals prepared in this way. While the HPLC method has high accuracy, it has the disadvantage that it cannot measure all of the large numbers of products produced due to the long measurement time. To solve this problem, a Raman spectrometer was installed in the pharmaceutical production process to measure the API content of all products. After installing a Raman spectrometer in the pharmaceutical production process, the verified quantitative model was used to measure API content in real time. A total of 4500 vials from three batches were tested, and the averages of each of the three batches were 100.0%, 100.2%, and 99.7%.

동적 신경망과 압축 센싱을 이용한 효율적인 이미지 분류 시스템

한수민(Sumin Han),최장희(Janghee Choi),금호현(Hohyun Keum),이현근(Hyunkeun Lee) 대한전자공학회 2023 대한전자공학회 학술대회 Vol.2023 No.6

This study presents an image classification system using the fusion of compressive sensing (CS) and a dynamic neural network. We developed an end-to-end deep learning framework to optimize the power consumption of the image classification system. The systems performance was analyzed through simulations, comparing accuracy for various compression ratios (CR) for CIFAR-10 datasets.

Quantitative analysis and validation of naproxen tablets by using transmission raman spectroscopy

Jaejin Kim,Janghee Han,Young-Chul Lee,Young-Ah Woo 한국분석과학회 2024 분석과학 Vol.37 No.2

A transmission Raman spectroscopy-based quantitative model, which can analyze the content of a drug product containing naproxen sodium as its active pharmaceutical ingredient (API), was developed. Compared with the existing analytical method, i.e., high-performance liquid chromatography (HPLC), Raman spectroscopy exhibits high test efficiency owing to its shorter sample pre-treatment and measurement time. Raman spectroscopy is environmentally friendly since samples can be tested rapidly via a nondestructive method without sample preparation using solvent. Through this analysis method, rapid on-site analysis was possible and it could prevent the production of defective tablets with potency problems. The developed method was applied to the assays of the naproxen sodium of coated tablets that were manufactured in commercial scale and the content of naproxen sodium was accurately predicted by Raman spectroscopy and compared with the reference analytical method such as HPLC. The method validation of the new approach was also performed. Further, the specificity, linearity, accuracy, precision, and robustness tests were conducted, and all the results were within the criteria. The standard error of cross-validation and standard error of prediction values were determined as 0.949 % and 0.724 %, respectively.

Clinical Characteristics of Trauma-Related Chronic Osteomyelitis in 3 Wild Raccoon Dogs (Nyctereutes procyonoides)

Ha, Minjong,Ahmed, Sohail,Lee, Do Na,Han, Janghee,Yoon, Junghee,Yeon, Seong-Chan The Korean Society of Veterinary Clinics 2022 한국임상수의학회지 Vol.39 No.3

Osteomyelitis typically occurs because of the direct inoculation of bacteria or fungi after penetrating trauma or surgical contamination or, by extension, from soft tissue infection. Osteomyelitis is rarely reported in wildlife animals, though severe chronic osteomyelitis cases do exist in wildlife owing to the scarcity of medical support in the wild environment. This report describes three cases of chronic osteomyelitis in wild raccoon dogs related to trauma. The typical symptoms of three reported cases were ataxia, stiffness, muscle atrophy, and lethargy. All three cases were relevant to traumatic or severe external injury, and skin infestation caused by ectoparasites was apparent on an ocular inspection. In the radiographic examination, diffuse sites of osteolytic lesions and remarkable periosteal responses were demonstrated around the injured limb in all three cases. Apparent neutrophilia with a left shift, lymphocytosis, and monocytosis in hematological examinations generally indicated chronic infection as shown in case 1 and 3. Treatment was attempted with broad-spectrum antibiotics and non-steroidal anti-inflammatory drugs, such as amoxicillin/clavulanic acid, enrofloxacin, clindamycin, and meloxicam. These treatment options helped improve the overall prognosis of chronic osteomyelitis, but the outcomes did not meet the treatment goal entirely. Osteomyelitis can be extremely challenging to treat, particularly in wild animals, because of their distinctive traits, such as masking phenomenon and uncontrolled exposure to ectoparasites. Earlier diagnosis with a radiographic examination, hematological examinations, and careful patient monitoring, followed by prolonged antibiotic therapy and restricted exercise, are the key factors leading to a better prognosis.

분산 공유 메모리 시스템에서 동적 공유 메모리 할당 기법이 거짓 공유에 미치는 영향

이종우(JongWoo Lee),김문희(MoonHee Kim),한장희(JangHee Han),지대구(DaeKu Ji),윤종완(JongWan Yoon),김장선(JangSeon Kim) 한국정보과학회 1997 정보과학회논문지 : 시스템 및 이론 Vol.24 No.12

거짓 공유는 공유 메모리 다중 처리기 시스템에서 여러 처리기들이 일관성 유지의 단위 메모리 영역을 공유함으로 인해 발생하는 현상으로써, 메모리 일관성 유지의 정확성에는 아무런 도움을 주지 못하면서 그 비용만 증가시키는 주요 요인이다. 특히 DSM(분산 공유 메모리) 시스템처럼 메모리 일관성 유지의 단위가 큰(일반적으로, 가상 페이지) 경우에는 그 피해가 더 커진다고 할 수 있다. 본 논문에서는 동적 공유 메모리 할당자를 통해 공유 데이타 영역을 생성하는 병렬 응용들을 대상으로 공유 메모리 할당 기법이 거짓 공유의 발생 빈도에 어떠한 영향을 미치는지 분석하고, 이를 토대로 분산 공유 메모리 시스템에서 거짓 공유 감소에 도움을 주기 위한 동적 공유 메모리 할당 기법을 제시한다. 본 논문에서는 거짓 공유에 영향을 미치는 동적 공유 메모리 할당 방식으로 “객체 크기 별 할당 방식”과 “태그 분리 할당 방식”, 그리고 “다중 페이지 걸침 최소화”를 제시하였으며, 이 기법들의 효용성을 검증하기 위해 실제 병렬 응용에 기반한 실행-기반 시뮬레이션 기법을 사용하였다. 이를 통해 우리는 이 세 가지 방식을 지원하는 할당 기법이 그렇지 않은 할당 기법에 비해 거짓 공유 현상을 적게 유발한다는 것을 확인하였다. False sharing is a result of co-location of unrelated data in the same unit of memory coherency, and is one source of unnecessary overhead being of no help to keep the memory coherency in multiprocessor systems. Moreover, the damage caused by false sharing becomes large in proportion to the granularity of memory coherency. In this paper we analyze the impact of dynamic shared memory allocation techniques on the degree of false sharing in parallel applications communicating with each process by dynamically allocated shared heap. And we propose several allocation techniques for reducing false sharing misses in page-based DSM systems. They include "avoiding multi-page spanning", "separate tag", and "sized or semi-sized" allocation techniques. We use execution-driven simulation of real parallel applications to evaluate the effectiveness of our allocation techniques. And we can find out that by using our dynamic shared memory allocation techniques a considerable amount of false sharing misses can be reduced and so the overhead of memory coherence protocol can also be alleviated.

NT - SPLASH : Win32 API를 이용한 Windows - NT 용 병렬 벤치마크 프로그램

이종우(JongWoo Lee),윤종완(JongWan Yoon),김문희(MoonHee Kim),지대구(DaeKu Ji),한장희(JangHee Han),김장선(JangSeon Kim) 한국정보과학회 1998 한국정보과학회 학술발표논문집 Vol.25 No.1A

Windows-NT 운영체제(이하 NT)는 최근 들어 PC 뿐만 아니라 공유 메모리 다중처리기(이하 SMP) 서버들에 의해서도 많이 사용되는 등 그 범용성이 확대되고 있는 추세이다. 본 논문에서는 SMP 서버 운영체제로서의 NT를 위한 병렬 벤치마크 프로그램 이식에 관한 내용을 다루고자 한다. SPLASH는 미국 스탠포드 대학에서 SMP 유닉스 환경을 위해 개발된 병렬 벤치마크 프로그램으로서 성능 평가를 위해 여러 연구들에 의해 흔히 사용되는 도구이다. 본 논문에서는 SPLASH에서 사용하는 m4 매크로들을 각 기능 별로 해당 Win32 함수로 수정하였으며, 실제 수행을 통해 일부 매크로의 일부 오류를 디버깅하였다. 이식을 위해 수정한 기능들로는 다중 스레드 관리 및 스레드 간 동기화, 스레드 간 공유 메모리 관리, 기타 라이브러리 API등 이었다. 또한 이식된 프로그램들의 정상 동작 여부를 판단하기 위해 이식된 프로그램들을 1 CPU NT 서버와 2 CPU NT 서버에서 수행시켰으며 이를 통해 이들의 정상 동작을 확인하였다. 본 논문이 기여하는 바는 NT를 위한 병렬 벤치마크 테스트 프로그램을 지원함으로써 NT 기반의 시스템 관련 연구 활성화에 도움을 줄 수 있다는 점이라 하겠다.

산불 진압을 위한 딥러닝 기반 소화탄 투하지점 자동 추천 시스템 가능성 연구

신성균(Sung Gyun Shin),김주연(Joo Yeon Kim),장승수(Seungsoo Jang),이민재(Min-Jae Lee),한상수(Sangsoo Han),최찬호(Chan-Ho Choi),김성겸(Sungkyum Kim),조우성(Woo-Sung Cho),이장희(Janghee Lee),김송현(Song Hyun Kim) 한국정보기술학회 2021 한국정보기술학회논문지 Vol.19 No.3

For suppression of wildfire, unmanned aerial vehicles (UAVs) have paid attention. Individual UAV for the fire suppression is generally controlled by human; however, it is difficult to utilize it for the environment including loss of communications as well as requiring large human resources for controlling multiple UAVs. This study aims at developing an automatic estimation system of release point for overcoming the operation problems of UAV in wildfire. For the automatic detection and localization of wildfire, semantic segmentation, which is one of the deep learning techniques, is used; the recommendation algorithm of the release point is proposed using the locailization information. After conducting the machine learning, the accuracy on the proposed release point was estimated over 90%, which agrees well with the location proposal of human. It is expected that the algorithm proposed in this study can be utilized for developing fully-automatic system of fire suppression with UAV.

산불의 효과적 진압을 위한 인공지능 및 영상기반 드론 임무제어 시스템

이민재(Min-Jae Lee),신상균(Sung-Gyun Shin),김주연(Joo-Yeon Kim),장승수(Seungsoo Jang),한상수(Sangsoo Han),최찬호(Chan-Ho Choi),조우성(Woo-Sung Cho),이장희(Janghee Lee),김송현(Song-Hyun Kim) 한국정보기술학회 2022 한국정보기술학회논문지 Vol.20 No.1

Suppression of forest fire with beyond visual line of sight drone operation is difficult because of factors including forest topography, obstacles and others. Since air support and human access at night are difficult for safety reason, drone has paid attention. This study proposes the drone control system which nonprofessional humans also use the drone for forest fire suppression. This system continuously tracks the fire origin selected by operator with deep-learning technique and guides drone to fire origin with collision avoidance. Additionally, this guidance technique apply to vision-based precision landing for minimizing the memory consumption of mission computer. it is expected that this system can be utilized for suppression of forest fire and non professional people also easily use in the future.