• 방송공학회논문지
• /
• 제28권3호
• /
• pp.302-313
• /
• 2023

본 연구는 과학기술정보통신부가 2017년부터 1조원 이상을 투자한 'AI Hub 댐' 사업에서 구축된 인공지능 모델 학습데이터의 품질관리를 자동화할 수 있는 프레임워크의 개발을 목표로 한다. 자율주행 개발에 사용되는 AI 모델 학습에는 다량의 고품질의 데이터가 필요하며, 가공된 데이터를 검수자가 데이터 자체의 이상을 검수하고 유효함을 증명하는 데는 여전히 어려움이 있으며 오류가 있는 데이터로 학습된 모델은 실제 상황에서 큰 문제를 야기할 수 있다. 본 논문에서는 이상 데이터를 제거하는 신뢰할 수 있는 데이터셋 정제 프레임워크를 통해 모델의 인식 성능을 향상시키는 전략을 소개한다. 제안하는 방법은 인공지능 학습용 데이터 품질관리 가이드라인의 지표를 기반으로 설계되었다. 한국정보화진흥원의 AI Hub을 통해 공개된 자율주행 데이터셋에 대한 실험을 통해 프레임워크의 유효성을 증명하였고, 이상 데이터가 제거된 신뢰할 수 있는 데이터셋으로 재구축될 수 있음을 확인하였다.

• Asia pacific journal of information systems
• /
• 제34권1호
• /
• pp.150-190
• /
• 2024

This article proposes a framework to elucidate the structural dynamics of the generative AI ecosystem. It also outlines the practical application of this proposed framework through illustrative policies, with a specific emphasis on the development of the Korean generative AI ecosystem and its implications of platform strategies at AI platform-squared. We propose a comprehensive classification scheme within generative AI ecosystems, including app builders, technology partners, app stores, foundational AI models operating as operating systems, cloud services, and chip manufacturers. The market competitiveness for both app builders and technology partners will be highly contingent on their ability to effectively navigate the customer decision journey (CDJ) while offering localized services that fill the gaps left by foundational models. The strategically important platform of platforms in the generative AI ecosystem (i.e., AI platform-squared) is constituted by app stores, foundational AIs as operating systems, and cloud services. A few companies, primarily in the U.S. and China, are projected to dominate this AI platform squared, and consequently, they are likely to become the primary targets of non-market strategies by diverse governments and communities. Korea still has chances in AI platform-squared, but the window of opportunities is narrowing. A cautious approach is necessary when considering potential regulations for domestic large AI models and platforms. Hastily importing foreign regulatory frameworks and non-market strategies, such as those from Europe, could overlook the essential hierarchical structure that our framework underscores. Our study suggests a clear strategic pathway for Korea to emerge as a generative AI powerhouse. As one of the few countries boasting significant companies within the foundational AI models (which need to collaborate with each other) and chip manufacturing sectors, it is vital for Korea to leverage its unique position and strategically penetrate the platform-squared segment-app stores, operating systems, and cloud services. Given the potential network effects and winner-takes-all dynamics in AI platform-squared, this endeavor is of immediate urgency. To facilitate this transition, it is recommended that the government implement promotional policies that strategically nurture these AI platform-squared, rather than restrict them through regulations and stakeholder pressures.

• 한국정보과학회 언어공학연구회:학술대회논문집(한글 및 한국어 정보처리)
• /
• 한국정보과학회언어공학연구회 2023년도 제35회 한글 및 한국어 정보처리 학술대회
• /
• pp.683-689
• /
• 2023

자연어 처리 분야는 최근에 큰 발전을 보였으며, 특히 초대규모 언어 모델의 등장은 이 분야에 큰 영향을 미쳤다. GPT와 같은 모델은 다양한 NLP 작업에서 높은 성능을 보이고 있으며, 특히 챗봇 분야에서 중요하게 다루어지고 있다. 하지만, 이러한 모델에도 여러 한계와 문제점이 있으며, 그 중 하나는 모델이 기대하지 않은 결과를 생성하는 것이다. 이를 해결하기 위한 다양한 방법 중, Retrieval-Augmented Generation(RAG) 방법이 주목받았다. 이 논문에서는 지식베이스와의 통합을 통한 도메인 특화형 질의응답 시스템의 효율성 개선 방안과 벡터 데이터 베이스의 수정을 통한 챗봇 답변 수정 및 업데이트 방안을 제안한다. 본 논문의 주요 기여는 다음과 같다: 1) QA Pair Passage RAG을 활용한 새로운 RAG 시스템 제안 및 성능 향상 분석 2) 기존의 LLM 및 RAG 시스템의 성능 측정 및 한계점 제시 3) RDBMS 기반의 벡터 검색 및 업데이트를 활용한 챗봇 제어 방법론 제안

• 서비스연구
• /
• 제14권1호
• /
• pp.13-26
• /
• 2024

멀티모달 (multi-modal) 생성이란 텍스트, 이미지, 오디오 등 다양한 정보를 기반으로 결과를 도출하는 작업을 말한다. AI 기술의 비약적인 발전으로 인해 여러 가지 유형의 데이터를 종합적으로 처리해 결과를 도출하는 멀티모달 기반 시스템 또한 다양해지는 추세이다. 본 논문은 음성과 텍스트 인식을 활용하여 인물을 묘사하면, 몽타주 이미지를 생성하는 AI 시스템의 개발 내용을 소개한다. 기존의 몽타주 생성 기술은 서양인들의 외형을 기준으로 이루어진 반면, 본 논문에서 개발한 몽타주 생성 시스템은 한국인의 안면 특징을 바탕으로 모델을 학습한다. 따라서, 한국어에 특화된 음성과 텍스트의 멀티모달을 기반으로 보다 정확하고 효과적인 한국형 몽타주 이미지를 만들어낼 수 있다. 개발된 몽타주 생성 앱은 몽타주 초안으로 충분히 활용 가능하기 때문에 기존의 몽타주 제작 인력의 수작업을 획기적으로 줄여줄 수 있다. 이를 위해 한국지능정보사회진흥원의 AI-Hub에서 제공하는 페르소나 기반 가상 인물 몽타주 데이터를 활용하였다. AI-Hub는 AI 기술 및 서비스 개발에 필요한 인공지능 학습용 데이터를 구축하여 원스톱 제공을 목적으로 한 AI 통합 플랫폼이다. 이미지 생성 시스템은 고해상도 이미지를 생성하는데 사용하는 딥러닝 모델인 VQGAN과 한국어 기반 영상생성 모델인 KoDALLE 모델을 사용하여 구현하였다. 학습된 AI 모델은 음성과 텍스트를 이용해 묘사한 내용과 매우 유사한 얼굴의 몽타주 이미지가 생성됨을 확인할 수 있다. 개발된 몽타주 생성 앱의 실용성 검증을 위해 10명의 테스터가 사용한 결과 70% 이상이 만족한다는 응답을 보였다. 몽타주 생성 앱은 범죄자 검거 등 얼굴의 특징을 묘사하여 이미지화하는 여러 분야에서 다양하게 사용될 수 있을 것이다.

• International journal of advanced smart convergence
• /
• 제11권4호
• /
• pp.10-19
• /
• 2022

AI-based Network Intrusion Detection Systems (AI-NIDS) detect network attacks using machine learning and deep learning models. Recently, unsupervised AI-NIDS methods are getting more attention since there is no need for labeling, which is crucial for building practical NIDS systems. This paper aims to test the impact of designing autoencoder models that can be applied to unsupervised an AI-NIDS in real network systems. We collected security events of legacy network security system and carried out an experiment. We report the results and discuss the findings.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2023년도 추계학술발표대회
• /
• pp.842-843
• /
• 2023

Covid-19으로 인한 언택트(Untact) 문화의 확산으로 키오스크 주문과 같은 비대면 서비스가 증가하였다. 본 연구에서는 비대면 서비스로 인해 발생하는 정보격차 및 접근성 문제를 해결하기 위해 AI 기술과 사용자 인터페이스를 결합하여 개인 맞춤화된 키오스크를 소개한다. 본 연구에서 개발된 AI 키오스크는 자연어 처리기술을 활용한 음성 주문을 통해 편리성을 증진하고, 딥러닝 기술을 이용한 연령대 인식, 사용자의 알레르기 정보를 고려한 메뉴 추천을 통해 사용자에게 맞춤화된 서비스를 제공한다. 개발된 키오스크를 통해 개인화된 서비스를 개선할 수 있으며 더불어 정보 취약계층 간의 정보격차를 해소할 수 있다.

• 농촌의학ㆍ지역보건
• /
• 제22권1호
• /
• pp.27-34
• /
• 1997

This study was performed to offer the fundamental data for preventing and controlling the cardiovascular diseases of rural community women. The number of 168 women aged 40 to 50s who live in Kumsan-gun, Chungnam were selected for analysis after medical examination by a Kumsan-gun health center during the period of November to December in 1996. Total cholesterol(TC), triglyceride(TG), high density lipoprotein cholesterol(HDL-C), blood pressure(BP), degree of obesity(BMI) and atherosclerosis index(AI) were measured and relation between these physical measurements to serum lipid levels were studied. The results were as follows; 1. Mean TC level, TG level, BP, BMI and AI were significantly increased with advancing age. 2. Mean TC level, TG level, BMI and AI of borderline BP group and hypertension group were significantly increased than those of normal BP group. 3. Mean TC level, TG level and AI of obesity group were significantly increased than those of normal BMI group. 4. Simple correlation analysis showed that the positive correlation between TC, TG, BP, BMI and AI, but the level of HDL-C was negative correlation with TG and AI. 5. In multiple regression analysis taking HDL-C level as the dependent variable and TG, TC, BP, BMI, Age AI as the independent variable, AI, TG, BMI, TC were significantly related to HDL-C.

• Korean Journal of Radiology
• /
• 제25권4호
• /
• pp.343-350
• /
• 2024

Objective: Artificial intelligence-based computer-aided diagnosis (AI-CAD) is increasingly used in mammography. While the continuous scores of AI-CAD have been related to malignancy risk, the understanding of how to interpret and apply these scores remains limited. We investigated the positive predictive values (PPVs) of the abnormality scores generated by a deep learning-based commercial AI-CAD system and analyzed them in relation to clinical and radiological findings. Materials and Methods: From March 2020 to May 2022, 656 breasts from 599 women (mean age 52.6 ± 11.5 years, including 0.6% [4/599] high-risk women) who underwent mammography and received positive AI-CAD results (Lunit Insight MMG, abnormality score ≥ 10) were retrospectively included in this study. Univariable and multivariable analyses were performed to evaluate the associations between the AI-CAD abnormality scores and clinical and radiological factors. The breasts were subdivided according to the abnormality scores into groups 1 (10-49), 2 (50-69), 3 (70-89), and 4 (90-100) using the optimal binning method. The PPVs were calculated for all breasts and subgroups. Results: Diagnostic indications and positive imaging findings by radiologists were associated with higher abnormality scores in the multivariable regression analysis. The overall PPV of AI-CAD was 32.5% (213/656) for all breasts, including 213 breast cancers, 129 breasts with benign biopsy results, and 314 breasts with benign outcomes in the follow-up or diagnostic studies. In the screening mammography subgroup, the PPVs were 18.6% (58/312) overall and 5.1% (12/235), 29.0% (9/31), 57.9% (11/19), and 96.3% (26/27) for score groups 1, 2, 3, and 4, respectively. The PPVs were significantly higher in women with diagnostic indications (45.1% [155/344]), palpability (51.9% [149/287]), fatty breasts (61.2% [60/98]), and certain imaging findings (masses with or without calcifications and distortion). Conclusion: PPV increased with increasing AI-CAD abnormality scores. The PPVs of AI-CAD satisfied the acceptable PPV range according to Breast Imaging-Reporting and Data System for screening mammography and were higher for diagnostic mammography.

• Korean Journal of Radiology
• /
• 제23권12호
• /
• pp.1241-1250
• /
• 2022

Objective: To conduct a simulation study to determine whether artificial intelligence (AI)-aided mammography reading can reduce unnecessary recalls while maintaining cancer detection ability in women recalled after mammography screening. Materials and Methods: A retrospective reader study was performed by screening mammographies of 793 women (mean age ± standard deviation, 50 ± 9 years) recalled to obtain supplemental mammographic views regarding screening mammography-detected abnormalities between January 2016 and December 2019 at two screening centers. Initial screening mammography examinations were interpreted by three dedicated breast radiologists sequentially, case by case, with and without AI aid, in a single session. The area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and recall rate for breast cancer diagnosis were obtained and compared between the two reading modes. Results: Fifty-four mammograms with cancer (35 invasive cancers and 19 ductal carcinomas in situ) and 739 mammograms with benign or negative findings were included. The reader-averaged AUC improved after AI aid, from 0.79 (95% confidence interval [CI], 0.74-0.85) to 0.89 (95% CI, 0.85-0.94) (p < 0.001). The reader-averaged specificities before and after AI aid were 41.9% (95% CI, 39.3%-44.5%) and 53.9% (95% CI, 50.9%-56.9%), respectively (p < 0.001). The reader-averaged sensitivity was not statistically different between AI-unaided and AI-aided readings: 89.5% (95% CI, 83.1%-95.9%) vs. 92.6% (95% CI, 86.2%-99.0%) (p = 0.053), although the sensitivities of the least experienced radiologists before and after AI aid were 79.6% (43 of 54 [95% CI, 66.5%-89.4%]) and 90.7% (49 of 54 [95% CI, 79.7%-96.9%]), respectively (p = 0.031). With AI aid, the reader-averaged recall rate decreased by from 60.4% (95% CI, 57.8%-62.9%) to 49.5% (95% CI, 46.5%-52.4%) (p < 0.001). Conclusion: AI-aided reading reduced the number of recalls and improved the diagnostic performance in our simulation using women initially recalled for supplemental mammographic views after mammography screening.

• Journal of Oral Medicine and Pain
• /
• 제47권2호
• /
• pp.107-108
• /
• 2022

Artificial intelligence (AI) refers to the use of machines to mimic intelligent human behavior. It involves interactions with humans in clinical settings, and augmented intelligence is considered as a cognitive extension of AI. The importance of AI in healthcare and medicine has been emphasized in recent studies. Machine learning models, such as genetic algorithms, artificial neural networks (ANNs), and fuzzy logic, can learn and examine data to execute various functions. Among them, ANN is the most popular model for diagnosis based on image data. AI is rapidly becoming an adjunct to healthcare professionals and is expected to be human-independent in the near future. The introduction of AI to the diagnosis and treatment of oral diseases worldwide remains in the preliminary stage. AI-based or assisted diagnosis and decision-making will increase the accuracy of the diagnosis and render treatment more precise and personalized. Therefore, dental professionals must actively initiate and lead the development of AI, even if they are unfamiliar with it.