• Annual Conference of KIPS
• /
• 2022.11a
• /
• pp.603-605
• /
• 2022

추천시스템은 개개인의 성향에 따른 맞춤화 추천이 가능하기 때문에 음악, 영상, 뉴스 등 많은 분야에서 관심을 받고 있다. 일반적인 추천시스템 모델은 블랙박스 모델이기 때문에 추천 결과에 따른 원인 도출을 할 수 없다. 하지만 XAI 의 모델은 이러한 블랙박스 모델의 단점을 해결하고자 제안되었다. 그 중 KGAT 는 Attention Score 를 기반으로 추천 결과에 따른 원인을 알 수 있다. 이와 같은 AI, XAI 등의 딥 러닝 모델에서 각각의 활성화 함수는 상황에 따라 상이한 성능을 나타낸다. 이러한 이유로 인해 데이터에 맞는 활성화 함수를 적용해보는 다양한 시도가 필요하다. 따라서 본 논문은 XAI 추천시스템 모델인 KGAT 의 성능 개선을 위해 여러 활성화 함수를 적용해보고, 실험을 통해 수정한 모델의 성능이 개선됨을 보인다.

• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.281-286
• /
• 2023

더욱 사람같은 대화 모델을 실현하기 위해, 페르소나 메모리를 활용하여 응답을 생성하는 연구들이 활발히 진행되고 있다. 다수의 기존 연구들에서는 메모리로부터 관련된 페르소나를 찾기 위해 별도의 검색 모델을 이용한다. 그러나 이는 전체 시스템에 속도 저하를 일으키고 시스템을 무겁게 만드는 문제가 있다. 또한, 기존 연구들은 페르소나를 잘 반영해 응답하는 능력에만 초점을 두는데, 그 전에 페르소나 참조의 필요성 여부를 판별하는 능력이 선행되어야 한다. 따라서, 우리의 제안 모델은 검색 모델을 활용하지 않고 생성 모델의 내부적인 연산을 통해 페르소나 메모리의 참조가 필요한지를 판별한다. 참조가 필요하다고 판단한 경우에는 관련된 페르소나를 반영하여 응답하며, 그렇지 않은 경우에는 대화 컨텍스트에 집중하여 응답을 생성한다. 실험 결과를 통해 제안 모델이 장기적인 대화에서 효과적으로 동작함을 확인하였다.

• Journal of Practical Engineering Education
• /
• v.15 no.1
• /
• pp.23-29
• /
• 2023

For given the importance of elementary school students developing the ability to solve problems using artificial intelligence (AI), problem-solving abilities should be developed using AI along with education to develop problem-solving abilities. Such students need a form that allows them to understand the concepts and principles of AI and to be easily educated in a fun way to understand basic understanding of how AI works. To this end, this study planned an 8-hour AI convergence program and operated based on self-driving cars, demonstrating that it was effective in improving elementary school students' problem-solving abilities, creativity, and AI understanding. As a result of operating the camp, students' understanding of AI was 3.56 (standard deviation 0.85), 4.00 (standard deviation 0.71), and t-value was -5.412 (p<0.001), indicating statistically improved understanding of AI, and high satisfaction and interest of students. In the future, it will be necessary to develop an educational program that allows elementary school students to devise their own ideas and create products to which AI models can be applied.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.14 no.5
• /
• pp.969-976
• /
• 2019

A lightweight artificial intelligence hardware has made great strides in many application areas. In general, a lightweight artificial intelligence system consist of lightweight artificial intelligence engine and preprocessor including feature selection, generation, extraction, and normalization. In order to achieve optimal performance in broad range of applications, lightweight artificial intelligence system needs to choose a good preprocessing function and set their respective hyper-parameters. This paper proposes a unified framework for a lightweight artificial intelligence system and utilization method for finding models with optimal performance to use on a given dataset. The proposed unified framework can easily generate a model combined with preprocessing functions and lightweight artificial intelligence engine. In performance evaluation using handwritten image dataset and fall detection dataset measured with inertial sensor, the proposed unified framework showed building optimal artificial intelligence models with over 90% test accuracy.

• The Journal of Bigdata
• /
• v.7 no.1
• /
• pp.1-14
• /
• 2022

The proof-of-work consensus algorithm used by most blockchains is causing a massive waste of computing resources in the form of mining. A useful proof-of-work consensus algorithm has been studied to reduce the waste of computing resources in proof-of-work, but there are still resource waste and mining centralization problems when creating blocks. In this paper, the problem of resource waste in block generation was solved by replacing the relatively inefficient computation process for block generation with distributed artificial intelligence model learning. In addition, by providing fair rewards to nodes participating in the learning process, nodes with weak computing power were motivated to participate, and performance similar to the existing centralized AI learning method was maintained. To show the validity of the proposed methodology, we implemented a blockchain network capable of distributed AI learning and experimented with reward distribution through resource verification, and compared the results of the existing centralized learning method and the blockchain distributed AI learning method. In addition, as a future study, the thesis was concluded by suggesting problems and development directions that may occur when expanding the blockchain main network and artificial intelligence model.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2022.06a
• /
• pp.445-446
• /
• 2022

In this paper, in order to improve port work preparation and berth scheduling efficiency in an artificial intelligence-based berthing monitoring system that can monitor the ship's berthing process, we develop a mooring line recognition model to check an exact berthing time. By improving the pre-designed AI model, it is possible to segment the mooring line from the input image, and to recognize when the mooring line arrives or falls on the berth, thereby providing the correct ship's berthing time. The proposed AI model confirmed by the results that mooring line recognition is possible with evaluation data about the actual berthing situation.

• Journal of the Korean Society of Radiology
• /
• v.83 no.6
• /
• pp.1208-1218
• /
• 2022

In the medical field, artificial intelligence has been used in various ways with many developments. However, most artificial intelligence technologies are developed so that one model can perform only one task, which is a limitation in designing the complex reading process of doctors with artificial intelligence. Multi-task learning is an optimal way to overcome the limitations of single-task learning methods. Multi-task learning can create a model that is efficient and advantageous for generalization by simultaneously integrating various tasks into one model. This study investigated the concepts, types, and similar concepts as multi-task learning, and examined the status and future possibilities of multi-task learning in the medical research.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2007.04a
• /
• pp.58-62
• /
• 2007

인간이 생성한 음성, 표정 영상, 문장 등으로부터 인간의 감정 상태를 인식하는 연구와 함께, 인간의 감정을 모방하여 다양한 외부 자극으로 감정을 생성하는 인공 감정(Artificial Emotion)에 관한 연구가 이루어지고 있다. 그러나 기존의 인공 감정 연구는 외부 감정 자극에 대한 감정 변화 상태를 선형적, 지수적으로 변화시킴으로써 감정 상태가 급격하게 변하는 형태를 보인다. 본 논문에서는 외부 감정 자극의 강도와 빈도뿐만 아니라 자극의 반복 주기를 감정 상태에 반영하고, 시간에 따른 감정의 변화를 Sigmoid 곡선 형태로 표현하는 감정 생성 모델을 제안한다. 그리고 기존의 감정 자극에 대한 회상(recollection)을 통해 외부 감정 자극이 없는 상황에서도 감정을 생성할 수 있는 인공 감정 시스템을 제안한다.

• Journal of the Korean Society of Radiology
• /
• v.18 no.3
• /
• pp.267-273
• /
• 2024

Using a cloud-based vertex AI platform that can develop an artificial intelligence learning model without coding, this study easily developed an artificial intelligence learning model by the non-professional general public and confirmed its clinical applicability. Nine dental diseases and 2,999 root disease X-ray images released on the Kaggle site were used for the learning data, and learning, verification, and test data images were randomly classified. Image classification and multi-label learning were performed through hyper-parameter tuning work using a learning pipeline in vertex AI's basic learning model workflow. As a result of performing AutoML(Automated Machine Learning), AUC(Area Under Curve) was found to be 0.967, precision was 95.6%, and reproduction rate was 95.2%. It was confirmed that the learned artificial intelligence model was sufficient for clinical diagnosis.

• Convergence Security Journal
• /
• v.24 no.3
• /
• pp.81-93
• /
• 2024

The purpose of this paper is to derive considerations when improving the Personal Information Protection Act based on the personal information protection life cycle of the generative artificial intelligence model as generative artificial intelligence models are introduced and used in Korea a lot. Through the study, the necessity of using open information in the collection stage, using personal information preservation technology in the learning stage, and preparing the basis for the development of protection technology in the holding stage was derived. It also revealed the necessity of managing the generated information in the generation and inference stage, re-learning in the limitation and destruction stage, and preparing a filtering basis. It is expected that the results of this study can be used to revise the Personal Information Protection Act and make policies in the future.