• The Transactions of the Korea Information Processing Society
• /
• v.13 no.10
• /
• pp.497-504
• /
• 2024

Generative Adversarial Networks (GANs) are typically described as a two-player game between a generator and a discriminator, where the generator aims to produce realistic data, and the discriminator tries to distinguish between real and generated data. However, this setup often leads to mode collapse, where the generator produces limited variations in the data, failing to capture the full range of the target data distribution. This paper proposes a new training system to mitigate the mode collapse problem. Specifically, it extends the traditional two-player game of GANs into a multi-player game and introduces a peer-evaluation method to effectively train multiple GANs. In the peer-evaluation process, the generated samples from each GANs are evaluated by the other players. This provides external feedback, serving as an additional standard that helps GANs recognize mode failure. This cooperative yet competitive training method encourages the generators to explore and capture a broader range of the data distribution, mitigating mode collapse problem. This paper explains the detailed algorithm for peer-evaluation based multi-GANs training and validates the performance through experiments.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.7
• /
• pp.53-71
• /
• 2024

Generative Artificial Intelligence (AI) has become the focus of societal attention due to its wide range of applications and profound impact. This paper constructs a comprehensive theoretical model based on the Unified Theory of Acceptance and Use of Technology 2 (UTAUT2), integrating variables such as Personal Innovativeness and Perceived Risk to study the key factors influencing enterprises' adoption of Generative AI. We employed Structural Equation Modeling (SEM) to verify the hypothesized paths and used the Bootstrapping method to test the mediating effect of Behavioral Intention. Additionally, we explored the moderating effect of Perceived Risk through Hierarchical Regression Analysis. The results indicate that Performance Expectancy, Effort Expectancy, Social Influence, Price Value, and Personal Innovativeness have significant positive impacts on Behavioral Intention. Behavioral Intention plays a significant mediating role between these factors and Use Behavior, while Perceived Risk negatively moderates the relationship between Behavioral Intention and Use Behavior. This study provides theoretical and empirical support for how enterprises can effectively adopt Generative AI, offering important practical implications.

• Journal of the Korea Society of Computer and Information
• /
• v.17 no.8
• /
• pp.31-40
• /
• 2012

In this paper, we propose a method of synthesizing faces of a user and an animation character using a 3D face model. The suggested method first receives two orthogonal 2D face images and extracts major features of the face through the template snake. It then generates a user-customized 3D face model by adjusting a generalized face model using the extracted facial features and by mapping texture maps obtained from two input images to the 3D face model. Finally, it generates a user-customized animation character by synthesizing the generated 3D model to an animation character reflecting the position, size, facial expressions, and rotational information of the character. Experimental results show some results to verify the performance of the suggested algorithm. We expect that our method will be useful to various applications such as games and animation movies.

• Proceedings of the Korean Society for Cognitive Science Conference
• /
• 2000.05a
• /
• pp.143-149
• /
• 2000

본고에서는 불어 자동처리 연구의 기초이면서, 불어 초·중급 학습에 가장 큰 걸림돌로 여겨지는 불어 굴절 변화형의 분석 및 생성 교육 시스템 Inflection-edu를 소개한다. inflection-edu는 부산대학교 언어 정보 연구실에서 개발한 불어 형태소 분석-생성기를 기반으로 하였으며, 동사 8,249개, 명사 29,059개, 형용사 9,957개와 그 굴절 변화형을 모두 분석 및 생성할 수 있으며, 학습자를 위한 굴절 규칙 231개를 포함한다. 제 2 장에는 분석과 생성을 위한 모델화(modelling) 방법론을 제시하고, 제 3 장에서는 이 결과를 불어 교육 시스템에 연동하기 위한 인터페이스를 제시하고, 제 4 장에서는 Inflection-edu의 인터페이스를 소개한다. 제 5 장에서는 남은 문제와 향후 응용 방향을 알아본다. 형태소 분석기와 생성기능이 교육 시스템에 통합된 Inflection-edu는 70년대 개발된 프로그램과 같이 단순하고 반복적인 교수-학습 작업을 제공하는 것이 아니다. 학습자의 요구에 정확하고 빠르게 피드백을 줄 수 있으며, 좀 더 큰 단위의 분석 및 생성이 가능하도록 하여, 좀 더 지능적인 언어 교육 시스템을 구현하는 것을 그 개발 목표로 하고 있다.

• Journal of KIISE:Computing Practices and Letters
• /
• v.15 no.3
• /
• pp.226-230
• /
• 2009

Intelligent Tutoring System (ITS) is a procedure of analyzing collected data for teaming, making a strategy and performing adequate service for learners. To perform suitable service for learners, modeling is the first step to collect data from the process of their learning. The model, however, cannot be authentic if collected data can contain learners' inconsistent behaviors or unpredictable learning inclination. This study focused on how to sort normal and abnormal data by analyzing collected data from learners through visualization. A model has been set up to assort unusual data from collected learner's data by using DOLLS-HI which makes possible to diagnose learner's learning propensity based on housing interior learning contents in the experiment. The created model has been confirmed its improved reliability comparing to previous one.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2002.12a
• /
• pp.199-202
• /
• 2002

신경회로망을 동적 정보처리에 응용하기 위해서는 비대칭 결합 신경회로망에서 생성되는 동적 상태천이에 관한 직관적 이해가 필요하다. 자기결합을 갖고 결합하중치가 비대칭인 순환결합형 신경회로망은 복수 개의 리미트사이클이 기억 가능하다는 것이 알려져 있다. 현재까지 이산시간 모델의 네트워크에 대한 상태천이 해석은 상세하게 이루어져 왔다. 그러나 연속시간 모델에 대한 해석은 네트워크 규모의 증가에 따른 급격한 계산량의 증가 때문에 연구가 그다지 활발하게 이루어지지 않고 있다. 본 논문에서는 각 뉴런이 최근접 뉴런에만 이진화된 결합하중 +1 및 -1로 연결된 연속시간모델 순환결합형 신경회로망의 동적인 상태천이 특성을 해석하여 이산시간 모델에서 기억 가능한 리미트사이클과의 차이점을 분석한다. 또한 연속시간 네트워크 모델에 카오스 신호를 인가하여 리미트사이클간의 천이를 제어할 수 있는 가능성을 분석하여 동적정보처리에 네트워크를 응용할 수 있는 가능성을 검토한다.

• Proceedings of the Korea Multimedia Society Conference
• /
• 2003.11b
• /
• pp.626-629
• /
• 2003

본 논문은 회화적 렌더링에 있어서 브러쉬 스트로크의 방향을 결정하는 새로운 방법을 제안한다. 전류가 흐르는 도선 주위에는 자기장이 생성된다는 물리적 이론을 기초로 자기장 모델의 벡터를 생성한다 이 모델을 이용하여 원형(circular) 스트로크나 방사형(emissive)의 벡터 필드를 만들어 내고 스트로크에 적용함으로써 고흐와 같은 화가의 브러쉬 기법을 효과적으로 표현할 수 있다.

• Journal of Internet Computing and Services
• /
• v.25 no.1
• /
• pp.79-90
• /
• 2024

This study focused on user trust in ChatGPT, a generative AI technology, and explored the factors that affect usage status and intention to continue using, and whether the influence of trust varies depending on the purpose. For this purpose, the survey was conducted targeting people in their 20s and 30s who use ChatGPT the most. The statistical analysis deploying IBM SPSS 27 and SmartPLS 4.0. A structural equation model was formulated on the foundation of Bhattacherjee's Expectation-Confirmation Model (ECM), employing path analysis and Multi-Group Analysis (MGA) for hypothesis validation. The main findings are as follows: Firstly, ChatGPT is mainly used for specific needs or objectives rather than as a daily tool. The majority of users are cognizant of its hallucination effects; however, this did not hinder its use. Secondly, the hypothesis testing indicated that independent variables such as expectation- confirmation, perceived usefulness, and user satisfaction all exert a positive influence on the dependent variable, the intention for continuance intention. Thirdly, the influence of trust varied depending on the user's purpose in utilizing ChatGPT. trust was significant when ChatGPT is used for information retrieval but not for creative purposes. This study will be used to solve reliability problems in the process of introducing generative AI in society and companies in the future and to establish policies and derive improvement measures for successful employment.

• Korean Journal of Food Science and Technology
• /
• v.52 no.6
• /
• pp.565-572
• /
• 2020

The 3-monochloropropane-1,2-diol (3-MCPD) is a contaminant that occurs in foodstuffs in its free form as well as in its bound form. The objective of this study was to evaluate the effects of emulsifier, frying temperature, and the amounts of salt and oil on the formation of bound 3-MCPD in a fried snack model. Emulsifier affected the formation of bound 3-MCPD; furthermore, it was observed that the largest amount of bound 3-MCPD was detected in the fried snack model when glycerin esters of fatty acids were used as emulsifiers. Frying temperature also affected the formation of bound 3-MCPD, which increased significantly as the frying temperature increased from 145 to 190℃. In addition, salt affected the formation of bound 3-MCPD. As the amount of salt increased, the amount of bound 3-MCPD also increased significantly. Moreover, it was observed that the amount of oil did not affect the formation of bound 3-MCPD. These results will aid in the reduction of bound 3-MCPD in fried snacks.

• The Journal of the Convergence on Culture Technology
• /
• v.10 no.4
• /
• pp.305-310
• /
• 2024

With the emergence of generative AI, a new era of coexistence with humanity has begun. The vast data-driven learning capabilities of AI are being utilized in various industries to achieve a level of productivity distinct from human learning. However, AI also manifests societal phenomena such as technophobia. This study aims to analyze collaborative AI models based on an understanding of AI and identify areas within the jewelry industry where these models can be applied. The utilization of collaborative AI models can lead to the acceleration of idea development, enhancement of design capabilities, increased productivity, and the internalization of multimodal functions. Ultimately, AI should be used as a collaborative tool from a utilitarian perspective, which requires a proactive, human-centric mindset. This research proposes collaborative AI strategies for the jewelry business, hoping to enhance the industry's competitiveness.