• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.11a
• /
• pp.475-477
• /
• 2022

최근 강화 학습을 적용한 게임 AI 에 대한 연구가 활발히 진행되고 있다. 하지만 대부분 상용게임은 유한 상태 머신(Finite State Machine, FSM)을 이용한 스크립트 기반 AI 를 사용하기 때문에 복잡한 환경의 게임에서 불안정한 상태로 인해 적절한 강화 학습의 수행이 어렵다. 따라서 본 논문에서는 상용 게임 강화 학습 적용을 위하여 상태 정규화 및 Dense 보상 기반 강화 학습 기법을 제안한다. 제안한 기법을 상용 농구 게임에 적용하고 학습된 모델의 성능을 기존 FSM 기반 AI 와 비교를 통해 성능이 약 80% 증가한 결과를 확인하였다.

• International Journal of Advanced Culture Technology
• /
• v.11 no.4
• /
• pp.417-422
• /
• 2023

AI blockchain technology, which is attracting attention as a core technology of the 4th Industrial Revolution, is a technology that can be used as an important means of innovation not only in the current gaming industry but also in various industrial fields. This paper extracts the platforms and types of blockchain games currently ranked within the top 100 on the blockchain app (DApp) sites State Of The DApps, DApp.com, and Dapp Rader and introduces the top games on major platforms. As a result of extracting platforms and types, the top games were mainly based on Ethereum, EOS, and Steam. However, the results showed that there are significantly more games based on the Ethereum platform, which are stable, easy to apply, and have a low barrier to entry due to the large number of users and DApps. We plan to improve awareness of blockchain games by studying the characteristics that only blockchain games have.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2024.01a
• /
• pp.375-377
• /
• 2024

본 논문에서는 OpenAI Gym 환경에서 프로그램으로 간단한 제어가 가능한 Mountain-Car-v0 게임에 대해 DQN(Deep Q-Networks) 강화학습을 진행하였다. 본 논문에서 적용한 DQN 네트워크는 입력층 1개, 은닉층 3개, 출력층 1개로 구성하였고, 입력층과 은닉층에서의 활성화함수는 ReLU를, 출력층에서는 Linear함수를 활성화함수로 적용하였다. 실험은 Mountain-Car-v0에 대해 DQN 강화학습을 진행했을 때 각 에피소드별로 획득한 보상 결과를 살펴보고, 보상구간에 포함된 횟수를 분석하였다. 실험결과 전체 100회의 에피소드 중 보상을 50 이상 획득한 에피소드가 85개로 나타났다.

• Journal of Korea Game Society
• /
• v.8 no.4
• /
• pp.95-105
• /
• 2008

In virtual games, Non-Playing Character(NPC)s as game elements tend to frequently communicate with game players. Although the artificial-intelligence (AI) algorithm widely used for games has been greatly developed, basic roles of NPCs have remained on the same. In a life game whose goal is to observe the actions and behaviors of the human-like NPCs, for example, their straightahead actions cause boredom. Actually, NPCs fail to display their various expressions that are characterized by humans. To make NPCs act like humans, several characters with a greater variety of characteristics need to be created. this paper proposes how NPCs both express the wide range of emotions using probability distribution and react based on their different characteristics. To verify the change of NPC actions, personalities were assigned according to the probability distribution and this algorithm was applied to a 3D game to validate the method suggested in this paper.

• Journal of Information Technology Services
• /
• v.16 no.4
• /
• pp.137-148
• /
• 2017

Researches on the Artificial Intelligence has been explosively activated in various fields since the advent of AlphaGo. Particularly, researchers on the application of multi-layer neural network such as deep learning, and various machine learning algorithms are being focused actively. In this paper, we described a development of an artificial intelligence Janggi game based on reinforcement learning algorithm and MCTS (Monte Carlo Tree Search) algorithm with accumulated game data. The previous artificial intelligence games are mostly developed based on mini-max algorithm, which depends only on the results of the tree search algorithms. They cannot use of the real data from the games experts, nor cannot enhance the performance by learning. In this paper, we suggest our approach to overcome those limitations as follows. First, we collects Janggi expert's game data, which can reflect abundant real game results. Second, we create a graph structure by using the game data, which can remove redundant movement. And third, we apply the reinforcement learning algorithm and MCTS algorithm to select the best next move. In addition, the learned graph is stored by object serialization method to provide continuity of the game. The experiment of this study is done with two different types as follows. First, our system is confronted with other AI based system that is currently being served on the internet. Second, our system confronted with some Janggi experts who have winning records of more than 50%. Experimental results show that the rate of our system is significantly higher.

• Journal of Korea Game Society
• /
• v.11 no.6
• /
• pp.61-70
• /
• 2011

In this paper, we propose a defense NPC control model in the soccer game by applying the Decision Tree learning algorithm. The proposed model extracts the direction patterns and the action patterns generated by many soccer game users, and applies these patterns to the Decision Tree learning algorithm. Then, the proposed model decides the direction and the action according to the learned Decision Tree. Experimental results show that the proposed model takes some time to learn the Decision Tree while the proposed model takes 0.001-0.003 milliseconds to decide the direction and the action based on the learned Decision Tree. Therefore, the proposed model can control NPC in the soccer game system in real time. Also, the proposed model achieves higher accuracy than a previous model (Letia98); because the proposed model can utilize current state information, its analyzed information, and previous state information.

• Journal of Korea Game Society
• /
• v.18 no.4
• /
• pp.75-82
• /
• 2018

With the development of information and communication technology, various methods have been tried to provide learners with a fun educational environment through fun and interest. It is a good example to utilize technologies such as games and robots in education for edutainment and game-based learning. In this study, we propose an intelligent rhythm education system using user data collection and analysis for humanoid robot rhythm generation. To do this, the user selects music and inputs rhythm information according to the selected music. The robot utilization data of this user extracts patterns through collection and analysis. Patterns are based on frequency, and FFT similarity comparison method is applied when past data is insufficient. The proposed method is validated through experiments of kindergarten children.

• Proceedings of the Korea Contents Association Conference
• /
• 2005.11a
• /
• pp.523-528
• /
• 2005

In this paper, we designed and made a bicycle with tangible interface using a game to make a tangible interface for health-care system. In particular, through a survey of the preference of users, we have showed a tangible machine which is possible to recognize users' control as a form of tangible interface, have presented the controllable method of a game contents, have provided the contents reality maximized using game AI and game physics.

• Journal of Korea Game Society
• /
• v.17 no.3
• /
• pp.63-70
• /
• 2017

In AI research Go is regarded as the most challenging board game due to the positional evaluation difficulty and the huge branching factor. MCTS is an exciting breakthrough to overcome these problems. The idea behind AlphaGo was to estimate the winning rate of a given position and then to lead deeper search for finding the best promising move. In this paper, using non-strategic MCTS we verified the fact that most pro players regard the best first move as Tengen (Origin of heaven) in $9{\times}9$ Go is correct. We also compared the average winning rates of the most popular first moves.

• Convergence Security Journal
• /
• v.22 no.2
• /
• pp.115-120
• /
• 2022

This study analyzed the recent trends in the sports environment to which big data and AI technologies, which are representative technologies of the 4th Industrial Revolution, and approached them from the perspective of convergence of big data and AI technologies in the sports field. And the results are as follows. First, it is being used for player and game data analysis and team strategy establishment and operation. Second, by combining big data collected using GPS, wearable equipment, and IoT with artificial intelligence technology, scientific physical training for each player is possible through user individual motion analysis, which helps to improve performance and efficiently manage injuries. Third, with the introduction of an AI-based judgment system, it is being used for judge judgment. Fourth, it is leading the change in marketing and game broadcasting services. The technology of the 4th Industrial Revolution is bringing innovative changes to all industries, and the sports field is also in the process. The combination of big data and AI is expected to play an important role as a key technology in the rapidly changing future in a sports environment where scientific analysis and training determine victory or defeat.