• Proceedings of the Korea Information Processing Society Conference
• /
• 2004.11a
• /
• pp.497-500
• /
• 2004

가상현실은 현실세계에서 경험하기 어려운 환경을 간접적으로 경험할 수 있는 가상의 공간이다. 이러한 가상현실에는 건물, 지형, PC(Playable Character), NPC(Non-Playable Character)등의 다양한 객체들이 존재하게 되며, PC와 NPC와 같은 객체들은 현실감을 주기 위해 인공지능을 가지게 된다. 현재까지 인공지능에 대한 많은 연구가 진행되었으며, 다양한 분야에서 활용되고 있다. 가상현실에서는 유한상태 기계(Finite Sate Machine, FSM), 유전자 알고리즘, 신경망 알고리즘 $A{\ast}$ 알고리즘 등이 활용되고 있으며, FSM은 비교적 알고리즘이 간단하고, 다른 알고리즘에 비해 구현이 간단해 널리 이용되고 있다. 본 논문에서는 FSM을 활용하여 여러 행동 패턴을 정의하고 행동 패턴간 천이가 이루어 짐으로, 객체의 행동을 다양하게 나타낼 수 있는 Multiple FSM은 제안한다.

• The Journal of Korea Robotics Society
• /
• v.14 no.3
• /
• pp.157-162
• /
• 2019

The video game Tetris is one of most popular game and it is well known that its game rule can be modelled as MDP (Markov Decision Process). This paper presents a DQN (Deep Q-Network) based game agent for Tetris game. To this end, the state is defined as the captured image of the Tetris game board and the reward is designed as a function of cleared lines by the game agent. The action is defined as left, right, rotate, drop, and their finite number of combinations. In addition to this, PER (Prioritized Experience Replay) is employed in order to enhance learning performance. To train the network more than 500000 episodes are used. The game agent employs the trained network to make a decision. The performance of the developed algorithm is validated via not only simulation but also real Tetris robot agent which is made of a camera, two Arduinos, 4 servo motors, and artificial fingers by 3D printing.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.5
• /
• pp.205-212
• /
• 2019

Scene graph is a kind of knowledge graph that represents both objects and their relationships found in a image. This paper proposes a 3D scene graph generation model for three-dimensional indoor environments. An 3D scene graph includes not only object types, their positions and attributes, but also three-dimensional spatial relationships between them, An 3D scene graph can be viewed as a prior knowledge base describing the given environment within that the agent will be deployed later. Therefore, 3D scene graphs can be used in many useful applications, such as visual question answering (VQA) and service robots. This proposed 3D scene graph generation model consists of four sub-networks: object detection network (ObjNet), attribute prediction network (AttNet), transfer network (TransNet), relationship prediction network (RelNet). Conducting several experiments with 3D simulated indoor environments provided by AI2-THOR, we confirmed that the proposed model shows high performance.

• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.9
• /
• pp.407-418
• /
• 2023

The Multi-Object Goal Visual Navigation(MultiOn) is a visual navigation task in which an agent must visit to multiple object goals in an unknown indoor environment in a given order. Existing models for the MultiOn task suffer from the limitation that they cannot utilize an integrated view of multimodal context because use only a unimodal context map. To overcome this limitation, in this paper, we propose a novel deep neural network-based agent model for MultiOn task. The proposed model, MCFMO, uses a multimodal context map, containing visual appearance features, semantic features of environmental objects, and goal object features. Moreover, the proposed model effectively fuses these three heterogeneous features into a global multimodal context map by using a point-wise convolutional neural network module. Lastly, the proposed model adopts an auxiliary task learning module to predict the observation status, goal direction and the goal distance, which can guide to learn the navigational policy efficiently. Conducting various quantitative and qualitative experiments using the Habitat-Matterport3D simulation environment and scene dataset, we demonstrate the superiority of the proposed model.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2004.11a
• /
• pp.249-258
• /
• 2004

오늘날 활발하게 이루어지고 있는 유비쿼터스 컴퓨팅 관련 기술 연구는 사용자가 시간과 장소에 구애받지 않고 네트워크에 접근해 다양한 컴퓨터 관련 서비스를 제공 받을 수 있는 방법에 초점을 맞추고 있다. 이 처럼 시간과 공간의 한계를 뛰어 넘은 네트워크로의 자유로운 접근은 일상 생활의 패러다임을 바꾸어 놓게 될 것이다. 유비쿼터스 컴퓨팅 기술을 통해 가장 큰 변화가 일어나는 분야는 일반 가정환경에서 일어나는 인텔리전트 홈 네트워크 (Intelligent Home Network) 라고 할 수 있다. 집에 들어오면, 자동으로 문을 열어주고, 불을 켜주며, 놓쳤던 TV 프로그램을 자동으로 녹화해 놓았다가 원하는 시간에 보여주고, 적당한 시간에 목욕물을 미리 받아준다. 또한 집밖으로 나가기 전, 일기예보에 따라 우산을 챙겨주고, 일정을 확인시켜주며 입고 나갈 옷을 골라줄 수도 있다. 이 모든 일들이 유비쿼터스 컴퓨팅 기술이 가져올 인텔리전트 홈 네트워크의 모습이다. 그러나, 모든 사용자에게 효과적인 서비스를 제공하기 위해서는 홈 네트워크 상의 자원 관리에서 일어날 수 있는 에이전트들간의 자원 접근 권한 충돌을 효율적으로 방지할 수 있는 기술이 필요하다. 유비쿼터스 컴퓨팅 환경에서 자원관리 특성은 점유의 연속성, 자원 사이의 연관성, 그리고 자원과 사용자 사 사이의 연계성의 3 가지 특성을 지니고 있다. 본 논문에서는 유비쿼터스 컴퓨팅 환경에서 일어날 수 있는 자원 충돌 상황을 효율적으로 처리하기 위한 자원 협상 방법을 제안한다. 본 방법은 자원 관리 특성을 바탕으로 시간논리에 기반을 둔 자원 선점과 분배 규칙으로 구성된다.트 시스템은 b-Cart를 기반으로 할 것으로 예측할 수 있다.타났다. 또한, 스네이크의 초기 제어점을 얼굴은 44개, 눈은 16개, 입은 24개로 지정하여 MER추출에 성공한 영상에 대해 스네이크 알고리즘을 수행한 결과, 추출된 영역의 오차율은 각각 2.2%, 2.6%, 2.5%로 나타났다.해서 Template-based reasoning 예를 보인다 본 방법론은 검색노력을 줄이고, 검색에 있어 Feasibility와 Admissibility를 보장한다.매김할 수 있는 중요한 계기가 될 것이다.재무/비재무적 지표를 고려한 인공신경망기법의 예측적중률이 높은 것으로 나타났다. 즉, 로지스틱회귀 분석의 재무적 지표모형은 훈련, 시험용이 84.45%, 85.10%인 반면, 재무/비재무적 지표모형은 84.45%, 85.08%로서 거의 동일한 예측적중률을 가졌으나 인공신경망기법 분석에서는 재무적 지표모형이 92.23%, 85.10%인 반면, 재무/비재무적 지표모형에서는 91.12%, 88.06%로서 향상된 예측적중률을 나타내었다.ting LMS according to increasing the step-size parameter $\mu$ in the experimentally computed. learning curve. Also we find that convergence speed of proposed algorithm is increased by (B+1) time proportional to B which B is the number of recycled data b

• Proceedings of the KSR Conference
• /
• 2007.05a
• /
• pp.1948-1953
• /
• 2007

Intelligent agent is to decide what customers need on the internet and offer them accurate information. In this paper, the system which can recommend the tourism items in terms of customer's needs is proposed by appling the intelligent agent to railway tourism system. Most of previous agents are focused on price. But, this study proposes the Railway tourism system which offers each customer the best suitable information based on quality of information and reputation. The customer's needs are analyzed through intelligent agent and the information which is suitable for customer's needs is obtained the Artificial Neural Network Model.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2001.05a
• /
• pp.676-679
• /
• 2001

As network security is coning up with significant problem after the major Internet sites were hacked nowadays, IDS(Intrusion Detection System) is considered as a next generation security solution for more reliable network and system security rather than firewall. In this paper, we propose the new IDS model which tan detect intrusion in different systems as well as which ran make real-time detection of intrusion in the expanded distributed environment in host level of drawback of existing IDS. We implement its prototype and verify its validity. We use pattern extraction agent so that we can extract automatically audit file needed in distributed intrusion detection even in other platforms.

• KIPS Transactions on Software and Data Engineering
• /
• v.3 no.8
• /
• pp.299-308
• /
• 2014

Various literatures related computing of information processing have been recently shown the researches inspired from the remarkably excellent human capabilities which recognize and categorize very complex visual patterns such as body motions and facial expressions. Applied from human's outstanding ability of perception, the classification function of visual sequences without context information is specially crucial task for computer vision to understand both the coding and the retrieval of spatio-temporal patterns. This paper presents a biological process based action recognition model of computer vision, which is inspired from visual information processing of human brain for action recognition of visual sequences. Proposed model employs the structure of neural fields of bio-inspired visual perception on detecting motion sequences and discriminating visual patterns in human brain. Experimental results show that proposed recognition model takes not only into account several biological properties of visual information processing, but also is tolerant of time-warping. Furthermore, the model allows robust temporal evolution of classification compared to researches of action recognition. Presented model contributes to implement bio-inspired visual processing system such as intelligent robot agent, etc.

• Journal of Internet Computing and Services
• /
• v.21 no.5
• /
• pp.1-7
• /
• 2020

Deep neural networks(DNN), which are used as approximation functions in reinforcement learning (RN), theoretically can be attributed to realistic results. In empirical benchmark works, time difference learning (TD) shows better results than Monte-Carlo learning (MC). However, among some previous works show that MC is better than TD when the reward is very rare or delayed. Also, another recent research shows when the information observed by the agent from the environment is partial on complex control works, it indicates that the MC prediction is superior to the TD-based methods. Most of these environments can be regarded as 5-step Q-learning or 20-step Q-learning, where the experiment continues without long roll-outs for alleviating reduce performance degradation. In other words, for networks with a noise, a representative network that is regardless of the controlled roll-outs, it is better to learn MC, which is robust to noisy rewards than TD, or almost identical to MC. These studies provide a break with that TD is better than MC. These recent research results show that the way combining MC and TD is better than the theoretical one. Therefore, in this study, based on the results shown in previous studies, we attempt to exploit a random balance with a mixture of TD and MC in RL without any complicated formulas by rewards used in those studies do. Compared to the DQN using the MC and TD random mixture and the well-known DQN using only the TD-based learning, we demonstrate that a well-performed TD learning are also granted special favor of the mixture of TD and MC through an experiments in OpenAI Gym.

• The Journal of the Convergence on Culture Technology
• /
• v.8 no.5
• /
• pp.697-703
• /
• 2022

Due to the recent economic downturn caused by Covid-19 and the unstable international situation, many investors are choosing the derivatives market as a means of investment. However, the derivatives market has a greater risk than the stock market, and research on the market of market participants is insufficient. Recently, with the development of artificial intelligence, machine learning has been widely used in the derivatives market. In this paper, reinforcement learning, one of the machine learning techniques, is applied to analyze the scalping technique that trades futures in minutes. The data set consists of 21 attributes using the closing price, moving average line, and Bollinger band indicators of 1 minute and 3 minute data for 6 months by selecting 4 products among futures products traded at trading firm. In the experiment, DNN artificial neural network model and three reinforcement learning algorithms, namely, DQN (Deep Q-Network), A2C (Advantage Actor Critic), and A3C (Asynchronous A2C) were used, and they were trained and verified through learning data set and test data set. For scalping, the agent chooses one of the actions of buying and selling, and the ratio of the portfolio value according to the action result is rewarded. Experiment results show that the energy sector products such as Heating Oil and Crude Oil yield relatively high cumulative returns compared to the index sector products such as Mini Russell 2000 and Hang Seng Index.