• Electronics and Telecommunications Trends
• /
• v.37 no.6
• /
• pp.32-42
• /
• 2022

Recently, decentralized approaches to artificial intelligence (AI) development, such as federated learning are drawing attention as AI development's cost and time inefficiency increase due to explosive data growth and rapid environmental changes. Collaborative AI technology that dynamically organizes collaborative groups between different agents to share data, knowledge, and experience and uses distributed resources to derive enhanced knowledge and analysis models through collaborative learning to solve given problems is an alternative to centralized AI. This article investigates and analyzes recent technologies and applications applicable to the research of multi-agent collaboration of AI bots, which can provide collaborative AI functionality autonomously.

• ETRI Journal
• /
• v.45 no.5
• /
• pp.862-873
• /
• 2023

Determining whether an autonomous self-driving agent is in the middle of an intersection can be extremely difficult when relying on visual input taken from a single camera. In such a problem setting, a wider range of views is essential, which drives us to use three cameras positioned in the front, left, and right of an agent for better intersection recognition. However, collecting adequate training data with three cameras poses several practical difficulties; hence, we propose using data collected from one camera to train a three-camera model, which would enable us to more easily compile a variety of training data to endow our model with improved generalizability. In this work, we provide three separate fusion methods (feature, early, and late) of combining the information from three cameras. Extensive pedestrian-view intersection classification experiments show that our feature fusion model provides an area under the curve and F1-score of 82.00 and 46.48, respectively, which considerably outperforms contemporary three- and one-camera models.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.20 no.3
• /
• pp.71-84
• /
• 2024

Artificial intelligence can be cited as a key linkage technology for expanding drones' application fields, and drones combined with artificial intelligence are expected to improve drones' operational capabilities based on algorithms that can solve complex tasks through learning. The purpose of this study is to analyze various latest research cases that apply deep reinforcement learning to drones to solve limitations for performing swarm flight and to propose a new research direction that applies them to multi-agent communication optimization technology. The process of the research is to investigate and analyze the methods for efficient operation of control and communication technologies required for swarm flight to be successful, and to apply algorithms that have the advantage of exchanging richer feedback between agents and having less learning than conventional methods when learning deep reinforcement learning algorithms. It is expected that the efficiency and performance of learning communication protocols optimized for swarm flight will be improved, which will increase the efficiency of mission performance when exploring or scouting large areas through swarm flight in the future.

• Proceedings of the IEEK Conference
• /
• 2002.07b
• /
• pp.1173-1176
• /
• 2002

Interaction builds up among agents in order to accomplish their goals. Argumentation is important for agent negotiation and interaction. In this paper, we discuss about the framework for multi-agent argumentation and the way multi-agents co-operates between each other. We identify aspects of classical argumentation theory that are suitable and useful for artificial agents.

• Journal of Korea Multimedia Society
• /
• v.13 no.12
• /
• pp.1820-1831
• /
• 2010

When interacting with unknown environments, an autonomous agent needs to decide which action or action order can result in a good state and determine the transition probability based on the current state and the action taken. The traditional multiple sequential learning model requires predefined probability of the states' transition. This paper proposes a multiple sequential learning and prediction system with definition of autonomous states to enhance the automatic performance of existing AI algorithms. In sequence learning process, the sensed states are classified into several group by a set of proposed motivation filters to reduce the learning computation. In prediction process, the learning agent makes a decision based on the estimation of each state's cost to get a high payoff from the given environment. The proposed learning and prediction algorithms heightens the automatic planning of the autonomous agent for interacting with the dynamic unknown environment. This model was tested in a virtual library.

• The Journal of Society for e-Business Studies
• /
• v.5 no.2
• /
• pp.49-60
• /
• 2000

A major topic in the field of network and telecommunications is doing business on the Word Wide Web(WWW), which is called Electronic Commerce(EC). Another major topic is blending Artificial Intelligent techniques with the WWW. To provide customer with the information of goods in suit with a customer liking, we propose multi agent system which is consist of customer agent and search agent etc. Also we use case-based reasoning for customer liking searching the information of goods and training through the reuse. This reuse make efficient management of information and a process of operation. In the relation between customer and goods, if there are some goods which is not search from case-base reasoning, we calculate satisfaction function for customer purchase goods. And to provide customer with the information of goods in the first of satisfaction function, This EC system can always provide the information of goods which is satisfied to customer.

• 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.

• International Journal of Internet, Broadcasting and Communication
• /
• v.14 no.2
• /
• pp.192-198
• /
• 2022

Companies are building and utilizing their own data analysis systems according to business characteristics in the distributed cloud. However, as businesses and data types become more complex and diverse, the demand for more efficient analytics has increased. In response to these demands, in this paper, we propose an unsupervised learning-based data analysis agent to which reinforcement learning is applied for effective data analysis. The proposal agent consists of reinforcement learning processing manager and unsupervised learning manager modules. These two modules configure an agent with k-means clustering on multiple nodes and then perform distributed training on multiple data sets. This enables data analysis in a relatively short time compared to conventional systems that perform analysis of large-scale data in one batch.

• Journal of Animal Science and Technology
• /
• v.48 no.3
• /
• pp.467-478
• /
• 2006

Maintenance of an optimal air quality in the enclosed pig building is potentially important in terms of pig performance and farmer health. The objective of this on-site experiment is to evaluate and compare efficiencies of currently utilized biological additives to reduce odor emissions from the enclosed pig building. As a result, generally all the additives except for salt water, artificial spice and essential oil were proved ineffective in reducing odor generation. The beneficial effects of salt water, artificial spice and essential oil on odor reduction were highlighted on ammonia, odor intensity and offensiveness, and sulfuric odorous compounds, respectively. To efficiently utilize odor masking agent such as the artificial spice, ventilation rate should keep slightly lower than the optimal level. Essential oil functioned well as not only masking agent but also antimicrobial agent for reducing odor. To precisely quantify odor concentration, it should be measured by not the odor sensor but the olfactometry technique.

• Journal of the Korean Ceramic Society
• /
• v.54 no.6
• /
• pp.492-498
• /
• 2017

A thermal barrier coating (TBC) with self-healing property for cracks was proposed to improve reliability during gas turbine operation, including structural design. Effect of healing agent on crack propagation behavior in TBCs with and without buffer layer was investigated through furnace cyclic test (FCT). Molybdenum disilicide ($MoSi_2$) was used as the healing agent; it was encapsulated using a mixture of tetraethyl orthosilicate and sodium methoxide. Buffer layers with composition ratios of 90 : 10 and 80 : 20 wt%, using yttria stabilized zirconia and $MoSi_2$, respectively, were prepared by air plasma spray process. After generating artificial cracks in TBC samples by using Vickers indentation, FCTs were conducted at $1100^{\circ}C$ for a dwell time of 40 min., followed by natural air cooling for 20 min. at room temperature. The cracks were healed in the buffer layer with the healing agent of $MoSi_2$, and it was found that the thermal reliability of TBC can be enhanced by introducing the buffer layer with healing agent in the top coat.