Kwang-Hyun Shim

Soft Geodesic Kernel K-Means

In this paper we present a kernel method for data clustering, where the soft k-means is carried out in a feature space, instead of input data space, leading to soft kernel k-means. We also incorporate a geodesic kernel into the soft kernel k-means, in order to take the data manifold structure into account. The method is referred to as soft geodesic kernel k-means. In contrast to k-means, our method is able to identify clusters that are not linearly separable. In addition, soft responsibilities as well as geodesic kernel, improve the clustering performance, compared to kernel k-means. Numerical experiments with toy data sets and real-world data sets (UCI and document clustering), confirm the useful behavior of the proposed method.

Vision Based Motion Tracking System for Interactive Entertainment Applications

Although the vision based motion recognition plays an important role in virtual reality and interactive entertainment area, it has not yet found wide usage. The first main reason is the unavailability of relatively low cost wireless tracking system that would operate well enough for inputting gestures. The second reason lies in the lack of naturalness in gesture design. That is, so far most gestures used are either static images or poses and too abstract losing its intended meaning and affordances, resulting in low usability and presence. In this paper, we propose architecture for a low cost re-configurable vision-based motion tracking system and motion recognition algorithm. The user wears one or more retro-reflective markers for achieving more exact tracking performance and make predefined motion gestures. Through object segmentation processing, 3D positions of the objects are computed. And then, a motion gesture corresponding on these 3D position trajectories is found by a simple correlation-based matching algorithm. Also, we demonstrate this system by applying it to virtual environment navigation and interactive entertainments.

A dynamic load balancing for massive multiplayer online game server

On-line games are becoming more popular lately as the Internet becomes popular, game platforms become diverse and a ubiquitous game environment is supported. Therefore, distributed game server technology is required to support large numbers of concurrent game users simultaneously. Especially, while game users are playing games, many unpredictable problems can arise, such as a certain server handles more server loads than recommended because many game users crowded into a specific region of a game world. These kinds of situations can lead to whole game server instability. In this paper, global dynamic load balancing model and distributed MMOG(Massive Multiplayer Online Game) server architecture are proposed to apply our load balancing algorithm. Many different experiments were carried out to test for efficiency. Also an example of applying real MMOG application to our research work is shown.

Reinforcement learning of intelligent characters in fighting action games

In this paper, we investigate reinforcement learning (RL) of intelligent characters, based on neural network technology, for fighting action games. RL can be either on-policy or off-policy. We apply both schemes to tabula rasa learning and adaptation. The experimental results show that (1) in tabula rasa leaning, off-policy RL outperforms on-policy RL, but (2) in adaptation, on-policy RL outperforms off-policy RL.

Adaptation of intelligent characters to changes of game environments

This paper addresses how intelligent characters, having learning capability based on the neural network technology, automatically adapt to environmental changes in computer games. Our adaptation solution includes an autonomous adaptation scheme and a cooperative adaptation scheme. With the autonomous adaptation scheme, each intelligent character steadily assesses changes of its game environment while taking into consideration recently earned scores, and initiates a new learning process when a change is detected. Intelligent characters may confront various opponents in many computer games. When each intelligent character has fought with just part of the opponents, the cooperative adaptation scheme, based on a genetic algorithm, creates new intelligent characters by composing their partial knowledge of the existing intelligent characters. The experimental results show that intelligent characters can properly accommodate to the changes with the proposed schemes.