AGENTS '00 Proceedings of the fourth international conference on Autonomous agents
IJCNN '00 Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks (IJCNN'00)-Volume 2 - Volume 2
Information Filter for Ambiguous Information Retrieval
WI-IAT '08 Proceedings of the 2008 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Volume 03
Search in the mood: the information filter based on ambiguous queries
International Journal of Computer Applications in Technology
Hi-index | 0.00 |
In this paper, we propose a 3-D virtual space creation system reflecting user's emotion by interactive evolutionary method. In this system, a virtual space is created using Open GL. And users can walk through in the virtual space. Five kinds of emotions, happiness, sadness, anger, fear, and disgust, are treated. They are selected based on researches on emotion. Researches which create space corresponding to user's favorite emotions are difficult because techniques reflecting emotions in a 3-D space have not been established. Our system evolves space interactively using genetic algorithm (GA) to reflect user's emotion. This GA uses user's evaluation as fitness. This makes the followings possible: (1) direct reflection of user's emotion: (2) making the system simple because calculation of fitness function is not necessary. In GA, initial genes are very important. The proposed system determines initial genes using Kohonen's Self-Organizing Feature Map (SOFM). SOFM is used to interpolate a high dimensional parameter space by learning. The SOFM in the system has been learned by a preliminary experiments: people give emotions to many sample spaces. It has been confirmed by experiments that the proposed system can create spaces having high evaluation.