Special Issue on Advanced Intelligent Systems
Myung-Geun Chun and Toshihiko Watanabe
The 12th International Symposium on Advanced Intelligent Systems (ISIS 2011) held at La Vie DfOr Resort, Suwon, Korea, on September 28 to October 1, 2011, featured presentations by researchers, engineers and practitioners on the latest accomplishments, innovations and applications in artificial intelligence, intelligent systems, and information technology. The 152 papers consisted of 110 regular papers, 39 organized session papers and 3 invited papers were contributed to the conference. The Program Committee requested ISIS 2011 reviewers to select papers for a special issue of the Journal of Advanced Computational Intelligence & Intelligent Informatics (JACIII), of which about 10 were accepted for publication in a twopart ISIS 2011 special issue, Vol.16, No.7, containing 5 papers. Part II will feature about 5 papers.
A brief review of Part I covers 5 papers:
First paper proposes a quantitative model for assessing the collision risk in maritime waterway traffic. The method proposes recent maritime traffic characteristics in time-variant CPA waterway environments and models a dynamic causation factor as a risk indicator. Second paper proposes multiagent query refinement realizing personalized query refinement by three strategies ? knowledge-based query expansion, user-device-based query and weighted query expansion strategy. These approaches determine the domain that the initial query belongs to and expand the query by comprehensively considering user interests. Third paper presents variable-step-size incremental conductance direct maximum power point tracking using fuzzy membership for a standalone photovoltaic system under rapidly changing irradiation. Fourth paper applies a natural actor-critic (NAC) and natural evolution strategies (NES) from natural-gradient-based machine learning to path-tracking control problems for autonomous vehicles. Fifth paper presents features for authenticating painting style, specifically that of Piet Mondrain. It demonstrates meaningful features using two supervised learning algorithms ? decision tree induction algorithm C4.5 and the feature generating machine (FGM) ? selecting important features in the course of learning.
We thank the reviewers for their time and effort in making this special issue available so quickly. We are also grateful to the JACIII editorial office for invaluable assistance and advice in putting the issue together.