Advances in Artificial Intelligence and Applications
Fundamenta Informaticae · 2009
All around the world, every year, scientists and practitioners discuss at numerous conferences dedicated to Artificial Intelligence the novel achievements and challenges.Such conferences are organized also in Poland, one of them is the International Symposium Advances in Artificial Intelligence and Applications (AAIA), organized annually within the International Multiconference on Computer Science and Information Technology framework.Selected participants of that Symposium have been invited to submit their papers to this issue of Fundamenta Informaticae.Two papers have been written especially to this issue of Fundamenta journal, three other were presented at the Symposium and are considerably extended.The first paper, entitled "Measuring Semantic Closeness of Ontologically Demarcated Resources" is written by two teams: from Poland and Korea.Fusion of their experiences has allowed to develop an agent-based system, aimed to support workers in an organization.One of key functionalities of this system is ontological matchmaking, understood as a way of establishing closeness between resources.The system recommends which, among available resources, are relevant / of interest to the worker.Authors approach to measuring semantic closeness between ontologically demarcated information objects is discussed in the paper.A Duty Trip Support application is used as a case study.It is worth mentioning that, in computer science, ontology is a representation of a set of concepts and relationships between them within a given domain.The second paper, "Designing Model Based Classifiers by Emphasizing Soft Targets", deals with training classifiers.Classification task is very popular in real-life problems.A number of different classification methods exists, each reveals some advantages and weakness.The authors explore the effectiveness of using emphasized soft targets with generative models, such as Gaussian Mixture Models (GMM), and Gaussian Processes (GP).Their approach seems to produce better performance and is less sensitive for parameters values.The third paper, Improved Resulted Word Counts Optimizer for Automatic Image Annotation Problem, uses classifiers (a family of classifiers) for automatic annotation of images.It is an important research topic in pattern recognition area.In the paper, a generic approach to find correct word frequencies is proposed.The Optimizer can be used with different automatic image annotators, based on various machine learning paradigms.In the paper, a new, improved authors former method, Greedy Resulted Word Counts Optimization is proposed.The proposed method is more intuitive and it reduces