Artificial intelligence and inteligent systems

The members of the group carry out research in the following fields: • machine learning and knowledge discovery; • genetic algorithms and search algorithms; • constraint programming and combinatorial optimization; • qualitative reasoning methods; • machine learning in biomedical informatics; • agent technologies and semantic web. The research is constantly motivated with practical uses and concrete applications. Selected scientific achievements for the year 2005: • Analysis and reconstruction of genetic networks on the basis of microarray profiles, published also in van Driessche, Demšar, Juvan, Zupan, et al. Epistasis analysis with global transcriptional phenotypes. Nature Genetics, May 2005, impact factor 24.69. • Intelligent visualization of data using machine learning, published in (Leban et al., VizRank: finding informative data projections in functional genomics by machine learning; and in Curk et al., Microarray data mining with visual programming) Bioinformatics, #1 ranking journal in computer science. • Argument-based machine learning: a new approach to machine learning where the expert in the domain being learned can comment the learning examples with his or her arguments (explaining selected details in the learning examples). • Genetic algorithm for the analysis of complex biological systems with EPR spectroscopy (two papers in leading journals), and a genetic algorithm for steel production optimization introduced into use in steel mills Acroni Jesenice (Slovenia) and Ruuki Steel (Finland). • Machine learning of qualitative models from numerical data with our original method Q2 with application in ecological domains (predicting ozone concentration and Savinja river flooding). • System for automatic discovery of abnormalities in scintigraphic images of skeletons; in collaboration with the Clinic for Nuclear Medicine in Ljubljana. • Contributions to technology of Slovenian language: automatic accentuation of Slovenian words, use of ontologies for interpretation, continuing development of system “GOVOREC”. International awards in 2005: • ECCAI award for the best doctoral dissertation in artificial intelligence in Europe in 2005: A. Jakulin, Machine Learning Based on Attribute Interactions. • First award in NIST (USA) competition of e-mail spam filters (A. Bratko and B. Filipič). Fields of applicative research: • medicine; • bioinformatics and functional genomics; • system identification and management; • environmental issues; • engineering applications (textile industry, mechanical engineering); • internet applications (e-mail spam filtering, intelligent browsers); • marketing and economy.

This research programme made contributions to several problems of current interest in the area of Artificial Intelligence. These contributions are important because of the impact and high applicability of this area of Computer Science. In our assessment, the results of this programme will significantly contribute to the development of the addressed topics of AI, development of practical programming tools, and confirmation of the applicability in practice of the scientific innovations. A number of results of this program are important for the development of machine learning in AI, both in terms of new approaches and algorithms (e.g. argument-based machine learning, qualitative learning, Q2 learning, methods of intelligent information visualisation), and powerful application tools (such as Orange system). Some of the application-oriented research of this program also contributes to other sciences, concretely for example to genetic research. Methods developed in this programme enable automatic construction of new scientific theories from experimental data (for example from observations from genetic experiments). Some scientific results of this programme are also significant for the progress in the theory of search methods in AI. This includes the analysis of “pathology” in search algorithms, as well as the paradox discovered in this programme, that in “real-time search” pessimistic heuristics are more successful, in contrast to the previously known benefits of optimistic heuristics in search algorithms in general. The question of what intelligence is and how it comes about is a key and yet unanswered question of science and philosophy. Some results of this programme also contribute to this question, and are, among others, relevant to the philosophical question of (im)possibility of automation of research, in particular of automatically constructing new scientific hypotheses. Our results indicate a positive answer to this question. The development of new methods in this programme is not only important for the progress in computer science. Approaches of AI, like machine learning, knowledge discovery in data, representation and use of domain knowledge, qualitative modelling and design of intelligent systems, are becoming, or have become, key part of the infrastructure of other sciences. Examples of recent such advances are in functional genomics (application of heuristic search methods) and modern approaches to clinical diagnosis (prediction models based on gene expression).

We strive to apply the findings of our research in practice as much as possible. Some examples in recent years: quality control in light construction sandwich panel production at Trimo, predicting the appearance of clothing on the basis of physical and mechanical properties of textiles at Mura (in collaboration with Textile Department at University of Maribor), the system for speech synthesis for Slovenian language GOVOREC, predicting the danger of flooding by Slovenian rivers. System GOVOREC enables computer speech synthesis of Slovenian texts. Most likelz, its biggest contribution is helping handicapped people in their everyday life. With the use of qualitative modelling we constructed a predictive model for the level of Savinja river. Since Savinja is a torrentous river its level can rapidly change. Our model can serve as an early warning system against flooding. Apart from scientific work, the members of the group teach various courses (mostly AI and programming) at the University of Ljubljana and to a lesser extent at the University of Maribor, University of Nova Gorica and the Jožef Stefan International Postgraduate School. The textbook Prolog Programming for Artificial Intelligence by Ivan Bratko, the head of the group, has been translated into a number of languages, and is the most widely used textbook on prolog programming language in the world. We are continuously developing an open-source machine learning suite Orange (www.ailab.si/orange), among the largest of its kind in the world. The suite boasts a graphical user interface that enables visual programming which in turn enables even non-programmers (like experts from other fields, e.g. biologists or geneticists) to create complex procedures from Orange’s numerous basic building blocks. In collaboration with biologists and geneticists we developed Orange modules for processing and analysis of bioinformatics and functional genomics data. The most high-profile result of this work is the finding that it is possible to reason about relations between genes on the basis of transcriptional profiles of mutants (Van Driessche et al., Epistasis analysis with global transcriptional phenotypes, Nature Genetics 37(5): 471-477, 2005). In the last five years the members of the group have earned two prestigious awards for the best European doctoral dissertation in artificial intelligence awarded by ECCAI, the European Coordinating Committee for Artificial Intelligence.