Olga Stepankova

Multi-agent systems: which research for which applications

For sometime now agent-based and multi-agent systems (MASs) have attracted the interest of researchers far beyond traditional computer science and artificial intelligence (AI). In this article we try to identify focal points of interest for researchers working in the area of distributed AI (DAI) and MAS as well as application-oriented researchers coming from related disciplines, e.g. electrical and mechanical engineering. We do this by presenting key research topics in DAI and MAS research and by identifying application domains in which the DAI and MAS technologies are most suitable. The research topics we discuss are separated into agent architectures and organisations, negotiation among agents, and self-adaptation of MAS using learning techniques. Regarding the application domains for these techniques we distinguish the application domains according to whether the agents control a physical or virtual body (Gestalt) or not. This separation of the application domains is not strict; it represents two ends of a continuum. On the one end of this continuum we have autonomous robot systems which act in a physical environment (sometimes referred to as hardware agents), and on the other end, we have abstract environments, such as in workflow systems, which rarely display the geometrical and physical aspects of the environment we are used to living in.

Sequential Data Mining: A Comparative Case Study in Development of Atherosclerosis Risk Factors

Sequential data represent an important source of potentially new medical knowledge. However, this type of data is rarely provided in a format suitable for immediate application of conventional mining algorithms. This paper summarizes and compares three different sequential mining approaches based, respectively, on windowing, episode rules, and inductive logic programming. Windowing is one of the essential methods of data preprocessing. Episode rules represent general sequential mining, while inductive logic programming extracts first-order features whose structure is determined by background knowledge. The three approaches are demonstrated and evaluated in terms of a case study STULONG. It is a longitudinal preventive study of atherosclerosis where the data consist of a series of long-term observations recording the development of risk factors and associated conditions. The intention is to identify frequent sequential/temporal patterns. Possible relations between the patterns and an onset of any of the observed cardiovascular diseases are also studied.

Role of Acquaintance Models in Agent-Based Production Planning System

This paper comments the role of acquaintance models in agent-based engineering solutions. We present a specific methodology,a tri-base acquaintance model, as formal model of agents’ mutual awareness. The model contains three separate knowledge structures for representing agents’ permanent, semi-permanent and temporary knowledge, respectively, and mechanism for administering, maintenance and exploration of the knowledge. The paper explains how utilisation of an acquaintance model contributes to communication savings and to reduction of overall distributed problem solving complexity. Utilisation of the tri-base acquaintance model is illustrated on ProPlanT multi-agent system for project-oriented production planning. The system architecture exploits several differenttypes of agents exploring the tri-base mechanism including the meta-agents who are used to adjust and tune the agents’ acquaintance models.

Tissue Specific Electrochemical Fingerprinting

Background Proteomics and metalloproteomics are rapidly developing interdisciplinary fields providing enormous amounts of data to be classified, evaluated and interpreted. Approaches offered by bioinformatics and also by biostatistical data analysis and treatment are therefore of extreme interest. Numerous methods are now available as commercial or open source tools for data processing and modelling ready to support the analysis of various datasets. The analysis of scientific data remains a big challenge, because each new task sets its specific requirements and constraints that call for the design of a targeted data pre-processing approach. Methodology/Principal Findings This study proposes a mathematical approach for evaluating and classifying datasets obtained by electrochemical analysis of metallothionein in rat 9 tissues (brain, heart, kidney, eye, spleen, gonad, blood, liver and femoral muscle). Tissue extracts were heated and then analysed using the differential pulse voltammetry Brdicka reaction. The voltammograms were subsequently processed. Classification models were designed making separate use of two groups of attributes, namely attributes describing local extremes, and derived attributes resulting from the level = 5 wavelet transform. Conclusions/Significance On the basis of our results, we were able to construct a decision tree that makes it possible to distinguish among electrochemical analysis data resulting from measurements of all the considered tissues. In other words, we found a way to classify an unknown rat tissue based on electrochemical analysis of the metallothionein in this tissue.

Maintenance of Discovered Knowledge

The paper addresses the well-known bottleneck of knowledge based system design and implementation – the issue of knowledge maintenance and knowledge evolution throughout lifecycle of the system. Different machine learning methodologies can support necessary knowledge-base revision. This process has to be studied along two independent dimensions. The first one is concerned with complexity of the revision process itself, while the second one evaluates the quality of decision-making corresponding to the revised knowledge base. The presented case study is an attempt to analyse the relevant questions for a specific problem of industrial configuration of TV transmitters. Inductive Logic Programming (ILP) and Explanation Based Generalisation (EBG) within the Decision Planning (DP) knowledge representation methodology, have been studied, compared, and tested on this example.

A Learning System for Decision Support in Telecommunications

We present a system for decision support in telecommunications. History data describing the operation of a telephone exchange are analyzed by the system to reconstruct understandable event descriptions. The event descriptions are processed by an algorithm inducing rules describing regularities in the events. The rules can be used as decision support rules (for the exchange operator) or directly to automate the operation of the exchange.

Use of brightness wavelet transformation for automated analysis of serum metallothioneins- and zinc-containing proteins by Western blots to subclassify childhood solid tumours.

In this study, we determined serum levels of metallothioneins (MTs) and zinc in children with solid tumours (neuroblastoma, Hodgkin lymphoma, medulloblastoma, osteosarcoma, Ewing sarcoma and nephroblastoma) by differential pulse voltammetry Brdicka reaction and ELISA. Zn(II) level in patients sera was 40% compared to controls, contrariwise, MT level was 4.2 × higher in patients. No significant differences among single diagnoses were found both for Zn(II) and MT. When determined Zn(II)/MT ratio, in controls its value was 24.6, but it was 2.6 in patients. After Western‐blotting with anti‐MT and anti‐Zn chicken antibodies, variable intensities of the bands within the samples were observed. The brightness curve obtained for each sample both for MT‐ and Zn blots was further analysed to produce a list of band positions together with some complementary information related to the intensity of the observed bands by the optimised algorithm. We constructed from those profiles decision trees that enable to distinguish different groups of tumours. The blood samples were heat‐treated, in which we supposed mainly MT, but samples contained other thermostable Zn‐containing proteins that were helpful for identification of embryonal tumours with 88% accuracy and for identification of sarcomas with 78% accuracy. In MT blots the accuracies were 53 and 45%, respectively. Simultaneous analysis of MT and Zn blots did not increased accuracy of identification neither in embryonal tumours (80%) nor in sarcomas. Those results are promising not only from diagnostic point of view but particularly in the area of studying of individual MT isoforms and their aggregates in malignant tumours.

Diabetes management in OLDES project

EU project OLDES (Older People’s e-services at home) develops easy to use and low cost ICT platform in order to offer a better quality of life to elderly people directly in their homes through innovative systems of tele-accompany, tele-assistance and tele-medicine. The elderly are able to access the services and send relevant medical data from their home by being connected to the central server via a low cost PC which is based on Negroponte paradigm. The OLDES platform interface uses television screens controlled through a remote control customized for the elderly. The feasibility of OLDES project is evaluated by the pilot study concentrating on compensation of diabetic patients. Compensation of diabetes is achieved by monitoring glucose glycemia level, blood pressure and weight. Moreover, the patient feeds into OLDES system daily consumption of food using interactive food scales and obtains advice if necessary.

On Updating the Data Warehouse from Multiple Data Sources

The paper discusses various problems related to feeding the data warehouse with data originated in various data sources. The main focus is concentrated on describing and handling of uncertainty, which comes from merging imperfect data sources. Uncertainty measures and their implementation approaches are suggested.

An Introduction to Qualitative Reasoning

The first section tries to indicate the reasons for study of qualitative reasoning within the field of Artificial Intelligence. The second section solves an example from physics as an illustartion of usefulness of qualitative arguments. The third section briefly reviews formal systems for qualitative reasoing. The last section offers some suggestions for further reading.