Programme of colloquia with abstracts for the Spring 2005 semester

1 March 2005: Prof. RNDr. Jan Kratochvíl DrSc., Faculty of Mathematics and Physics, Charles University, Prague; The frequency assignment problem: from mobile phones to graph isomorphism
8 March 2005: Prof. Renate Motschnig, University of Vienna; Soft skills for IT professionals: initial experience in combining theory and practice; Abstract: Whilst designing the ‘Computer Science and Business Informatics’ curriculum at the University of Vienna, we asked IT professionals the question: ‘What skills does industry require of IT graduates?’ Social skills topped the list, followed by the ability to work in teams and think in abstract terms.
This presentation focuses on our theory-guided, practical approach to developing soft skills within academic courses on project management, web engineering and person-centred communication. We find it particularly challenging and relevant to address social skills using humanistic educational principles, whilst combining face-to-face sessions with e-learning elements. We support e-learning with open-source CEWebS (Cooperative Environment Web Services) modules developed at our department. We will share selected course scenarios, practical experiences and evaluation results. Finally, we discuss initial findings on the conditions under which students’ motivation is increased and they report benefiting more than in conventional courses.
15 March 2005: Prof. M. Dokulil, Doc. I. Kopeček, Dr. J. Thomas, Faculty of Informatics, Masaryk University, Brno; Bach, Gödel and Informatics (Dedicated to the 320th anniversary of Bach’s birth and to the 70th anniversary of Gödel’s papers on the consistency of the axiom of choice.); Abstract: What is it about Bach’s music that so fascinates and enchants so many mathematicians and computer scientists? What are the parallels between Bach’s music and Gödel’s work, and between Baroque music and modern-day informatics? There are many such questions, which are both controversial and provocative, whilst at the same time being inspiring and fascinating. This presentation will touch upon some of these points and attempt to illustrate them, using the languages of both music and informatics.
22 March 2005: Doc. Luděk Matyska CSc., Dr Eva Hladká PhD., UVT+FI, Masaryk University, Brno; Large data volumes, distance and space – a challenge for distributed environments; Abstract: Collaborative environments are becoming increasingly integrated with large-scale distributed systems (Grids). In these environments, a person acts as an end node, where sensory interaction is mediated through devices that digitise audio, video and, in some cases, other (e.g. haptic) signals. These digital data are further processed, transmitted over a network and, at the other end (or ends in the case of multipoint collaboration), converted back into sensory signals. The data could also be combined with output from programmes that directly generate the digital equivalent of sensory signals. The continuous improvement in the quality of capture and projection systems leads to a substantial increase in the volume of data that must be generated, processed, transmitted and ultimately replicated within the network. The classical Internet protocols, which are widely deployed and studied, were designed for speeds in the range of tens of Mbps, whilst the actual data streams in high-end collaboration environments are at the Gbps level (for example, High Definition TV could generate a stream of around 1.3 Gbps in one direction). The fundamental challenge lies in capture (or generation), (multicast) transport, minimising latency over long distances, and converting the data back to analogue (sensory) signals at the receiving nodes, whilst maintaining the highest quality of service. This has led to research and development in optical networks for high-speed data transfer, overlay networks to provide multicast and internal network processing capabilities, and encoding and decoding protocols – part of this research is being carried out at FI in collaboration with the Institute of Computer Science and CESNET. The presentation will conclude with a discussion of the activities required to support an international demonstration of a multicast, low-latency HDTV-based collaborative environment, spanning Brno, Baton Rouge and San Diego.
29 March 2005: Doc. Michal Kozubek, PhD., Faculty of Informatics, Masaryk University, Brno; Processing of new types of image data in optical microscopy; Abstract: New trends in molecular biology are leading to new methods of acquiring image data in optical microscopy. This is giving rise to new types of image data which must be handled by computer image analysis. Typical examples include DNA microarrays (monitoring the activity of thousands of genes simultaneously on a single slide) and in vivo imaging (observing the behaviour of living cells and their components over time). These two types of study currently dominate the field worldwide. The Laboratory of Optical Microscopy at the Faculty of Information Studies, Masaryk University, is also intensively engaged in these two areas, from the perspectives of molecular biology, optics and computer science. The lecture will focus on the computer science aspect.
5 April 2005: Doc. Pavol Voda, CSc., FMFI, Comenius University, Bratislava; Specification and Verification of Declarative Programs; Abstract: We will present and demonstrate a functional programming language and proof system called CL (Clausal Language), which treats programmes as functions over natural numbers. The representation of the usual data structures used in computer programming as natural numbers is quite simple and very natural. CL is similar in spirit to LISP, but it is a thoroughly modern programming language. CL is implemented as a server that communicates with the user via a web browser (we use Mozilla). Programmes and their properties are presented in a highly readable mathematical form using MathML, an extension of XML and HTML. The representation of data structures as natural numbers allows the use of the basic logical theory of arithmetic – known as Peano arithmetic – as the formal theory within which programmes written in CL are specified and verified to be correct. CL is used at the University of Bratislava in three courses on declarative programming, mathematical logic, and the specification and verification of programmes.
12 April 2005: Prof. Ing. Bořivoj Melichar, DrSc., Faculty of Electrical Engineering, Czech Technical University in Prague; Repetitions in Text and Finite Automata; Abstract: This presentation outlines a general method for finding repetitions in text. It begins with a classification of repetitions. It then presents a general method for finding exact repetitions in a single string and in a finite set of strings. This utilises a method of analysing d-subsets arising from the determinisation of a non-deterministic factor automaton. These d-subsets contain all the information relating to factor repetitions. The principle of analysing these subsets is then applied to find approximate repetitions for various distances in general and ordered alphabets. The possibility of finding repetitions in the case of ‘don’t care’ symbols is also mentioned.
19 April 2005: Doc. Ivan Kopeček, CSc., Faculty of Informatics, Masaryk University, Brno; Dialogue Modelling; Abstract: Natural language dialogue is a highly complex communicative process of great importance, not only with regard to human–computer interaction, but also to other fields such as collaboration theory, speech recognition and synthesis, user modelling, computational psychology and many others. To understand and manage it, abstract models are often used, enabling the situation to be simplified and allowing the application of powerful general theories, including the theory of formal languages and automata, universal algebra and game theory. In this presentation, some of the formal approaches will be outlined with the aim of demonstrating the beauty and variety of the related problems. An application designed for the synthesis of a dialogue system will also be presented.
26 April 2005: Doc. Luboš Brim, CSc., Faculty of Informatics, Masaryk University, Brno; Parallel and Distributed Verification Methods; Abstract: Automated verification is a computationally intensive process. From a practical point of view, therefore, the availability of sufficient computing resources is very often a limiting factor. This lecture will provide an overview of current developments in the use of the combined power of parallel computers to solve difficult verification problems. The techniques and algorithms used will be discussed, and the results achieved will be evaluated. We shall also attempt to outline possible directions in which research in this field might proceed.
3 May 2005: Doc. RNDr. Roman Ďurikovič, Ph.D., Comenius University, Bratislava; Soap dynamics: bubble formation, clustering, collision and popping; Abstract: What happens when a soap bubble floats in the air? How do bubbles coalesce to form beautiful three-dimensional clusters? The physics-based model and animation described here provide the answers. We present a comprehensive computer simulation of soap bubbles from a dynamic perspective, which should be of great interest to physicists and mathematicians. We discuss the dynamic formation of irregular bubble clusters and how to animate bubbles. The resulting model takes into account surface tension, film elasticity and shape variations caused by gravity, external wind and stress forces.
10 May 2005: Prof. Edmund Burke, University of Nottingham, UK; New Heuristic and Meta-heuristic Approaches in Cutting and Packing; Abstract: In this talk, I will provide an overview of our recent research on cutting and packing, which has led to the establishment of a new spin-off company. I shall outline some of the new heuristic approaches that we have developed for rectangular and irregular packing. I will demonstrate the strength of these approaches, both in terms of the improved quality of the solutions produced and enhancements to the representation of solutions. Using a large set of test data taken from the literature, we have compared the new approaches to previously published methods from the cutting and packing research community over the last 20 years. We have outperformed all published benchmarks in the literature for both rectangular and irregular problems, achieving an average improvement of 5 per cent on each, alongside shorter run times. I will also provide an overview of the extensions to this work being carried out through new projects and outline new opportunities that have arisen.
17 May 2005: Prof. Vilém Novák, DrSc., University of Ostrava; On the Principles and Potential of Fuzzy Logic; Abstract: Fuzzy logic can be characterised as a specialised mathematical discipline whose aim is to provide tools for modelling the phenomenon of vagueness and for modelling the way humans reason, which relies heavily on natural language. We distinguish between fuzzy logic in the narrower sense (FLn) and the broader sense (FLb). In the narrower sense, it refers to the entire system of multi-valued formal logics. FLb is an extension of FLn to include certain further aspects: a model of the semantics of a subset of natural language (in particular, so-called evaluative linguistic expressions), the interpretation of so-called fuzzy IF-THEN rules, and their application both in a model of human reasoning and for the approximation of functions. The lecture will include a demonstration of the LFLC 2000 software, which implements the aforementioned theoretical results, and will discuss the possibilities for its practical application.