Program seminářů pro rok 2002/2003

Podzim 2002

26. 9. 2002

Úvodní seminář podzimního semestru

Program:
Informace o koncepci semináře v podzimním semestru.
Domluva programu semináře.
Diskuse.

3. 10. 2002

L. Bártek

Vývoj dialogového rozhraní pro knihovní systém

Abstrakt
V Laboratoři řeči a dialogu je vyvíjeno dialogové rozhraní pro přístup ke knihovním systémům. Aktuálně používaná dialogová strategie implementovaná na bázi technologie VoiceXML nutí v některých situacích uživatele odpovídat na řadu otázek, z nichž některé mohou být vzhledem k jeho požadavku nerelevantní. Přednáška se bude věnovat jednak řešením tohoto problému alternativními strategiemi a dále pak použitým metodám.

10. 10. 2002

Setkání a diskuse s oborovou radou FI MU

Kromě doktorandských studetnů jsou zváni i všichni další zájemci o studentskou vědeckou činnost. Využijte možnosti získat informace, formulovat své názory, problémy, otázky, a zůčastnit se otevřené a neformální diskuse.

17. 10. 2002

Prof. J. Gruska

How to get famous

Abstract
Some thoughts and examples how one can get easily famous (under certain conditions).

24. 10. 2002

P. Švéda

Bezpečnostní aspekty využívání elektronického podpisu

Abstract
Elektronický podpis má zakotvený legislativní rámec (zákonem i provozní vyhláškou) avšak v souvislosti s praktickým využitím a masovým rozšířením vyvstává řada otevřených, opomíjených a dosud neřešených otázek. Podpis je vytvářen počítačem, avšak to zdaleka není jediný kámen úrazu. Přednáška bude diskutovat bezpečnostní aspekty využívání elektronického podpisu v praxi a poukáže na možné přístupy řešení jednotlivých izolovaných problémů.

31. 10. 2002

P. Adámek

Vyhledávání v prostředí XML dokumentů

Abstrakt
TBA

7. 11. 2002

M. Povolný

Electronic dictionaries

Abstract
Electronic dictionaries and lexical databases are an important part of the NLP infrastructure. In the Natural Language Processing laboratory we are working on a set of tools for preparation, querying and presentation of electronic dictionaries and encyclopaedia intended for common use by an unskilled user as well as on preparation of databases of semantic relations for other NLP tools. I will present the basic tools and concepts that we use and also some plans for the near future.

14. 11. 2002

J. Bouda

Encryption of quantum information

Abstrakt:
The quantum cryptography is a fast developing branch of quantum information processing. The most fundamental results of quantum cryptography include quantum distribution of key, quantum secret sharing, quantum oblivious transfer and other cryptographic protocols. The fast evolution of quantum cryptography enforced the evolution of the encryption of quantum information defined in terms of the so-called quantum private channel, in which the plaintext is quantum information, the transmission channel is quantum and the key is either quantum or classical. As a consequence, we don't have to care about secure distribution of quantum information, but we can concentrate our attention on the protocol itself.

P. Tišnovský
VUT Brno, FIT

Modelování a vizualizace elastických těles pomocí hybridních částicových systémů

Abstract:
V této prezentaci budou podány informace o vytváření modelů těles pomocí tzv. hybridních částicových systémů. Hybridní částicový systém, jež v sobě spojuje vlastnosti volných a vázaných částic, je vhodné použít v případě potřeby modelování elastických nebo deformovatelných těles. Na prezentaci bude popsán způsob modelování těchto těles a simulace deformací při působení vnějších sil na těleso. Budou také nastíněny způsoby vizualizace těles modelovaných částicovým systémem.

21. 11. 2002

J. Dohnal

Searching in Metric Spaces

Abstract
In order to speedup retrieval in large collections of data, index structures partition the data into subsets so that query requests can be evaluated without examining the entire collection. As the complexity of modern data types grows, metric spaces have become a popular paradigm for similarity retrieval. A new index structure, called D-Index, combines a novel clustering technique and the pivot-based distance searching strategy to speed up execution of similarity range and nearest neighbor queries for large files with objects stored in disk memories. Contrary to tree organizations, the D-Index structure is suitable for dynamic environments with a high rate of delete/insert operations.

28. 11. 2002

F. Procházka

UMT - Architecture and applications

Abstract
We will present the tool which is used as a framework for building knowledge oriented systems. This tool is called the Diamond tool and it is the member of the Universal Modelling Tools (UMT) class. The key features of the Diamond tool are
- ability to manage a meta* modelling (the number of the "meta" prefixes is not restricted)
- ability to use this tool as a modelling tool and as a production system at the same time
We will show how we can use these features and how it is implemented. We will mention some applications that use the Diamond tool. The Diamond tool is implemented in Prolog and the version 1.3.0 is currently available.

5. 12. 2002

D. Svoboda

Segmentation of tissue nuclei acquired from confocal microscope

Abstrakt:
We've been developing a large project in Laboratory of Optical Miscroscopy which offers simplier analysis of biological specimens acquired from confocal microscope. There is a necessary part of this project - it's a graphics library which supports many non-trivial 3D-image analysis functions. The aim of this talk is to introduce a draft of one module of this library. The module is about how to segment tissue nuclei and analyze them. Because the problem of tissue segmentation is very hard and nowadays there is no good algorithm dealing this problem, we'll be discussing other similar more or less succesful methods that has already been published.

12. 12. 2002

J. Stibor

Searching for groups of chromosomes in the digital biomedical image

Abstract
Nowadays it is necessary to process a large amount of image data in many branches of medicine. There are two possible approaches of the processing - either human controlled or fully automatic processing. The automatic approach is for many reasons better. I will present the fully automatic algorithm searching for groups of chromosomes in digital biomedical image. Input images were obtained from Laboratory of Optical Microscopy, FI MUNI. For their further analysis it is needed to locate areas containing biological material (in this case chromosomes) and this is just the work for mentioned algorithm.

POSTER SESSION

Jiri Barnat: How To Distribute LTL Model-Checking Using Decomposition of Negative Claim Automaton
Jan Pesl: Implementing neural network computations in Maple
Jan Strejcek: Boundaries of Equivalence Checking
Pavel Kolcarek: Multiresolution methods in force feedback
Jitka Zidkova: An Implementation of a Distributed Modular Model Checking Algorithm
Lubomir Markovic: Objects with Changeable Roles
David Safranek: SGCCS
Radek Oslejsek
David Antos, Petr Sojka
Jan Flasar, Ludek Pokluda: VRECKO
Miroslav Hrad: Time Stamping
Michal Batko: Scalable and Distributed Data Structures
Zdenek Kaminski: Combo 6
Jaroslav Silbersky: Environmental decision support systems

Jaro 2003

26. 2. 2003

Úvodní seminář jarního semestru

Program:
Informace o koncepci semináře v jarním semestru.
Domluva programu semináře.
Diskuse.

5. 3. 2003

J. Stibor

Searching for groups of chromosomes in the digital biomedical image

Abstract
Nowadays it is necessary to process a large amount of image data in many branches of medicine. There are two possible approaches of the processing - either human controlled or fully automatic processing. The automatic approach is for many reasons better. I will present the fully automatic algorithm searching for groups of chromosomes in digital biomedical image. Input images were obtained from Laboratory of Optical Microscopy, FI MUNI. For their further analysis it is needed to locate areas containing biological material (in this case chromosomes) and this is just the work for mentioned algorithm.

12. 3. 2003

James Thomas

Presentations in English and English in Presentations

Abstract
As a good command of English is necessary to survive in academia in the modern world, this one-off lecture-workshop will touch on a few issues relevant to non-native speakers writing and speaking in English for academic purposes. Having proofread many academic papers written by Czechs in English, and having attended numerous lectures and presentations given by Czechs in English, it is hoped that some of the conclusions from this data will be salutary. Furthermore, some general observations about the mode of presentations will be proffered

19. 3. 2003

Jan Frisch

Gentle introduction to the NET and Grid Management

Abstract
The presentation will give an introductory overview of the network, computers, clusters and Grids management. Selected techniques of network monitoring and management through the standardized protocols and services will be discussed, followed by presentation of their use in the large scale computing and storage systems, with the emphasis on heterogeneous environments. An illustration of how these methods can be applied for integration of distributed systems and creation and management of "virtual organziations" will be also presented. The last part will be dedicated to expected contribution for end users and project oriented teams and to future trends.

26.3. 2003

M. Kubásek

Introduction to the Semantic Web and Ontologies

Abstract
The semantic web is a vision about the future of the World Wide Web brought forward by the inventor of the web, Tim Berners-Lee. It is not an utopic vision but more the feeling that the web has created enormous possibilities and that the right thing to do is to make use of these possibilities. The fundamental part of the semantic web are ontologies, which provide description of knowledge representation at meta-level. Ontologies add explicit metainformation describing semantics to data. Knowledge in ontologies is formalized using five kinds of components: classes, relations, functions, axioms and instances. Nowadays, computers are going through the change from standalone isolated devices to entry points to a world information exchange network. Therefore, support for effective data, information and knowledge exchange is becoming the key effort in contemporary computer technology. Ontologies can offer such support. Ontologies enable also the reuse of existing knowledge bases in knowledge engineering.

2. 4. 2003

Eva Lasarcyk

Studies of Phonetics and Computational Linguistics in Bonn

Abstrakt:
The talk gives brief information about Bonn University, teaching and research, especially at the presenter's home institute "Institut für Kommunikationsforschung und Phonetik". The talk will concentrate on speech processing aspects as well as formal semantics, corpus linguistics, and text technology/hypermedia (project "milca"). Depending on the audience's interests, the presenter is prepared to talk about studying informatics at her university as well. Finally, a very brief overview about studying in Bonn and in Germany in general is provided.

L. Svoboda

Dialogový system UIO

Abstract:
Prezentace se bude týkat dialogového systému UIO. Systém je navržen pro zodpovídání otázek v českém jazyce. Jde o doménově orientovaný systém s pevnou bází znalostí. Na použitou množinu slov nejsou kladeny žádné omezující požadavky. Přitom je brán ohled na jednoduchost konfigurace domén, jejich snadnou editaci a "učení" nových vět.

9. 4. 2003

Jan Pešl

Teaching math with MappleNet project

Abstract
Computer-assisted math teaching becomes very important by expansion of information technologies. The offer of synoptic visualisation of math definitions and computations becomes math much more comprehensible for students. The authors of computer algebra systems are therefore developing new software projects related to math teaching. In case of Maple, there was introduced a project called MapleNet. It offers a complete software platform to enhance mathematics and related courses over the Web. It is suited to meet the emerging challenges of distance, continuing, professional and many other non-traditional forms of education and training. It offers innovative client-server based software components to make it easy for instructors to create and distribute electronic supplements and enhancements to courses. MapleNet provides students with a rich interactive environment to explore difficult concepts. Its applications and lesson components can be easily launched form any web browser and do not require any advanced computing knowledge to run, so students can spend their time learning rather than operating software. MapleNet combines Maple software engine on server with Java-based Web interfaces.

16. 4. 2003

L. Pokluda, J. Sochor

Spatial haptic orientation for visually impaired people

Abstract
Is it possible to show space relations to blind people? Techniques and experiments described in this talk explore several ways blind people can access spatial information using a one point force-feedback device. The presentation focuses on techniques supporting their orientation in buildings. Preliminary results show that a wider spectrum of different techniques suitable for blind people is needed.

23.4. 2003

Petr Kuba

Exploiting Sampling and Meta-learning for Parameter Setting for Support Vector Machines

Abstrakt:
It is a known fact that good parameter settings affect the performance of many machine learning algorithms. Support Vector Machines (SVM) and Neural Networks are particularly affected. In this paper, we concentrate on SVM and discuss some ways to set its parameters. The first approach uses small samples, while the second one exploits meta-learning and past results. Both methods have been thoroughly evaluated. We show that both approaches enable us to obtain quite good results with significant savings in experimentation time.

Michal Batko

TBA

30. 4. 2003

David Antoš

Associative Memories for IP Packet Routing

Abstract
The lecture describes the design of a Lookup Processor in the Combo6 IPv6/IPv4 accelerating card for IP routing. Header matching in IPv6 requires more than 440 bits to be checked when packet filtering is requested.
It is advantagous to use Content Addressable Memory (CAM) to perform a match, unfortunately, the width of the widest market available CAM is not enough. More advanced processing is needed. We propose a lookup processor based on CAM combined with a subsequent tree structure. For the purposes of the router, the structure has to be dynamically (re)computed out of configuration, routing table, and firewall setting.

7. 5. 2003

M. Marciniszyn

Encryption of Classical Information: Using Quantum Channel to Detect Eavesdropping

Abstract
We present a cryptographic system which uses quantum channel to transmit classical information between two parties. The transmitted information is perfectly secure in the sense that it is encrypted using the classical Vernam cipher. Any eavesdropping of the message has large probability of detection and therefore the key can be reused when no eavesdropping is detected. This system can be used also as the key distribution system.

14. 5. 2003

M. Hrad

TBA

21.5. 2003

J. Flasar

Improved interaction in local region

Abstrakt:
The interaction is almost accomplished with the virtual objects which the user sees. In this talk, we will show the methods for interaction in the user's local region. Than we will desribe design of collection of methods that allow user to interact not only with visible objects but even with objects that are hidden behind obstacles. At first, we will mention some techniques that have been using for solving the problem of interaction with objects hidden behind obstacle. Then we will describe new approaches that more improve interaction in this situation. It means, the techniques (approaches) in question either temporary change actual user's location to better location for interaction or reveal the hidden area by some way.

L. Markovič

Objects with Roles & VRECKO

Abstract:
A brief introduction of a technique of creating applications with dynamic behaviour -- "Objects with Roles" and its application in the HCI Lab's system VRECKO will be presented.

POSTER SESSION

V. Kadlec: PACE - Parser comparison and Evaluation
D. Svoboda, P. Matula: Multilevel Adaptive Thresholding and Dual Simplex Mesh Deformation as a Tool for Tissue Reconstruction
Miroslav Hrad: Time Stamping
Martin Vach: Biometrics
Pavel Kolcarek: Aplikace LOD v silové zpětné vazbě
Vlastislav Dohnal: D-index: Indexing Metric Spaces
Tomas Hnilica: WSDATA
Filip Prochazka: Diamantový modelovací nástroj
E. Hladká, P. Holub: AccessGrid Node at Faculty of Informatics
P. Holub: Protocols for High-Speed Networks
Petr Adamek
Martin Povolný: DEB
Stanislav Barton: Semantic Web Technology
Petr Sojka : Presentations in Spring 2003