UW Database Research Group Events -- Spring 2001

[Please remove <h1>]

Winter 2001

Note: Events of interest to the Database Research Group are posted to the uw.cs.database newsgroup and to subscribers of the dbgroup mailing list. If you wish to subscribe to the dbgroup mailing list, send mail to majordomo@db with "subscribe dbgroup" in the message body. To unsubscribe, send "unsubscribe dbgroup" to the same address.
DB group meetings
The DB group meets most Friday afternoons at 2pm, usually in DC1331. See the list of current events for times and locations of upcoming meetings. Each meeting lasts for an hour and features an informal presentation by one of the members of the group. Everyone is welcome to attend. These talks are intended to raise questions and to stimulate discussion rather than being polished presentations of research results. Speakers are determined using a rotating speaker list, which can be found on the DB group meeting page
DB seminar series
The DB seminar series features visiting speakers. These seminars are more-or-less monthly, and are usually scheduled on Monday mornings at 11am. See the list of current events for times and locations of upcoming seminars. The full schedule can be found on the DB seminar series page.

Recent and Upcoming Events


DB meeting: Friday, September 21st, 2:00 pm, DC3303 (the DB Lab)
Speaker: Kickoff meeting - no speaker.
Topic: Announcements and introductions.
Snacks: Yes!  Bring your own mug...

DB seminar: Monday, September 24, 11:00AM, DC1304
Speaker: Anthony Bonner, University of Toronto
Title: Large-Scale Physical Genome Mapping

DB meeting: Friday, September 28th, 2:00 pm, DC1331
Speaker: Airi Salminen
Topic: Modelling databases of multi-faceted XML documents
Abstract: A well-defined database system is based on a well-defined data model. The complexity of XML-related data repositories and the need to integrate the management of structured documents with the management of other types of data creates a special challenge for the underlying data model. Different data models have been suggested for XML data, for example, the XQuery 1.0 and XPath 2.0 Data Model. The models may be sufficient for developing capabilities dealing with the hierarchic structure of elements. To be able to manage XML documents with their multiple facets as a database, however, requires a richer data model. In the talk I first briefly discuss requirements for the data model. Then I explore the capabilities to use the production rules of the XML 1.0 specification as a basis for the model. The model would serve as a universal model from where views describing the different realities represented by XML documents could be derived.

References: 

A.Salminen and F.W.Tompa, Requirements for XML Document Database Systems, ACM Symposium on Document Engineering, Atlanta, Nov. 2001 (to appear). 

A.Salminen and F.W.Tompa, Grammars++ for Modelling Information in Text, Information Systems, Vol. 24, No. 1 (1999), 1-24.

Snacks: Huizhu Liu

DB meeting: Friday, October 5th, 2:00 pm, DC1304 
Speaker: Chiasen (Charles) Chung
Topic: Topic-Oriented Collaborative Web Crawling
Snacks: Ken Salem
Note: this is an MMath thesis presentation

DB meeting: Friday, October 12th, 2:00 pm, DC1331
Speaker: Vlado Keselj
Topic: Question-Answering Using Modular Stochastic HPSG Grammars
Snacks: Charlies Chung
Abstract: I will describe the architecture of a question answering system.  The key novel solutions implemented in the system are:
  • a "light" HPSG formalism with a stochastic component
  • grammar modules for different domains, and
  • just-in-time sub-grammar extraction for faster NL processing.
Some encouraging preliminary evaluation results will be presented and analyzed.

DB seminar: Monday, October 15, 11:00AM, DC1304
Speaker: Qiang Zhu, The University of Michigan - Dearborn
Title: Local Cost Estimation for Global Query Optimization in a Multidatabase System

DB meeting: Friday, October 19th, 2:00 pm, DC1331
Speaker: Gord Cormack
Topic: Web-Boosted Question Answering
Snacks: Vlado Keselj
Abstract: I will discuss our participation in the TREC 10 question answering task.  The objective of the task is to find short answers to factual questions within a fixed corpus of news data.  The performance of our system is much improved when Web data are included with the corpus data in initial phases of processing, and removed in a final filtering step.

DB meeting: Friday, October 26th, 2:00 pm, DC1331
Speaker: Heng Yu
Topic: A DBMS Perspective on Update of Kernel Data in Linux
Snacks: Gord Cormack
Abstract: The information integration of legacy Embedded Control Program (ECP) is of great significance in industry and has received good attention in research. To provide a DBMS wrapper for ECP, it is necessary to carry on reverse engineering based on ECP source code in order to find hidden database features. Linux operating system is taken as a specimen of ECP for analysis. The focus is on the system call fork() to explore abstraction both from kernel data structures to database scheme and from control flow in program to transaction model. The protection facilities in Linux is surveyed and the protection pattern in the study case is analysed. The traditional transaction model saga is used to model the transaction feature in linux system call fork().

DB meeting: Friday, November 2nd, 2:00 pm, DC1331
Speaker: Kevin See
Topic: With the increasing use of the XML technology, an efficient way to store and query the XML data will be crucial. There are several ways to manage the XML document. The simplest way is to store it as a file in the operating system level. Another approach is to use existing database technology (relational database and object-oriented database) as the backend using some form of mapping. Yet another approach is to develop a new storage technology designed specifically to store XML. In this presentation, we will focus on two of  the newest mapping proposals that employ a relational database as a backend. The desiderata of an XML benchmark will be addressed along with existing proposals of XML benchmark.
Snacks: Heng Yu
Note: this is an MMath essay presentation

DB meeting: Friday, November 9th, 2:00 pm, DC1331
Speaker: Lei Chen
Topic: In this talk, I will give a brief overview of MPEG-7.  The following topics will be addressed:
  •      Motivation and Objectives of MPEG-7
  •      MPEG-7 structure
  •      MPEG-7 data definition language
  •      MPEG-7 video and audio descriptors
  •      MPEG-7 multimedia description schemes
  •      Research issues regarding to MPEG-7
Snacks: Grant Weddell

DB seminar: Monday, November 12th, 11:00AM, DC1304
Speaker: Sylvia Osborn, University of Western Ontario
Title: Database Applications of Role-Based Access Control

DB meeting: Friday, November 23rd, 2:00 pm, DC1331
Speaker: Anil Goel
Topic: I'll discuss some published work on the problem of optimizing and executing "TOP N" queries.
Snacks: Lei Chen

DB meeting: Friday, November 23rd, 2:00 pm, DC1331
Speaker: Ian Davis
Topic: Integrating Xpath with SQL or, if you prefer, SQL with Xpath
Snacks: Anil Goel
Abstract: Xpath is a specification which allows fragments of text to be recovered and operated upon.  Xpath restricts the information recovered according to a complex set of considerations which include structural requirements, relationships between fragments of text, types of text, and textual content.  Xpath further allows texts to be restricted by sequences of boolean predicates which may themselves employ knowledge of the texts being restricted in deriving a true (include) or false (exclude) result.

The boolean predicates as provided in Xpath approximate a simplified subset of the predicates available in SQL2.  In this talk I will focus on how Xpath can be extended to support a direct mapping of text into relations usable by SQL2, and how the boolean predicates employed by Xpath can be generalised to support arbitrary SQL2 predicates.

I will also focus on how the internal runtime optimisations necessary to support indexing of text accessed by external SQL where clauses, can be transformed into boolean predicates embedded within the Xpath expressions returning these texts.


DB seminar: Monday, December 3, 11:00AM, DC1304
Speaker: Jose Blakeley, Microsoft Corp.
Title: Programmability in the SQL Server DBMS

DB meeting: Friday, December 7th, 2:00 pm, DC1331
Speaker: Hui Zhang
Topic: Translating FLWR Expressions into Extended Relational Algebra
Snacks: Ian Davis
Abstract: As XML becomes more widespread as a standard representation for data, XML-based query languages and their evaluations are increasingly important. Certainly, there are many alternative ways to process XML queries. In common with other researchers, we wish to capitalize on the extensive work invested in relational database technology. In particular, we take an algebraic approach to translate FLWR expressions of W3C's XQuery to an extended relational algebra with support for the structured text datatype. In this talk, I will present our approach and study several optimization cases.

This page is maintained by Frank Tompa and Ken Salem.