Information Retrieval List Digest 090 (November 4, 1991)
URL = http://hegel.lib.ncsu.edu/stacks/serials/irld/irld-090

IRLIST Digest
November 4, 1991
Volume VIII, Number 47
Issue 90
**********************************************************
I.  NOTICES
    A. Meeting Announcements/Calls for Papers
       1. Combinatorial Pattern Matching, Tucson, Arizona,
          April 29-May 1, 1992
II. QUERIES
    B. Requests for Information
       1. TOPIC
IV. PROJECT WORK
    B. Bibliographies
       1. IR-Related Dissertation Abstracts
**********************************************************
I. NOTICES
I.A.1.
Fr: Udi Manber <udi@cs.arizona.edu>
Re: Call for Papers - Combinatorial Pattern Matching

                         CALL FOR PAPERS
                 COMBINATORIAL PATTERN MATCHING
          Tucson, Arizona, USA, April 29 - May 1, 1992

The third symposium on Combinatorial Pattern Matching will be
held in Tucson, Arizona, USA, on April 29 - May 1, 1992.  (This
is the end of the week preceding the Theory of Computing
conference to be held in Victoria, BC, May 3-5.) The first two
meetings were held at the University of Paris in 1990 and at the
University of London in 1991.  Papers in all areas related to
combinatorial pattern matching and its applications will be
considered, including, but not limited to, string algorithms,
pattern recognition, applications in molecular biology, text
searching, information retrieval, symbolic computing, and data
compression.

Two types of submissions will be considered.
  1. Original research unpublished elsewhere.
  2. Survey of important results, especially recent ones.

Proceedings, probably in the Springer-Verlag series on Computer
Science, will be published.  The goal is to keep the attendance
at about 50-70 participants to allow ample time for informal
interaction.  Thus, attendance will be by invitation only.  If
you are interested in attending the meeting without submitting a
paper, contact us via the address below.  Some financial
assistance for students will probably be available.

To submit a paper, please send 10 copies of an extended abstract
(5-10 pages) to the address below by December 16, 1991.  This is
a strict deadline.  Papers arriving after that date will not be
considered.  Please include e-mail address if possible.  Authors
will be notified by February 10, 1992.

                       Udi Manber, CPM 92
                    Dept. of Computer Science
                      University of Arizona
                      Tucson, AZ 85721 USA
               e-mail address: udi@cs.arizona.edu.

Program Committee:  A. Apostolico, M. Crochemore, Z. Galil,
G. Gonnet, D. Gusfield, D. Hirschberg, U. Manber, E. Myers,
F. Tompa, and E. Ukkonen.
**********************************************************
II. QUERIES
II.A.1.
Fr: Lou Rosenfeld <lou@zug.csmil.umich.edu)
Re: just checking.../Verity TOPIC

I was pleased to stumble upon this newsgroup in my .newsrc, but
saddened that it seems kind of dead right now.  Anyone out there
a veteran of a more fecund period of activity on this newsgroup?
I'm interested in information retrieval theory, but more
interested in practical applications.  Would this be the right
newsgroup to discuss applications?

Here's my stab at reviving this group:  I am working with the
Verity TOPIC package on a number of projects at the University of
Michigan.  We will be testing the applicability of TOPIC to
dynamic data sources (ie, setting up current awareness services
for Usenet and listserv postings), and using TOPIC as a brokering
service for information on research grants and colleagues for
faculty here at UM.

TOPIC itself is a somewhat interesting program in that:  it
handles physically distributed information in almost any format
(thus providing a common interface to many different databases);
it will soon be available with a graphical windows-based
interface on all major platforms; and offers "Topic" searching,
as well as keyword and Boolean.  "Topics" are graphical
representations of Boolean searches which allow the weighting of
search terms and the use of probabilistic and statistical
algorithms; thus, retrieved documents are ranked by "relevance".
We are particularly interested in studying these application of
these non-Boolean algorithms, the user's interaction with ranked
retrieval, and the semantic and other linguistic relationships
that can be built into "Topics".  The group I represent is
comprised of researchers from the School of Information and
Library Studies, the College of Engineering, and others at
Michigan.

Anyone with experience with TOPIC, ideas, suggestions,
criticisms, flames?

--Lou Rosenfeld, Univ. Michigan
**********************************************************
IV. PROJECT WORK
IV.B.1.
Fr: Susanne M. Humphrey <humphrey@nlm.nih.gov>
Re: IR-Related Dissertation Abstracts

The following are citations selected by title and abstract as
being related to Information Retrieval (IR), resulting from a
computer search, using BRS Information Technologies, of the
Dissertation Abstracts Online database produced by University
Microfilms International (UMI).  Included are UMI order number,
title, author, degree, year, institution; number of pages, one or
more Dissertation Abstracts International (DAI) subject
descriptors chosen by the author, and abstract.  Unless otherwise
specified, paper or microform copies of dissertations may be
ordered from University Microfilms International, Dissertation
Copies, Post Office Box 1764, Ann Arbor, MI 48106; telephone for
U.S.  (except Michigan, Hawaii, Alaska):  1-800-521-3042, for
Canada:  1-800-268-6090.  Price lists and other ordering and
shipping information are in the introduction to the published
DAI.  An alternate source for copies is sometimes provided.
Dissertation titles and abstracts contained here are published
with permission of University Microfilms International,
publishers of Dissertation Abstracts International (copyright by
University Microfilms International), and may not be reproduced
without their prior permission.

AN University Microfilms Order Number ADG91-07195.
AU HINDES, MARY ANN.
TI THE SEARCH PROCESSES AND ATTITUDES OF STUDENTS ACCESSING CD-ROM
   RESOURCES: A CASE STUDY OF TWO HIGH SCHOOL MEDIA CENTERS.
IN University of Georgia Ed.D. 1990, 175 pages.
DE Education, Technology.  Information Science.
AB The purpose of this study was to examine the search processes of
   students accessing CD-ROM resources and to describe their attitudes
   concerning this approach to computer-based information retrieval.
   This study employed a case study research design to examine student
   use of CD-ROM technology in two high school media centers.  Four
   media specialists and twenty-two students were observed and
   interviewed during March 1990.  The researcher observed fifty-five
   instances of students accessing the available CD-ROM resources.

   Three categories emerged from the analysis of the data concerning the
   students' search processes and two categories emerged from the data
   concerning students' attitudes towards computer-based searching of
   CD-ROM resources.  The three categories of the Search Process
   actually reflect a process, viz., initiation of the search,
   modifications to the search, and selection of citations.  The Search
   Initiation phase of the search process is described by the following
   properties which reflect the students' approach: the Shotgun
   Approach, the Educated Guess, and the Resourceful Approach.  The
   second phase of the search process, called Search Modification,
   addressed that part of the search involving students' manner of
   Broadening the Search or Narrowing the Search.  The third phase of
   the search process was the students' Selection of Citations.
   Students selected citations based on the Availability of the
   full-text articles in their local media centers or solely on the
   seeming Aptness of the information.  From the data concerning
   students' attitudes towards the use of CD-ROM resources, two major
   categories developed.  Students felt Comfortable using CD-ROM
   technology and were Enthusiastic about this form of computer-based
   information retrieval.  Students considered the systems easy to
   access and developed a personalized relationship with the computer.
   Enthusiastic responses were generated by the sense of independence
   students derived from using CD-ROM resources.

   Two general conclusions of the study were (a) Students experienced
   difficulty performing searches, a difficulty exacerbated by their
   failure to use the tutorials or help screens included in the search
   systems; and (b) despite the difficulties encountered and their
   failure to take advantage of the full capabilities of the systems,
   students liked this form of information retrieval, preferring it to
   print-based searching.

AN University Microfilms Order Number ADG91-06437.
AU KIM, CHONGTAI.
TI DEVELOPMENT OF A NATURAL LANGUAGE INTERFACE TO A SLEEP EEG DATABASE.
IN The University of Florida Ph.D. 1990, 175 pages.
DE Engineering, Electronics and Electrical.  Artificial Intelligence.
AB A natural language (NL) system called SEEGER (Sleep $\underline{\rm
   EEGer}$) has been developed to retrieve information from a sleep
   database without requiring any database expertise or computer
   programming knowledge of the operator.  A new knowledge-based
   structure for building a NL system has been developed and designed
   into SEEGER.  It provides a more convenient method for incorporating
   domain knowledge than the previous approaches to building NL systems.
   In addition, it provides a more efficient database mapping method for
   those NL requests that rely on bottom-up parsing more often than
   top-down parsing.  SEEGER handles ungrammatical input, makes
   inferences, and disambiguates some complicated word senses.  It
   engages in an interactive dialogue to correct spelling errors, to
   clarify ambiguous or incomplete queries, and to enter a synonym when
   a word is not in the lexicon.  The database is then retrieved by
   retrieval query created from SEEGER.  The sleep database has been
   designed to contain information about subject record, channel
   recording, epoch summaries of waveform occurrences in EEG/EOG/EMG,
   sleep stage parameters, and night parameters.  It has been
   implemented with the commercially available dBASE IV database
   management system.

   SEEGER is now in the experimental stage, and it was evaluated using
   four users familiar with sleep data analysis.  It gave correct
   answers at the performance level of 63%, achieving the preliminary
   design goal.  The entire program has been written in about 10,000
   lines of LISP.  The system performance was analyzed in terms of the
   elements of the syntactic variations, semantic complexities, and
   interactive dialogues based on the test results.  These tasks show
   that future improvements can lead to a sleep NL system performing at
   the level of a data processing technician, but it would take at least
   two man-years of effort, assuming the developers already possess some
   background in artificial intelligence programming.  SEEGER has been
   implemented in a personal computer; it requires 6 megabytes of system
   memory and 15 megabytes of hard disk storage for the LISP environment
   and dBASE IV.  Finally, the guidelines for future development are
   suggested.

AN University Microfilms Order Number ADG91-09215.
AU VASSILAS, NIKOLAOS.
TI PERFORMANCE OF NEURAL ASSOCIATIVE MEMORIES FOR CHARACTER RECOGNITION
   AND ASSOCIATIVE DATABASE RETRIEVAL.
IN University of Minnesota Ph.D. 1990, 105 pages.
DE Engineering, Electronics and Electrical.  Computer Science.
   Artificial Intelligence.
AB Neural associative memories have been extensively used in
   applications that perform classification or some kind of input-output
   mapping.  The main advantage of neural associative memories over
   conventional computer memories is their robustness in noisy
   environments, generalization and fast retrieval capabilities.  In
   this work we use a robust associative memory model to investigate the
   performance, capacity and saturation effects of three of the most
   popular associative memory paradigms, namely: back-propagation
   networks, correlation matrix memories and generalized inverse
   memories.  The comparison studies use binary character recognition
   and spelling correction as application domains.  Our analytic and
   experimental modeling shows that correlation matrix memories used as
   classifiers exhibit overall superiority over other training rules.
   The correlation matrix memories are then used as the building block
   of a new neural model for semantic categorization and concept
   formation.  The proposed model represents a hierarchical two-level
   associative memory, where the first level memory stores categories
   (concepts) of semantically related items, and each of the second
   level memories stores items (prototypes) forming a given concept.
   Memory construction (or learning) is based on supervised learning.
   Besides its ability to form and store semantic categories of input
   patterns, the proposed hierarchical model offers substantial scaling
   advantages over traditional single-level associative memory systems
   in terms of software implementations.  Experimental results (assuming
   fixed prototypes) for binary character recognition and spelling
   correction applications demonstrate excellent quality of associative
   retrieval and superior scaling properties of the proposed model.
   Finally, we propose an iterative scheme for adaptive prototype
   formation from repeated presentations of noisy inputs, and present
   theoretical analysis and proof of convergence.  We also find optimal
   adaptation parameters that provide the best convergence rate.
**********************************************************
IRLIST Digest is distributed from the University of California,
Division of Library Automation, 300 Lakeside Drive, Oakland, CA.
94612-3550.

Send subscription requests to: LISTSERV@UCCVMA.BITNET
Send submissions to IRLIST to: IR-L@UCCVMA.BITNET

Editorial Staff:
 Clifford Lynch lynch@postgres.berkeley.edu or calur@uccmvsa.bitnet
 Nancy Gusack ncgur@uccmvsa.bitnet
 Mary Engle engle@cmsa.berkeley.edu or meeur@uccmvsa.bitnet

The IRLIST Archives will be set up for anonymous FTP, and the
address will be announced in future issues.  To access back
issues presently, send the message INDEX IR-L to
LISTSERV@UCCVMA.BITNET.  To get a specific issue listed in the
Index, send the message GET IR-L LOG ***, where *** is the month
and day on which the issue was mailed, to LISTSERV@UCCVMA.BITNET.

These files are not to be sold or used for commercial purposes.
Contact Nancy Gusack or Mary Engle for more information on
IRLIST.  The opinions expressed in IRLIST do not represent those
of the editors or the University of California. Authors assume
full responsibility for the contents of their submissions to
IRLIST.