Meeting Agenda
Wednesday, March 6, 1996
- Location:
- DC 1304
- Time:
- 12:30 PM
- Chair:
- John Kominek
1. Adoption of the Agenda - additions or deletions
2. Coffee Hour
- Coffee hour this week:
- Navid Sadikali
-
- Coffee hour next week:
- any volunteers?
3. Next meeting
- Date:
- March 13, 1996
- Location:
- DC 1304
- Time:
- 12:30 PM
- Chair:
- Rob Kroeger
- Technical presentation:
- Gilles Khouzam
4. Forthcoming
- Chairs:
-
- Iain Little
- Steve Mann
- Mike McCool
- Tech Presenters:
-
- Gilles Khouzam
- John Kominek
- Rob Kroeger
5. Technical Presentations
- Presenter:
- Rick Kazman
-
- Audio Enhanced 3D Interfaces for Visually Impaired Users
-
Three dimensional computer applications such as CAD packages are often
difficult to use because of inadequate depth feedback to the user. It
has, however, been shown that audio feedback can help improve a user's
sense of depth perception. This talk describes an experiment which
evaluates the use of three different audio environments in a 3D task
undertaken by visually impaired users. The three audio environments map
tonal, musical, and orchestral sounds to an (x, y, z) position in a 3D
environment. In each environment the user's task is to locate a target
in three dimensions as accurately and quickly as possible. This
experiment has three important results: that audio feedback improves
performance in 3D applications for all users; that visually impaired
users can use 3D applications with the accuracy of sighted users; and
that visually impaired users can attain greater target accuracy than
sighted users in a sound-only environment.
[This work was done jointly with Stephen Mereu.]
6. General Discussion Items
- Lab PC's: Iain Little -- virus checking and disk space
7. Action List
8. Director's Meeting
9. Seminars
DEPARTMENT OF COMPUTER SCIENCE
UNIVERSITY OF WATERLOO
SEMINAR ACTIVITIES
SCIENTIFIC COMPUTATION SEMINAR
-Wednesday, March 6, 1996
Peter A. Forsyth, Department of Computer Science,
University of Waterloo, will speak on ``Object Oriented
Methods in CFD''.
TIME: 4:00-5:00 p.m.
ROOM: MC 5158
ABSTRACT
This talk will discuss the use of object oriented
methods (C++) in CFD. Example computations for a C++
implementation of solution of the compressible Navier-
Stokes equations on unstructured meshes will be
presented. A full Newton approach is used, with the
Jacobian being solved using a PCG-like iterative
method. Timings will be presented for implementations
of the matrix solution algorithm in both C++ and
Fortran. It is shown that, provided certain language
features of C++ are avoided, there is practically no
difference in execution time between C++ and Fortran.
An overview of the advantages and disadvantages of C++
for numerical computation will also be presented.
=======================================================
DEPARTMENT OF COMPUTER SCIENCE
UNIVERSITY OF WATERLOO
SEMINAR ACTIVITIES
THEORY SEMINAR
-Wednesday, March 6, 1996
Radu Neglescu, graduate student, Dept. Comp. Sci.,
Univ. Waterloo will speak on ``Process Spaces''.
TIME: 3:30-4:30 p.m.
ROOM: DC 1304
ABSTRACT
We present process spaces, a unified theory of
interacting systems. The main new trait, abstract
executions, leads to a simple and general set
formalism. For concurrent systems, process spaces
yield decoupled and homogeneous studies of diverse
correctness concerns. Applications to other
interacting systems, such as electrical networks, are
also possible. Some algebraic properties of process
spaces and their practical significance are discussed.
=======================================================
DEPARTMENT OF COMPUTER SCIENCE
UNIVERSITY OF WATERLOO
SEMINAR ACTIVITIES
COMPUTER SCIENCE SEMINAR
-Thursday, March 7, 1996
Russell Greiner, Siemens Corporate Research, Princeton,
NJ, will speak on ``Learning Effective Performance
Elements''.
TIME: 4:00-5:00 p.m.
ROOM: DC 1350 *NOTE ROOM*
ABSTRACT
Many learning systems search through a space of
possible performance elements, seeking an element whose
expected utility, over the distribution of problems, is
high. As the task of finding the globally optimal
element is often intractable, many practical learning
systems instead hill-climb to a local optimum. Even
this is problematic as the learner typically does not
know the underlying distribution of problems, which is
essential for computing the utility of an element, and
hence for determining which element is optimal. This
research addresses the task of approximating this
hill-climbing search when the utility function can only
be estimated by sampling. We present a general
algorithm, PALO, that returns an element that is, with
provably high probability, essentially a local optimum.
We then demonstrate the generality of this algorithm by
presenting four distinct applications, which find
* an efficient ordering of a set of clauses [speedup learning];
* an accurate revision of a rule base [theory revision];
* a categorical approximation to a
theory [completeness/tractability tradeoff];
and
* an effective subset of a landmark-set
[position registration for navigation].
We use both formal theorems, and empirical results, to
quantify the effectiveness of these learning
algorithms.
Before covering this material, I will first provide
brief synopses of several other recent results,
including
* optimal solution caching (aka view realization);
* sequential learning techniques; and
* learning "default" classifiers (for labeling incomplete data).
=======================================================
DEPARTMENT OF COMPUTER SCIENCE
UNIVERSITY OF WATERLOO
SEMINAR ACTIVITIES
COMPUTER SCIENCE SEMINAR
-Monday, March 11, 1996
Praveen Seshadri, University of Wisconsin, will speak
on ``Sequence Data Management''.
TIME: 4:00-5:00 p.m.
ROOM: DC 1302
ABSTRACT
Database systems are increasingly trying to support new
kinds of complex data. My research studies how a
database system can support sequences , a specific
- -
kind of complex data representing ordered information.
There are two distinct aspects to this talk. The first
aspect deals with the data model for sequences, an
algebraic framework for sequence queries, and
techniques for query optimization and evaluation.
The second aspect deals with the practical design of a
database system supporting sequences. The most
important design issue is how sequences interact with
relations and other complex data types. I tackle this
issue using a novel loosely-coupled paradigm based on
an Enhanced notion of Abstract Data Types(E-ADTs). The
SEQ database system has been implemented using this
design paradigm, resulting in efficient support for
sequences. Extrapolating from this experience, I
propose that E-ADTs may well be the design of choice
for next-generation database systems supporting complex
data.
=======================================================
An Algebra for Structured Text Search
by
Gord Cormack
Abstract
A query algebra is presented that expresses searches on structured
text. In addition to traditional full-text boolean queries that search
apre-defined collection of documents, the algebra permits queries that
harness document structure. The algebra manipulates arbitrary intervals of
text, which are recognized in the text from implicit or explicit markup.
The algebra has seven operators, which combine intervals to yield new ones:
containing, not containing, contained in, not contained in, one of, both of,
followed by. The ultimate result of a query is the set of intervals that
satisfy it.
For more information and a look at the direct application of this algebra
try "The Multitext NetNews Searcher" (http://multitext.uwaterloo.ca).
Wednesday, 13 March, 1996
16:30
MC 4059
All are welcome.
Doughnaughts and Tea will be served.
oo$$$$$ooo o$$$$$oo$$$$$oo
o$$"" $$$" o$$"" $$$$$$ ""$$o C O M P U T E R
o$$ $$" $$" o$" "$$ $$o
$$ $$ $$o $$ $$o $$ S C I E N C E
$$o "$$ $$ o$$ $$ $$
$$o "$o $$" o$$ o$$ o$$" C L U B
"$$ooo$$$$$$ooo$$" $$$$ooo$$$
A Student Chapter of the ACM
=======================================================
10. Lab Cleanup (until 1:30 or 5 minutes)