[ont.events] U of Toronto Computer Science activities, Oct. 14-17

clarke@utcsri.UUCP (Jim Clarke) (10/03/86)

              (SF = Sandford Fleming, 10 King's College Road)
              (GB = Galbraith Building, 35 St.George Street)

COLLOQUIUM, Tuesday, October 14, 11 am, SF 1101

                           Professor Rudi Mathon
                  Computer Science, University of Toronto

                 "The Beauty and Complexity of Iteration"
                             (abstract below)

AI SEMINAR, Tuesday, Oct. 14, 3 pm, GB 119

                               Dr. Dan Fass
                        New Mexico State University

                           "Collative Semantics"

THEORY SEMINAR, Thursday, Oct. 16, 3 pm, GB 220

                         Professor Patrick Dymond
             Visiting from University of California, San Diego

   "Speedup of Sequential Multitape Turing Machines using Parallelism"
                             (abstract below)

NUMERICAL ANALYSIS, Thursday, Oct. 16, 4 pm, GB404

                           Professor Ken Jackson
                  Computer Science, University of Toronto

                      "Parallel Runge-Kutta Methods"
                             (abstract below)


                                 ABSTRACTS

                           Professor Rudi Mathon
                  The Beauty and Complexity of Iteration
                       (Julia sets and Computer Art)

     The analysis and impact of competition in real world or model systems
is a scientific topic of great challenge in numerous disciplines such as
biology, chemistry, ecology, economics etc.  Recent computer experiments
have demonstrated that even simple systems with competition can lead to
spatial patterns of great complexity and beauty. While such experiments
will continue to enhance our intuition, in the future they might develop
into a sophisticated art form.  A presentation of some award winning pic-
tures is included.

                         Professor Patrick Dymond
              Speedup of Sequential Multitape Turing Machines
                             Using Parallelism

     Joint research with Martin Tompa will be described, which investigates
the utility of unbounded parallelism in accelerating sequential (Turing
machine) computations.  We show that
                Tm-TIME(T) is included in PRAM-TIME(T**1/2)
and that
                Tm-TIME(T) is included in ATM-TIME(T/logT).
These improve on the previous best results for these problems, due to Hop-
croft, Paul and Valiant, and to Paul and Reischuk.  The proof of the second
inclusion refines a theorem of Hopcroft, Paul and Valiant relating Turing
machine time to space, and uses a new kind of 2-person pebbling game which
may be of independent interest.  For example, the theorem of Paterson and
Valiant that any combinational circuit of size T computes a function of
depth complexity O(T/logT) can be proved as a direct corollary of this
approach.

                           Professor Ken Jackson
                       Parallel Runge-Kutta Methods

     In this seminar, we consider whether it is possible to reduce the time
required to take one integration step with a pth-order Runge-Kutta formula
by doing several function evaluations in parallel.  Some negative results
are given for Runge-Kutta formulas in the standard one-step form, but it is
shown that a significant speed-up can be gained by using a predictor-
corrector formulation.
-- 

Jim Clarke -- Dept. of Computer Science, Univ. of Toronto, Canada M5S 1A4
              (416) 978-4058
{allegra,cornell,decvax,linus,utzoo}!utcsri!clarke

armin@utai.UUCP (Armin Haken) (10/08/86)

> 
>               (GB = Galbraith Building, 35 St.George Street)

 	The following seminar has been changed:

> AI SEMINAR, Tuesday, Oct. 14, 3 pm, GB 119
> 
>                                Dr. Dan Fass
>                         New Mexico State University
> 
>                            "Collative Semantics"

	The revised information is:

AI SEMINAR, Tuesday, Oct. 14, 3 pm, GB 119

                      Dr. Evangelos Milios
                     University of Toronto

             "Signal Processing and Interpretation
              Using Multilevel Signal Abstractions"


> Jim Clarke -- Dept. of Computer Science, Univ. of Toronto, Canada M5S 1A4
>               (416) 978-4058
> {allegra,cornell,decvax,linus,utzoo}!utcsri!clarke

Armin Haken, DCS, University of Toronto, (416)978-6277