Department of Mathematics and Statistics
Calendar of Events
September 1994 - July 1996
Thursday, September 22
- Colloquium
Statistical Estimation of Fractal Dimension
Professor Peter Hall
Australian National University
September 22, 1994 3:10-5:30 p.m. ** NOTE THURSDAY**
Sidney Smith Hall **Room SS2117**
Two 50 minute talks with an intermission. The first talk will be for
somewhat more general audiences; the second talk will be more mathematical.
ABSTRACT
Statistical scientists are becoming increasingly interested in modelling
real phenomena using self-similar processes. Applications are particularly
diverse; they range from the classification of metal or plastic surfaces in
terms of roughness, to the analysis of wave characteristics in the open sea,
and the study of nervousness in exchange rate markets. This talk was
motivated by such applications, and the work that will be described is
largely the result of collaborative research with CSIRO's Division of
Mathematics and Statistics.
A variety of stationary processes, including many Gaussian processes,
have sample paths that exhibit fractal-like behaviour. They may be used to
model line-transect samples drawn from irregular surfaces. There is a
plethora of different estimators that may be employed to estimate fractal
dimension; they include methods based on the variogram, the periodogram, box
counting or level crossing. Some of these will be discussed. Multivariate
data are becoming increasingly common, and applications in that context are
of particular interest, not least because of the potential for anisotropy of
fractal features. These matters will be addressed.
Thursday, October 20
- Colloquium: Professor Michael Barr (McGill University)
Time: 3.30-4.30pm
Place: S174R (Ross Building)
Title: Acyclic Models
Abstract:
Acyclic models is a powerful technique in algebraic topology and
homological algebra in which facts about homology theories are
verified by first verifying them on ``models'' (on which the homology
theory is trivial) and then showing that there are enough models to
present arbitrary objects. One version of the theorem allows one to
conclude that two chain complex functors are naturally homotopic and
another that two such functors are object-wise homologous. Neither is
entirely satisfactory. The purpose of this paper is to provide a
uniform account of these two, fixing what is unsatisfactory and also
finding intermediate forms of the theorem.
Friday, October 21
- Topology Seminar
At 1:30 p.m. in N638R
George Delistathis
A compact Hausdorff space in which no Baire set is regular open
Monday, October 24
- Applied Math Seminar:
Dr. Hu
3:30-4:30
"Continuous Runge-Kutta methods for delay differential equations"
Abstract:
We consider continuous Runge-Kutta methods for vanishing
delay differential equations. If the lag term smaller than
the stepsize when solving these problems, the standard explicit
equations associated with the underlying CRK formula become
implicitly defined . To solve these implicit equations, we
develop an iteration scheme based on the use of special 2-step
interpolant to generate the initial guess.
Numerical tests show favorable
aspects compared with other approaches.
Friday, October 28
- Statistics Seminar
Professor Michael Evans (Department of Statistics, University of Toronto)
will speak on
"Bayesian Hypothesis Testing Procedures Derived Via
The Concept of Surprise"
at 10:30 a.m. in N638 Ross.
ABSTRACT
We consider the general problem of testing a null hypothesis $H_0$ and
the related problems of model checking and model selection all within
the Bayesian context. We approach these problems via the basic concept
of surprise. A general theory is developed that allows complete freedom
to the statistician for the choice of the prior and the choice of the
set of alternatives to which $H_0$ is compared. As a special case the
theory leads to a combination of the posterior probability and Bayes
factor approaches to Bayesian hypothesis testing. It is shown that a
natural choice of prior and of the set of alternatives leads to a
resolution of a conflict between Bayesian and classical frequency
theory; namely the avoidance of Lindley's paradox. This is achieved
within the context of proper Bayesian inference which is not the case
for other attempts at resolving these problems. Further we discuss the
use of surprise in the model checking procedures suggested in Box
(1980). The relevance of the principle of least surprise, as stated in
Good(1988), to almost all problems of Bayesian inference is noted. It
is shown that a defect in previous approaches to surprise; namely the
change of variable problem, can be completely and appropriately avoided
within the context of Bayesian inference.
- SEMINAR IN HISTORY AND PHILOSOPHY OF MATHEMATICS AND
MATHEMATICS EDUCATION
Professor Solomon Marcus (Universite Laval) will give a talk on
MATHEMATICAL MISTAKES AS A SOURCE OF CREATIVITY at 2:00 p.m. in N501
Ross.
ABSTRACT
Mathematical mistakes are classified into syntactic, semantic and pragmatic,
and they are examined at three different levels: the creative work of
professional mathematicians, the teaching of mathematics and the learning of
mathematics. Several historical examples are analyzed in order to show that
mistakes are genuine in most pioneering works and they become a source of
creativity. Then it is shown that mistakes play a fundamental role in both
creative teaching and creative learning. Unfortunately, the teaching of
mathematics is in most cases concerned almost exclusively with syntactic
mistakes, while semantic mistakes are ignored or marginalized and pragmatic
mistakes are left to the job of the so-called specialists in pedagogy.
Friday, October 28
Tuesday, November 1
- Automatic Group Seminar: 9:00 to 10:00, N638,
D.Promislow will continue speaking.
Friday, November 4
- STATISTICS SEMINAR
Professor K. S. Brown (Department of Statistics and Actuarial Science,
University of Waterloo) will speak on
"Methodological Issues in the Evaluation of Behavioural Interventions"
at 10:30 a.m. in N638 Ross.
ABSTRACT
Studies which evaluate the effectiveness of behavioural interventions
such as smoking prevention programs encounter methodological issues which are
not seen in traditional trials of therapeutic agents. In this talk we will
discuss and illustrate design and analysis issues such as the use of
randomization of intact units, or clusters, to intervention or control
conditions; the choice of response variable(s); the longitudinal follow-up of
individuals; the use of data collected from multiple sources including
individuals, providers, and the system; and the modeling of the underlying
behavioural process. These issues will be illustrated using data from three
longitudinal studies of the effectiveness of school-based smoking prevention
curricula.
- Topology Seminar: At 1:30 p.m. in N638R,
Arnie Miller (University of Wisconsin, Madison visiting York University)
"On almost continuous functions"
Wednesday, November 9
- Analysis Seminar:
Dr. Minh Dzung Ha from the University of Toronto will give a talk
entitled " Convergence of Operators in Ergodic Theory" at 1:30 p.m. in N638R.
Thursday, November 17
- Probability Seminar:
"Self-avoiding walks: Critical exponents and Monte Carlo methods"
Neal Madras (Department of Mathematics and Statistics,
York University), 4:00-5:00 p.m.,
Sidney Smith Hall **ROOM SS2130**
ABSTRACT
The self-avoiding walk is a simple combinatorial model that has important
applications in statistical mechanics and polymer science. Of primary
importance is the model's scaling behaviour, which is characterized by
a set of critical exponents. Mathematically rigorous results about
scaling behaviour are in woefully short supply, so physicists and
chemists often study this model by a variety of nonrigorous methods.
This talk will focus on the use of Monte Carlo methods: How does
one efficiently generate self-avoiding walks at random? What can one say
rigorously about such methods? I shall also describe some recent work
with B. Li and A.D. Sokal that resolves a disagreement between
different methods on the value of a certain critical exponent,
while simultaneously highlighting an error that has persisted in the
polymer literature for many years.
This seminar is joint with the the Department of Mathematics
Friday, November 18
- Statistics Seminar:
Professor Peter D. M. Macdonald
(Department of Mathematics and Statistics, McMaster University)
will speak on "Mark-recapture estimation with small numbers of recaptures
and an uncertain number of marks" at 10:30 a.m. in N638 Ross.
ABSTRACT
When marked animals are sparsely distributed in a large population and
the expected number of recaptures is small, population estimates based
on the Poisson or negative binomial distributions are appropriate. I
will review these methods and consider ways to reduce bias.
In many mark-recapture studies the number of marked animals may not be
known exactly at a given time because marked animals are lost through
migration and mortality. It may be possible to estimate the number of
marks from sampling data, but these estimates will be subject to
sampling error. I will show how to adjust for this uncertainty when
computing interval estimates of population size. Shortest confidence
intervals for the negative binomial distribution will be compared to
likelihood intervals.
These methods are used to give improved estimates of migrating salmon
smolt numbers from the data of Macdonald \& Smith ({\em Biometrics} 36,
401-417, 1980). The annual sockeye salmon ({\em Oncorhynchus nerka}) smolt
runs from Babine Lake, British Columbia, Canada, have been sampled by
mark-recapture in a consistent way from 1961 to the present, giving a
unique long-term data set.
Monday, November 21
- Probability Seminar:
"Capacities of uniform random sets" by Thomas S. Salisbury. 3:30 p.m. in
N638 Ross.
Abstract: Consider a strongly subadditive finite capacity on the
unit interval, which doesn't charge points. Split the interval into
equal parts, and form a random set by including each interval
independently with fixed probability. I'll show, using the FKG inequality,
that the capacity of this set converges in probability to a constant, as
the number of parts goes to infinity. In other words, through a strongly
subadditive pair of glasses, the random set looks nearly deterministic. I'll
also try to convince you that this is not true if our glasses are only
subadditive, making an interesting distinction between the two
properties. I'll try to relate this to some recent work on a set
theoretic problem of Juris Steprans.
Friday, November 25
- Statistics Seminar:
Professor D. L. McLeish (Department of Statistics and Actuarial Science,
University of Waterloo) will speak on
"How to Simulate if you must; Projection, Variance Reduction
and Sensitivity Analysis"
Time: 10:30 a.m.,
Place: N638 Ross.
ABSTRACT
Many stochastic systems of interest to practitioners are sufficiently
complex that they cannot be studied analytically. Even in cases where
a problem may be analytically tractable, it is often of importance to
study the solution under some perturbation of the model, usual
incorporating stochastic components. In such cases, Monte Carlo
simulation is frequently used. Applications of simulation abound,
from problems involving the assessment of estimators and tests (e.g.
the bootstrap), to obtaining conditional distributions, optimizing
systems (e.g. Gibbs sampling and simulated annealing).
Unfortunately, reported simulations often suffer from a lack of
extendibility. What does a simulation at one value of a parameter
enable you to say about other values? A simulation at a single value of
a parameter $~ \theta_0 ~$ does, in fact, carry information on the
expected value of the quantity of interest $~T(X)~$, say, not only for
that value of the parameter. Information is available on the whole
function $~m(\theta) = E_\theta[T(X)]$ and all of its derivatives.
Such information can be extracted using importance sampling
techniques. This involves conducting the simulation at some value of
the parameter {\em other than the one in which our interest lies}.
Unfortunately, sample means obtained from importance sampling are often
very noisy and variance reduction is required to make this a useful
technique.
In this talk, we review some variance reduction techniques and show
how they are all essentially a projection. A particular example
corresponds to conducting the simulation conditionally on the score
function for the input variables. We show how this leads to
substantial reduction in variance both for estimating the expected
value$~E_\theta[T(X)]$ and its derivative.
- Seminar in History and Philosophy of Mathematics and Mathematics
Education:
Emeritus Professor Abe Shenitzer (York University) will give a talk on
REMARKS ON THE EVOLUTION OF SOME IMPORTANT CONCEPTS AND/OR AREAS OF
MATHEMATICS at 2:00 p.m. in N501 Ross.
Abstract
Integration goes back to the ancient Greek problem of quadrature. A
modern high point of this theory is found in the work of Lebesgue.
Geometric transformations were used to great effect by Apollonius. A
modern high point of this idea is found in Klein's Erlangen Program.
Gauss's notion of congruence modulo n is the basis of numerous
fundamental equivalence relations throughout mathematics and of quotient
spaces.
Topology was launched by Riemann and greatly advanced by Poincare'.
The theory of functions of a complex variable goes back to Euler. A
high point in the evolution of this subject was the proof of the Klein-
Poincare'-Kobe uniformization theorem intimately related to the topic of
automorphic functions.
Differential geometry became a full-fledged theory with the publication
of Gauss's paper on curves and surfaces. It was greatly advanced by Riemann.
In this talk we'll try to shed some light on (some of) these topics.
- Analysis Seminar: Professor Man-Duen Choi (University of
Toronto) at 3:30 p.m. in N638R.
The title of his talk is "TWO BY TWO MATRICES MADE DIFFICULT".
Abstract: These may provide true illustrations for some not easy results in
operator theory.
Friday, December 2
- Seminar in the History and Philosophy of Mathematics and
Mathematics Education:
Professor Vladimir Tikhomirov (Moscow State University) will give a talk
on "KOLMOGOROV: THE PERSON AND HIS WORK" at 2:00 - 3:00 p.m. in N501 Ross.
ABSTRACT
Professor Tikhomirov is a noted Russian mathematician whose primary
research interests are approximation theory and optimal control. He is also
interested in the historical aspects of mathematics and, as is proved by
his _Stories About Maxima and Minima_ (AMS/MAA 1990), is a master of
mathematical exposition aimed at a broad mathematical public. Given all these
gifts and the fact that he was a student of Kolmogorov's, Professor Tikhomirov
is certain to teach us a great deal about the work and personality of his
famous teacher.
Thursay, December 8
- Colloquium: Prof. Tikhomirov "Survey of the theory of
extrema".
Monday, May 1, 1995
Thursday, May 4, 1995
- Survey Paper for M.A. Degree: Jabina Ng will speak on "Markov chains applied to an Inventory Control Problem"
on Thursday, May 4, 1995 at 11:00 a.m. in N538 Ross.
Friday, May 5, 1995
- Probability seminar: Professor Ildar Ibragimov will give a probability seminar on May 5 at 10:30
in N638 Ross entitled "Bahadur efficiency in estimation theory".
Tuesday, May 9, 1995
- Meeting of the Statistics Section, 3:00 p.m., N627 Ross. See agenda.
Thursday, May 11 1995
- Canadian Undergraduate Mathematics Conference: May 11-14, 1995, U of Manitoba.
See Description.
Friday, May 19, 1995
- Seminar Requirement for M.A. Degree: Angela Amoroso will give a talk entitled
"An Introduction to Regression Graphics" at 10:30a.m. in N638 Ross. See abstract.
- Seminar Talk: Xingfu Zou will speak on "Allee Effect And Hopf Bifurcation
In Single Species Population Over A Patch Environment" at 3:00p.m. in N638 Ross.
See abstract.
Monday, May 29, 1995
- Seminar Requirement for M.A. Degree: Jianrong Wu will speak on
"Approximations to Noncentral Distributions" at 10:00 a.m. in N638 Ross.See abstract.
Friday, June 9, 1995 and Saturday, June 10, 1995
Wednesday, June 21
- Graduate Seminar:
Jianrong Wu will give a talk entitled:
A MODEL OF THE BEHAVIOR OF STOCK PRICES
at 2:00 p.m. in N638 Ross.
See ABSTRACT
Thursday, June 29
- Combined APMA - EATS seminar:
In 317 Petrie, at 2:30pm,
"THE GENERALIZED LAGRANGIAN FORMULATION OF COMPUTATIONAL FLUID DYNAMICS"
Professor W.H. Hui (The Hong Kong University of Science and Technology).
See Abstract.
Friday, June 30
- Dissertation Exam:
Artut Tomita will defend his Ph.D. Dissertation entitled
"COUNTABLE COMPACTNESS AND RELATED PROPERTIES IN GROUPS AND SEMIGROUPS:
FREE ABELIAN GROUPS AND THE WALLACE PROBLEM"
at 2:00 p.m. in S136 Ross.
Thursday, July 6
- Graduate Seminar:
Angela Amoroso will give a talk entitled
"TRIMMED MEAN OR SAMPLE MEDIAN?" at 3:00 p.m. in N638 Ross.
See abstract
Friday, July 7
- Probability Seminar:
Amir Dembo (Stanford University):
"Tail estimates for one-dimensional random walk in random environment"
10:30 a.m. in N638 Ross Building
- Dissertation Exam:
Jiaming Sun will defend his Ph.D. dissertation entitled
"LARGE DEVIATIONS AND EXPONENTIAL CONVERGENCE FOR RANDOM SEQUENCES"
at 1:00 p.m. in S128 Ross.
Wednesday, July 5
- Probability Seminar:
Tetsuya Hattori (Utsunomiya University) and
Kumiko Hattori (University of Tokyo):
"The self-avoiding walk on the Sierpinski gasket"
1:30 - 3:00 p.m., in 2005 Vari Hall. (
The talk will be in two parts: Tetsuya will speak for 40 minutes, then
we'll have a 10 minute break, and then Kumiko will speak for 40 minutes.)
Wednesday, July 12
- Graduate Seminar:
Charles Lam will give a talk entitled
"LOCALLY WEIGHTED LIKELIHOOD ANALYSIS"
at 10:00 a.m., in N638 Ross.
See abstract
From: "Primrose Miranda"
Date: Fri, 1 Sep 1995 15:25:18 -0400
Subject: Dissertation Subject Oral Exam
The Dissertation Subject Oral Exam of Mr. Xingfu Zou entited "Travelling Waves
in Delayed Lattice Differential Equations and Reaction-Diffusion Equations"
will be held on September 6, 1995 at 3:00 p.m. in N638 Ross.
>
>
1.Importance of the project:
>
Reaction-diffusion equations (RDE's) and lattice differential equations
> (LDE's) have found a considerable amount of
> interest in recent years. RDE's arise naturely in a variety of models from
> theoretical physics, chemistry and biology, e.g., the dynamics of nuclear
> reactors,
> chemical reactions in distributed media and combustion theory, ecological
> interactions in spatially distributed populations, motion of bacteria by
> chemotaxis, nerve pulse propagation, models from neurophysiology and
> population genetics.
> LDE's can be viewed as the discretization of
> RDE's along a lattice, and also have been proposed
> as models in various contexts, e.g., for population genetics where spatially
> discrete populations of disploid individuals are considered, and for
> conduction in myelinated nerve anxons.
>
> Among the various aspects of RDE's and LDE's is the phenomenon of traveling
> wave solutions which account for the propagation of the patterns (waves),
> and thus, is of great importance in understanding the established models
> and in
> designing a practical system. Taking systems of excitable cells in
> neurophysiology and cardiophysiology as examples, it is often the case that
> propagation failure leads to the failure of the systems, and in the case of
> the cardiac action potential, this can be fatal.
>
> Most of the works so far on traveling waves for RDE's and LDE's ignore the
> presence of time delay in reaction terms. But time delays always
> exist in reality. For example, animals need time to grow mature,
> resources once consumed
> take time to recover, and neurons also take time to respond to signals due to
> finite speed of the signals.
>
>
>
2. Previous works:
>
The proposed aim of this dissertation is to investigate the traveling waves
> for $\underline{systems}$ of RDE's and LDE's
> $\underline{with\,\,\, delay.}$
> For scalar RDE without delay, the problem
> has been thoroughly attacked using phase plane techniques: a minimal speed
> has
> been found explicitly in terms of the roots of an algebraic equation of
> degree 2. For scalar RDE with delay, there is an excellent pioneer work by
> K.W. Schaaf [Trans. Amer. Math. Soci. {\bf 302}(1987) 587-615 ]
> which obtains several similar results: the minimal speed is
> proven to exist, even though it can not be expressed explicitly due to the
> transcendence of the characteristic equations caused by time delay.
One of the major technical tools in Schaaf's paper is the phase plane
analysis, but this
> technique can not be applied to systems of RDE's.
> There has been few works on some specific systems of dimension 2
> (without delay), using comparison method, degree approach and Conley index,
> etc.. A newly published monigrapg [ Aizik I. Volpert, Vitaly A. Volpert and
> Vladimir A. Volpert: Traveling Wave Solutions of Parabolic Systems ]
> systematically attacks this problem
> for general systems without delay, by introducing a modification of
> Leray-Schauder degree. Taking into account the
> time delay, there has been no work on the existence
> and/or non-existence of traveling wave solutions of systems of RDE's
and LDE's,
> to the best of my knowledge. Moreover,the stability of RDE's and LDE's
has not been attacked at all.
>
> As far as LDE's are concerned, the corresponding wave equations become
> functional differential equations of mixed type for which no fundamental
> theory
> is available yet, and again, no phase plane can be used in this case. Thus
> LDE's are actually much harder to deal with. Little work has been done on
> theexistence of traveling wave solutions for LDE's without delay, not to
> mention LDE's in the presence of delay.
>
>
>
3. Aim of the project:
This dissertation plans to study the
> existence and/or
> non-existence and the stability of traveling wave solutions for
> $\underline{systems}$ of RDE's
> and LDE's $\underline{with \,\,\,delay}$. The comparison and monotonicity
> methods, and the topological degrees will be developed to tackle
the problem, and the applications of the general theory to
some physical systems and
> population systems in heterogeneous evironments, and to neural networks etc.
> will be persuaded.
>
From: "A. Ivic Weiss"
Date: Tue, 5 Sep 1995 11:22:40 -0400
Subject: Geometry Seminar
GEOMETRY SEMINAR
Tuesday, September 12, 1995, 4 p.m. N638 Ross
Professor Ludwig Danzer, University of Dortmund
Title: A space filler of Euclidean 3-space not tiling space
periodically or quasi-periodically
From: Elliott Pearl
Subject: Topology Seminar
Date: Tue, 5 Sep 95 13:33:22 EDT
Monday, September 11, 1995
at 12 noon in N638 Ross
Ulrich Heckmanns (Mathematisches Institut der Universitaet Muenchen)
On the topology of partial-order valued ultrametric spaces.
This year the Topology seminar will usually be held Fridays at noon.
All are welcome to attend.
The seminar's WWW home page can be found at the department's home page,
http://www.math.yorku.ca/, by selecting Seminars and Colloquia, then Topology.
From: Elliott Pearl
Subject: Topology seminars
Date: Thu, 14 Sep 95 21:34:02 EDT
Friday, September 22, 1995 at noon in N638R
George Delistathis (York University)
A regular space with a countable network and different dimensions, pt. 1.
Our speaker will continue the next week here.
The speaker will also be giving two talks at UofT on the same subject.
The York talks will cover the easier points of the construction, while the
UofT talks will cover, independently, the more difficult and lengthy bits.
From: "Primrose Miranda"
To: grads@mathstat.yorku.ca, faculty@mathstat.yorku.ca
Subject: Seminar requirement for M.A. degree
CHARLES LAM
will be giving a talk entitled
COMPARISONS OF APPROXIMATE CONFIDENCE INTERVALS
FOR DISTRIBUTIONS USED IN LIFE-DATA ANALYSIS
on Thursday, September 21 at 10:00 a.m. in N638 Ross
ABSTRACT
In this seminar, the accuracy of approximate confidence intervals for
parameters and quantiles of the small extreme value and normal distributions
will be discussed. The interval estimates are based on (a) the asymptotic
normality of the maximum likelihood estimator (b) the asymptotic X squared
distribution of the likelihood ratio (LR) statistic, (c) a mean and correction
to the signed square roots of the LR statistic, and (d) the Bartlett correction
to the Bartlett correction to the LR statistic. Since these methods also apply
to the Weibull and the lognormal distributions, an example will be given in the
lognormal distribution. Finally, a study design is also provided for comparing
all these methods for estimating the confidence intervals.
MASTERS STATISTICS STUDENTS ARE EXPECTED TO ATTEND THE SEMINAR
From: Man Wah Wong
Subject: Analysis Seminar
Date: Mon, 18 Sep 95 22:32:15 EDT
Speaker: Gheorghe Mihai Iancu
(York University)
Title: Global Solutions of Semilinear Evolution Equations in Hilbert Spaces
Date: Thursday, September 21, 1995
Time: 3:00 p.m.
Place: N638 Ross
From: "Huaxing Xia"
Date: Mon, 18 Sep 1995 22:51:21 -0400
Subject: Dynamical Systems Seminar
The following are the abstracts of the two talks which will
given on Tuesday, September 26, 1996, THE FIRST DAY OF
DYNAMICAL SYSTEMS SEMINAR.
**************************************************************
* Competition in a Chemostat: *
* The Effect of Delayed Response in Growth *
*
*
* Gail S. K. Wolkowicz, McMaster University *
* 10:00-10:50, N638, Ross Bldg. *
***************************************************************
The chemostat, a laboratory apparatus used for the continuous
culture of microorganisms will be described. After first
considering the basic mathematical model for $n$ species of
microorganisms competing exploitatively for a growth-limiting
nutrient, under the simplifying assumption that growth based on
consumption occurs instantaneously, two different approaches that
include the effect of delayed response to growth are considered.
The first approach assumes a discrete time delay and the model
involves functional differential equations and the second approach
assumes a distributed time delay and involves integro-differential
equations. What is known about the global dynamics of the various
models will be discussed and then the implications for applying these
results in ecology will be explored.
---------------------------------------------------------------------
---------------------------------------------------------------------
******************************************************************
* Uniform Persistence and Periodic Coexistence *
* States in Periodic Semiflows with Appliactions *
*
*
* Xiao-Qiang Zhao, Academia Sinica *
* 11:00-11:50, N638, Ross Bldg. *
*******************************************************************
A central problem in population dynamics is to study the asymptotic
behavior of the model systems. Uniform persistence(or permanence)
characterizes the long-term coexistence of interacting species. In this
talk, we will first give a brief review of persistence theory in biological
dynamical systems. Then we will focus on the uniform persistence and
repellors and periodic coexistence in infinite-dimensional periodic
semiflows. Under a general abstract setting, we prove that the uniform
persistence of a periodic semiflow is equivalent to that of its associated
Poincare map, and that the uniform persistence implies the existence
of a periodic coexistence state. This result, as we will illustrate,
also provides a dynamical approach to some static problems. As appliactions,
we will discuss some global behaviors of the periodic Kolmogorov two species
reaction-diffusion systems and, if time permits, a periodic competitor-
competitor-mutualist parabolic system.
Date: Wed, 20 Sep 1995 13:03:54 -0400
From: "Neal Madras"
Subject: Probability seminar
There will be a Probability Seminar on Friday, September 22, at 1:30 p.m.
in N638 Ross Building. The speaker will be V. Statulevicius, who will
speak on "Problems of Large Deviations with Non-normal Approximation"
(joint work with A. Aleskeviciene, both of the Lithuanian Academy of Science).
From: Man Wah Wong
Subject: Analysis Seminar
To: faculty@mathstat.yorku.ca, grads@mathstat.yorku.ca
Date: Mon, 25 Sep 95 0:00:38 EDT
Speaker: Gheorghe Mihai Iancu
(York University)
Title: Asymptotic Behaviour of Global Solutions of Semilinear Evolution
Equations in Hilbert Spaces
Date: Thursday, September 28, 1995
Time: 2:00 p.m.
Place: 225 Bethune College
From: "Debbie Keltz"
Date: Fri, 22 Sep 1995 13:37:08 -0400
To: faculty@mathstat.yorku.ca, grads@mathstat.yorku.ca,
guests@mathstat.yorku.ca, webmaster@mathstat.yorku.ca
Subject: COLLOQUIUM NOTICE
Professor Jason Levy
of
Oklahoma State University
will speak on
"Algebraic Number Theory and the Langlands Programme"
Place: N203 Ross
Date: October 5, 1995
Time: 4:00 p.m. - 5:00 p.m.
ABSTRACT
Number theory has experienced great strides in recent years, most strikingly
with the recent proof of Fermat's Last Theorem by Wiles. Much of this progress
fits into a general framework known as the Langlands program. We will discuss
the progress of algebraic number theory from Gauss's quadratic reciprocity to
class field theory to the Langlands program.
Refreshments will be served in N620 Ross at 3:30 p.m.
From: Nantel Bergeron
Subject: Algebra Seminar
To: faculty@mathstat.yorku.ca (All dept faculty)
Date: Tue, 26 Sep 95 15:14:16 EDT
Everyone is invited to our new Algebra Seminar.
Tentativaly it will be every Monday at 3:30 PM
in the Seminar room N638 Ross.
Our first speakers are:
R. G. Burns and Y. Medvedev (York U.)
Groups Satisfying Positive Laws
Monday October 2, in room 638 North Ross, 3:30 p.m.
Abstract: Every abelian group A satisfies the law xy=yx
(for all x,y in A). This law is called positive (or semigroup)
since no negative power of the variable x, y are involved.
Mal'cev showed that in fact any group which is an extension of a
nilpotent group by a group of finite exponent (that is,
satisfying x^n = 1 for some integer n > 0) satisfies some
positive law u=v where u,v are words in the variables
x, y, z, ... not involving any of the negative powers
x^{-1}, y^{-1}, z^{-1}, ...
QUESTION: Is the converse true: must every group satisfying
a positive law be nilpotent-by-finite exponent?
We describe some progress on this and a related question
of Bergman.
From: "Neal Madras"
Upcoming probability seminar at York University:
Friday, Oct. 6: John Verzani (York University): "IBM, SIBM and IBS"
1:30 p.m. in N638 Ross Building
Abstract: In this talk I will introduce iterated Brownian motion (IBM), the
model for super iterated Brownian motion (SIBM) and a construction of SIBM
via the Iterated Brownian snake (IBS).
From: Man Wah Wong
Subject: Analysis Seminar
To: faculty@mathstat.yorku.ca, grads@mathstat.yorku.ca
Date: Sun, 1 Oct 95 0:20:05 EDT
Speaker : Dr. Jingzhi Tai
(University of Toronto)
Title : The Explicit Solution of the Neumann Problem for the Cauchy-Riemann
Operator on the Non-Isotropic Siegel Domain
Date : Friday, October 6, 1995
Time : 4:30 p.m.
Place : N638 Ross
Abstract: The Neumann problem for the Cauchy-Riemann operator on (0,q) forms
in the strictly pseudoconvex non-isotropic Siegel domain is solved. The metric
used on this domain is invariant under the action of the Heisenberg group in
which the non-isotropic dilations are built in. The fundamental solution of
the related differential equation is derived using the Laguerre calculus. An
explicit formula for the kernel of the Neumann operator is given.The solution
of the corresponding heat equation is constructed by calculating the heat
kernel. The fundamental solution of the Laplace operator on the Heisenberg
group is also obtained.
From: "Xiaomin Dong"
Subject: Category Seminar
Title: Span --| Map adjointness in preordered categories
Speaker: Xiaomin Dong
Place: Room N638 Ross Building
Time: 2:30--3:30pm
Date: October 5, 1995
From: "Israel Kleiner"
Seminar in History and Philosophy of Mathematics and Mathematics
Education
Professor Judy Pelham, Department of Philosophy, York University
RUSSELL'S PROGRESS TOWARD "PRINCIPIA MATHEMATICA"
Date: Friday, October 13, 1995
Time: 3:00-4:00 p.m.
Place: N638 Ross
ABSTRACT
In 1901 Bertrand Russell discovered the famous paradox of the set of
all sets that are not members of themselves. This discovery
presented a formidable obstacle to Russell's plan to prove arithmetic
reducible to logic. For approximately the next five years Russell,
together with Alfred N. Whitehead, worked to find a solution to the
"Contradiction" (as they called it) that did NOT employ a typed
theory of logic. In this talk I explain the origin of this paradox
in Russell's logic, and outline his various attempts to resolve it.
I then go on to discuss in more detail one of Russell's lesser known
attempts which he called "The substitution theory".
From: "David Promislow"
Date: Tue, 3 Oct 1995 11:40:21 -0400
Subject: Automatic group seminar
The Automatic group seminar will be continuing. The time is Wed. 2:00 to 3:00, in N638.
The first meeting will be on Wed. Oct. 11. Bob Burns will be speaking on Braid
Groups.
From: "Debbie Keltz"
Subject: COLLOQUIUM NOTICE
Professor Doris Schuttschneider
of
Bethlehem, Pennsylvania
will speak on
"Ingenious Mathematical Amateurs: M.C. Escher (artist) and
Marjorie Rice (homemaker)"
Place: N203 Ross
Date: October 12, 1995
Time: 4:00 p.m. - 5:00 p.m.
ABSTRACT
It is generally believed in the mathematical community that it is impossible
today for someone without formal credentials in mathematics to engage in
mathematical research or to make any contributions to mathematics. Yet there
are subjects with open problems that need no accumulated mathematical arsenal
to understand and to attack. It is even possible that someone innocent of
training may have fresh insight that leads to fruitful results. I offer two
recent examples to illustrate this phenomenon-one the well-known Dutch graphic
artist, M.C. Escher, and the other an unknown San Diego homemaker, Marjorie
Rice. Each tackled problems that ask which types of shapes can tile the plane,
and in what manner. It is illuminating to see how each made the mathematical
problems their own, asking questions in a way that made sense to them, without
particularly caring how mathematicians might approach the problem. Each
developed an unorthodox notation that was an essential ingredient of their
methodical investigations. Each worked alone, essentially in secret, rewarded
by the exhilaration of finding some answers to a large puzzle. It is
interesting to contrast their questions and methods with those of
mathematicians and scientists who have investigated similar questions. The
professionals can learn something from amateurs.
Refreshments will be served in N620 Ross at 3:30 p.m.
From: Augustine Wong
Subject: Joint Econ & Stats Seminar
Hi,
The second joint Economics & Statistics Seminar for this term will be
held 2 weeks from now:
Speaker: Professor Steven Marron (University of North Carolina)
Title: Band Width Selection
Time: 4:00pm to 5:30pm
Date: Wednesday, October 18, 1995
Place: 1063 Vari Hall
Hope to see you there.
From: Augustine Wong
Subject: Stats Seminar Oct 13
Hi,
Finally we will have our first statistics seminar for this
term on Oct 13, 1995 (next Friday). Our speaker is Professor
Georges Monette. Enclosed is the latex file of the abstract
and the detailed date, time and place of the seminar.
Hope to see you there.
In the mean tiem, I would invite all faculty members/visitors/graduate
students to contribure a talk or to suggest a speaker. Please
let me know as soon as possible.
Professor Georges Monette
Department of Mathematics and Statistics
will speak on
"Is there a role for statisticians in pay equity?"
Friday, October 13, 1995, 10:30 a.m., N638 Ross
ABSTRACT
In studying a salary structure for the purposes of pay equity, the
statistician is expected to help produce gender-neutral measures of the
value of jobs and to estimate the shape and size of the gap between
salaries of women and men for work of comparable value.
The statistician is expected to provide an "objective and unbiased"
component to a very complex process that is otherwise fraught with
potential biases and subjective elements.
Unfortunately, when real salary data are analyzed with modern
statistical tools, they reveal many features that produce more
questions than answers. As statisticians, we feel we must have
an important role to play but we often face the dilemma that what
we can offer is very different from what is expected.
In this talk I will discuss some of my experiences analyzing
real salary data. I will look at a number of questions arising
in pay equity, such as:
gender-neutral job value: how can statistics help?
the importance of understanding basic regression
diagnostic plots and their often troubling revelations
factor analysis and regression: a problematic interaction
visualizing conditional effects: presenting 2-dimensional pictures of higher-dimensional phenomena
From: Man Wah Wong
Subject: Colloquium
Professor Y. T. Chan
Department of Electrical and Computer Engineering
Royal Military College, Kingston
will speak on
THE WAVELET TRANSFORM
Place: N638 Ross Building
Date: Friday, October 13, 1995
Time: 1:30 p.m.
ABSTRACT
Transformation of a signal from one domain to another can often reveal
properties that are not apparent in the original domain. For example, the
Fourier transform (FT) finds the spectral properties of a function of time
and has a wide use in sciences and engineering. When the spectrum of a
function changes with time, it is necessary to keep track of the changes
and the short time Fourier transform (STFT) is an adaptation of the FT for
this purpose.
The recent introduction of the wavelet transform (WT) provides an
alternative to the STFT for time-frequency analysis. Its major difference
from the STFT is in the dilation of a function of time before
transformation. It can also provide a constant-Q analysis where the ratio
of the frequency resolution to the centre frequency is a constant.
I'll discuss the principles and properties of WT, its relationship to
multiresolution analysis, and conclude with some application examples in
electrical engineering.
Refreshments will be served in N620 Ross Building at 2:30 p.m.
Please make note of the changes in day, time, place of the talk and
refreshments.
From: Nantel Bergeron
Everyone is invited to our Algebra Seminar
in the Seminar room N638 Ross.
Our speaker:
Y. Medvedev (York U.)
Groups Satisfying Positive Laws (Part II)
Monday October 16, in room 638 North Ross
ABSTRACT
We continue to talk on groups satisfying a positive law.
We will discuss how one could come to the main conjecture on
a structure of groups satisfying a positive law.
We talk on relations with Rings theory, Lie theory, results
of Gromov-Milnor on growth functions in groups, local and global
theorems for groups, and Engel groups.
From: "Neal Madras"
Upcoming probability seminar at York University:
Friday, Oct. 20: Gleb Haynatzki (McMaster University): "A Birth and a Death
Process in a Random Environment" (see abstract below)
1:30 p.m. in N638 Ross Building
Thursday, Nov. 16: Murad S. Taqqu (Boston University): "On-Off Models for
Generating Long-Range Dependence" (see abstract below)
4:00 p.m. location T.B.A.
(Department Colloquium; refreshments at 3:30 p.m. in N620 Ross)
ABSTRACT for G. Haynatzki:
The linear death and the linear birth processes with constant
parameters are classical models in the theory of continuous time Markov
chains. Here we shall consider these processes in a "random environment".
This term designates the situation when a stochastic process Y(t)
developing independently influences a second process X(t), which is
dependent on Y(t); in most cases this means that the independent process
is in fact a parameter of the dependent one.
To solve the above two problems, we first derive the Kolmogorov
forward equations for the state probabilities of the processes. Then we
transform these equations to equations for the Laplace transforms and
represent them in matrix form. In the case of the death process, we solve
the equations recursively so that the solution is in an easily inverted form.
In the case of the birth process, we use the p.g.f. approach and derive a
formal solution in terms of Laplace transforms.
ABSTRACT for M. Taqqu:
Consider an alternating renewal process whose interrenewal times have
an infinite variance. A reward, proportional to the length of a
renewal, is offered. The constant of proportionality alternates between
1 ("on") and 0 ("off"). That constant of proportionality (reward)
is thus 1 throughout a renewal, then 0 throughout the next
renewal, then 1 again, and so on. The aggregation of many such
renewal reward processes converges in distribution, after time
renormalization, to fractional Brownian motion, a Gaussian process with
long-range dependence. We will explain why this is so.
From: "Debbie Keltz"
Subject: Colloquium Notice
Professor Dieter Pumpluen
of
Fern Universitat and York University
will speak on
Superconvex Sets and Banach Spaces
Place: N203 Ross Building
Date: October 19, 1995
Time: 4:00 p.m. - 5:00 p.m.
ABSTRACT
Countably convex or superconvex sets in linear topological spaces are a natural
generalization of convex sets. There are numerous examples of superconvex sets
in mathematics. For investigating the structure of superconvex sets one
intruduces superconvex modules, i.e. sets admitting countably convex
combinations but which are not necessarily subsets of linear spaces. There
exists a close connection between this theory and the theory of base normed
Banach spaces.
Refreshments will be served in N620 Ross at 3:30 p.m.
From: Elliott Pearl
Subject: topology seminar
Friday, October 20, 1995 at 12 noon in N638
Camilo Costantini (Univ. Torino) speaks on
some generalizations of compact and totally bounded metric spaces.
From: "Israel Kleiner"
Subject: Seminar in History and Philosophy of Mathematics and Mathematics Education
Professor Israel Kleiner
Department of Mathematics and Statistics, York University
will give a talk on
THE GENESIS OF THE ABSTRACT RING CONCEPT
Date: Friday, October 27, 1995
Time: 3:00-4:00 p.m.
Place: N638 Ross Bldg.
ABSTRACT
The notion of a ring emerged as a central concept of algebra in the early decades of the 20th century,
but its origins date back to the first half of the 19th century. I will describe highlights of this
evolving process.
From: "Huaxing Xia"
Subject: Second Day of Dynamical Systems Seminar
Dear Colleages:
On October 31 (Tuesday), the last day of this month, we will have
our second meeting of the dynamical systems and applications seminar. There
will be two talks in the morning, followed by
no-host lunch and informal discussions in the afternoon. Professor Bill
Langford, the Chair of the
Department of Mathematics and Statistics of University of Guelph, will first
speak from 10:10-11:00am
on some recent generalizations and applications of the Hopf bifurcation
theorem. From 11:10-12:00am
Mr. Xingfu Zou of York will talk on convergence (global stability) results in
analog neural network models.
The seminar will take place at Ross S125. We look forward to seeing
you there.
From: Nantel Bergeron
Everyone is invited to our Algebra Seminar
in the Seminar room N638 Ross.
Our speaker:
H. D. Pumpluen (Visiting York U.)
Convex Modules
Monday October 23, in room 638 North Ross, 3:30 p.m.
Abstract: There are many types of convex sets, which are used in
mathematics, as e.g. convex, absolutely convex, (countably) convex,
totally convex sets etc. For any of these types of convex sets there
is a corresponding type of convex module, which is a canonical
generalization of the respective classical case. Convex modules of
certain types are used in economics and physics. Any of the above
mentioned types of convex modules has a canonical and close
connection to types of linear spaces used in functional analysis.
In the two lectures at first a short sketch about the development of
the theory of convex modules is given. Then selected results on the
connection to functional analysis and structure theorems are discussed.
P.S. The first lecture is the Coloquium Talk.
From: "Nantel Bergeron"
Subject: Algebra Seminar: D. Pumplum (Part II)
Everyone is invited to our Algebra Seminar
in the Seminar room N638 Ross.
Our speaker:
H. D. Pumplum (Visiting York U.)
Convex Modules (Part II)
Monday October 30, 3:30 PM in room 638 North Ross
Abstract: There are many types of convex sets, which are used in
mathematics, as e.g. convex, absolutely convex, (countably) convex,
totally convex sets etc. For any of these types of convex sets there
is a corresponding type of convex module, which is a canonical
generalization of the respective classical case. Convex modules of
certain types are used in economics and physics. Any of the above
mentioned types of convex modules has a canonical and close
connection to types of linear spaces used in functional analysis.
In the two lectures at first a short sketch about the development of
the theory of convex modules is given. Then selected results on the
connection to functional analysis and structure theorems are discussed.
The (Part I) was Monday Oct. 23, In Part II some example will be worked out.
Wednesday, November 1, 1995
- Automatic Group Seminar: Robert Burns (York University)
will continue to speak on "Braid Groups" at 2:30 p.m. in N638 Ross.
Thursday, November 2, 1995
- Departmental Colloquium: S.T. Ali (Concordia University) will speak on "Coherent States:
Mathematical Aspects and Physical Applications" at 4:00 p.m. in N203
Ross. See abstract.
Friday, November 3, 1995
- Topology Seminar: Camillo Costantini (University of Torino)
will continue to speak on "Some Generalizations of Compact and
Totally Bounded Metric Spaces" at 12:00 noon in N638 Ross.
- Probability Seminar: Jeffrey S. Rosenthal (University of
Toronto) will speak on "Optimal Scaling for Langevin Algorithms" at
1:30 p.m. in N638 Ross.
See abstract.
Monday, November 6, 1995
- Algebra Seminar: J. McCool (University of Toronto) will
speak on "The Non-finite Presentability of IA(F3) and
GL2(Z[t,t^{-1}])" at 3:30 p.m. in N638 Ross. See abstract.
Wednesday, November 8, 1995
- Automatic Group Seminar: Robert Burns (York University)
will continue to speak on "Braid Groups" at 2:30 p.m. in N638 Ross.
Thursday, November 9, 1995
- Departmental Lecture Series: Joan Wick Pelletier (York
University) will speak on "Involutive Quantales and Applications to
Analysis (Part 1)" at 1:30 p.m. in N638 Ross. See abstract.
Friday, November 10, 1995
- Statistics Seminar: Helene Massam (York University) will
speak on "Covariance Estimation in Lattice Conditional Independence
Models" at 10:30 a.m. in N638 Ross. See abstract.
- Topology Seminar: Camillo Costantini (Univ. Torino) will
continue to speak on "Some generalizations of compact and totally bounded
metric spaces" at 12 noon in N638 Ross.
- Seminar in History and Philosophy of Mathematics and
Mathematics Education: Abe Shenitzer (York University) will speak
on "A Few Expository Briefs" at 3:00 p.m. in N638 Ross. See abstract.
Monday, November 13, 1995
- Probability Seminar: Dana Randall (Institute for Advanced
Study, Princeton University), will speak on "Markov Chain Algorithms
for Planar Lattice Structures" at 1:30 p.m. in N638 Ross. See abstract.
- Algebra Seminar: Krystyna Bugajska (York University) will
speak on "Cyclic Cohomology - Physical Motivations" at 3:30 p.m. in
N638 Ross. See
abstract.
Thursday, November 16, 1995
- Category Seminar: Walter Tholen (York University) will
speak on "Tychonoff's Theorem Sixty Years Later" at 1:00 p.m. in N638
Ross. See abstract.
- Joint Economics and Statistics Seminar: Rene Garcia
(University of Montreal) will speak on "Asymmetric Effects of Monetary
Policy" at 4:00 p.m. in 1154 Vari Hall.
- Departmental Colloquium: Murad S. Taqqu (Boston University)
will speak on "On-Off Models for Generating Long-Range Dependence" at
4:00 p.m. in N203 Ross.
See abstract.
Friday, November 17, 1995
- Analysis Seminar: Gheorghe Mihai Iancu (York University)
will speak on "Semilinear Wave Equations and Weyl Quantizations" at
4:00 p.m. in N638 Ross. See abstract.
Monday, November 20, 1995
- Algebra Seminar: Nantel Bergeron (York University) will
speak on "Descent Algebra I" at 3:30 p.m. in N638 Ross. See abstract.
Thursday, November 23, 1995
- Category Seminar:
Prof. Joan Wick Pelletier (York University) will give the second of a
series of four or five lectures on " Involutive quantales and
applications to analysis" at
1:30 in N638 Ross. Everybody is welcome to attend. See
abstract.
- Departmental Colloquium: Peter C. Greiner (University of
Toronto) will speak on "Hamiltonian Mechanics and Fundamental Solutions
for Subelliptic Operators" at 4:00 p.m. in N203 Ross.
Friday, November 24, 1995
- Statistics Seminar: Robert Tibshirani (University of
Toronto) will speak on "Regression Shrinkage and Selection via the
Lasso" at 10:30 a.m. in N638 Ross. See abstract.
- Probability Seminar: Amram Meir (York University) will
speak on "Degree Distribution in Families of Weighted Rooted Plane
Trees" at 1:30 p.m. in N638 Ross. See abstract.
- Seminar in the History and Philosophy of Mathematics and
Mathematics Education: Kurt Ramskov (University of Aahrus) will
speak on "The Forgotten Bohr" at 3:00 p.m. in N638 Ross. See abstract.
- Analysis Seminar: Sid Scull (York University, Atkinson
College) will speak on "Group Representations and Wavelets" at 4:00
p.m. in N638 Ross. See
abstract.
Monday, November 27, 1995
- Probability Seminar: Valentin Petrov (St. Petersburg
University, University of Guelph) will speak on "On Strong Limit
Theorems for Sums of Random Variables" at 1:30 p.m. in N638 Ross. See abstract.
- Algebra Seminar: Nantel Bergeron (York University) will
speak on "Descent Algebra II" at 3:30 p.m. in N638 Ross. See abstract.
Friday, December 1, 1995
- Analysis Seminar:
Professor Sid Scull (York University) will give the second of his
series of talks entitled
"Group Representations and Wavelets"
at 4:00 p.m. in N638 Ross Building.
See abstract.
Wednesday, December 6, 1995
- Statistics Section Meeting:
There will be a meeting to choose a Hiring Committee
at 10:30 a.m.in N627R.
See message.
- Women's Remembrance Day:
No classes between 11:30 and 1:30 p.m.
- Last day of classes:
for Terms F and Y - except those which
normally meet on Mondays or on Wednesdays between 11:30 and 1:30.
Thursday, December 7, 1995
- Classes
normally scheduled for Mondays in Terms F, Y will be held on this day.
- Ph.D. Dissertation Colloquium:
George Delistathis will speak on
"A Regular Space with a Countable Network and Different
Dimensions -- Compact Spaces with Few Regular Open Baire Subsets"
at 12:00 noon in 105 Admin. Studies Building.
See abstract.
Friday, December 8, 1995
- Classes:
normally scheduled between 11:30 and 1:30 p.m. on
Wednesdays in Terms F and Y will be held on this day.
- Deffered: The Restructuring Committee planned for
today at 9:30 will be held, tentatively, at 2 p.m. on Tuesday,
December 12.
See notice.
- Dynamical Systems Seminar:
Sue Ann Campbell (University of Waterloo) will speak on
"Resonant Hopf Birfurcation in a Delay Differential Equation"
at 10:00 a.m. in 300 Administrative Studies Building.
See abstract.
- Dynamical Systems Seminar:
Marc Chamberland (McMaster University) will speak on
"A Dynamical Systems Approach to the 3x+1 Problem"
at 11:00 a.m. in 300 ASB.
See abstract.
- Topology Seminar:
Camillo Costantini (Univ. Torino) continues to speak on
generalized compactness at 12:00 noon in N638 Ross.
- Pure Math Curriculum Committee:
will meet
at 2:30 pm in North 627 Ross.
See agenda
- Analysis Seminar:
Professor Sid Scull will give the third and last of his series
of talks entitled "Group Representations and Wavelets" in N638
Ross Building at 4:00 p.m.
See notice.
Monday, December 11, 1995
- Statistical Consulting Service: Bi-weekly coordination
meeting at 10:30 a.m. in 262 Admin. Studies Bldg.
- December Department Party: starts at 1 p.m.
- Examinations: first day of examinations and
last day to submit term work.
Tuesday, December 12, 1995
- Restructuring Committee: Tentative meeting 2 p.m. in
N627 Ross.
See notice.
Friday, December 15, 1995
- Statistics Section:
Meeting of the Statistics Section from
9:30 to 11:00 am in N627R.
See agenda.
Tuesday, December 19, 1995
- Applied Mathematics Curriculum Committee
will meet at 10 am in Room 258 Petrie.
See agenda.
Thursday, December 21, 1995
Thursday, January 4, 1996
- Restructuring Committee:
Retreat to discuss: "THE FUTURE OF MATHEMATICS AND STATISTICS AT YORK"
from 9:30 am to 4:OO pm, Room 305 York Lanes.
See agenda.
Monday, January 8, 1996
Wednesday, January 10, 1996
- Meeting of Statistics Section:
There will be a meeting of the Statistics Section from
12:30 to 1:30 p.m. in N627R.
See agenda.
- Graduate Seminar: Matthew Elmslie will speak on
"Ramsey Numbers" at 2:30 p.m. in N638 Ross.
See abstract.
Thursday, January 11, 1996
- Colloquium:
Professor John C. Stillwell of the
Department of Mathematics,
Monash University, Australia,
will speak on
"WHAT IS CALCULUS?"
at 4:00 p.m. in N203 Ross Building.
See abstract.
Friday, January 12, 1996
- Graduate Seminar:
Christopher Benson will give a talk entitled
"A Proof of Brown's Algorithm for Solving a Two-Person Game"
at 2:00 p.m. in N638 Ross
See abstract.
-
Analysis Seminar:
Man Wah Wong will speak on sampling theorems and signal
analysis at 4:00 pm in N638 Ross.
See proposal for seminar.
Friday, January 19, 1996
- Statistics Seminar: Professor Eli Posner,
Department of Statistics,
University of Toronto, will speak on
"Nonparametric Regression for Arbitrary Random Processes
with Applications to Nonlinear System Output Prediction,"
at 10:30 a.m. in N638 Ross. See
abstract.
- Topology Seminar: Xiao-Qiang Zhao,
Institute of Applied Mathematics, Academia Sinica, Beijing, China,
will talk on "The Theory of Monotone Dynamical Systems," at 12:30 p.m.
in N638 Ross. See abstract
- Seminar in History & Philosophy of Mathematics and Math Education:
Professor John Stillwell, Department of Mathematics, Monash University,
Melbourne, Australia, will give a talk entitled
"From Pythagorean Triples to Eulerian Quadruples," at 2:45
(PLEASE NOTE TIME CHANGE) in N638 Ross.
See abstract.
- Analysis Seminar: Professor Man Wah Wong, York University,
will continue to talk on
"Signals, Wavelets and Localization Operators," (joint work with
Zhiping He) in N638 Ross.
See abstract
Saturday, January 20, 1996
- Classes:
Last day for registered students to enrol in Term W
courses without the approval of the course director.
Tuesday, January 23, 1996
- Department meeting: will be held at 4 p.m. in S173 Ross.
See Council Home Page
for agenda and minutes of last meeting.
Wednesday, January 24, 1996
Friday, January 26, 1996
- Graduate Social: At 1 p.m. in N620 Ross.
See poem.
There will be cards, drink (and not just your regular kind)
and food to spare, so bring the kids and have some fun.
We are looking forward to seeing as many of you as possible!
We go till we drop -- so come and join in!
- Classes:
Last day to enrol in Term W courses with the approval of
the course director.
- Statistics Seminar: Professor S.R. Chamberlin, York University,
will speak on "The Third Property of Mathematical Likelihood"
at 10:30 a.m. in N638 Ross.
See abstract.
- Topology Seminar: Alexander V. Arhangelskii,
Ohio University and Moscow State University,
will speak on
"Relative topological properties: some results and problems"
at 12:30 p.m. in N638 Ross.
- Ph.D. Defense: George Delistathis
will defend his Ph.D. dissertation entitled
"A Regular Space with a Countable Network and Different Dimensions -
Compact Spaces with Few Regular Open Baire Subsets,"
at 2:15 p.m. in S130 Ross. Anyone interested is welcome to
attend.
Tuesday, January 30, 1996
- Fourth Dynamical Systems Day: Professor Raymond Kapral,
University of Toronto, will talk on
"Chaotic Chemical Fronts,"
at 10:10 a.m. in N638 Ross.
See abstract.
- Fourth Dynamical Systems Day: Professor Kaijun Zhan,
University of Guelph, will talk on
"Generalized Hopf Bifurcations and
Vortex-induced Vibration Problems,"
at 11:10 a.m. in N638 Ross.
See abstract.
Thursday, February 1, 1996
- Computer Announcement: The Mathstat machines will go down
briefly around 5 p.m. to allow the installation of a new CD-ROM.
Friday, February 2, 1996
- Topology Seminar: Jianping Zhu
will speak on "Q-points and composants of $H^*$".
at 12:30 p.m. in N638 Ross.
- Computer Science Distinguished Lecture Series:
Bart Selman,
AT&T Bell Laboratories, will give a talk entitled
"Stochastic Search and Phase Transitions:
Artificial Intelligence Meets Physics," at 1:30 p.m. in CCB115.
See abstract.
- Graduate Seminar:
Eng Ong will give a talk entitled
"Lidstone's Theorem" at 2:00 p.m. in S536 Ross.
See abstract.
- Pure Math Curriculum Committee: will meet at 2:30 p.m.
in N627 Ross.See agenda.
- Seminar in History & Philosophy of Mathematics & Math. Educ:
Professor P Rajagopal,
Department of Mathematics & Computer Science,
Atkinson College, York University,
will give a talk on
"Indian Mathematics in the 1850s," at 2:45 in N638 Ross.
See abstract.
- Colloquium:
Professor Mark A. Kon,
Department of Mathematics,
Boston University,
will give a talk entitled
"Can we recover a function from the maxima of its wavelet expansion?"
at 4 p.m. in N638 Ross. See abstract.
Tuesday, February 6, 1996
- Applied Math/Probability Seminar:
Chris Soteros, University of Saskatchewan, will talk about
"Collapsing Lattice Animals" at 1:30 p.m. in N638 Ross.
See abstract.
Thursday, February 8, 1996
- Colloquium:
Professor H. Freedman of the University of Alberta will speak
on "Mathematical Models describing interactions between
Agriculture, Industry and Environment," at 4:00 p.m. in N203R.
See abstract.
Friday, February 9, 1996
- Topology Seminar: Jianping Zhu will continue to speak on
"Q-points and composants of H^*" at 12:30 p.m. in N638 Ross.
- Graduate Seminar: Tina Haller will talk on
"Clinical Trials" at 2:00 p.m. in S536 Ross.
See abstract.
- Algebra/Category Seminar: Walter Tholen will give a talk
entitled "Separation versus Connectedness," at 2:30 p.m. in
N637 Ross.
See abstract.
- Analysis Seminar:
Zhiping He of York University will give the first of a series of lectures
entitled "An Introduction to the Continuous Wavelet Transform"
at 4:00 p.m. in N638 Ross.
See abstract.
Saturday, February 10, 1996
- Classes:
Last day to withdraw from Term Y courses without
receiving a grade.
Wednesday, February 14, 1996
- Executive Committee:
meets from 12:30 to 2:00 in N627 Ross.
Friday, February 16, 1996
- Graduate Social: celebrating "Valentine's Day" at
1:00 pm in N620 Ross. Refreshments will be served.
- Topology Seminar: Xiao-Qiang Zhao will resume a series of
talks on "The Theory of Monotone Dynamical Systems," at 12:30 p.m.
in N638 Ross. See abstract
- Algebra/Category Seminar: Walter Tholen will give part II
of a talk entitled "Separation versus Connectedness" at
2:30 pm in N638 Ross.
See abstract.
- Analysis Seminar: Zhiping He of York University will give
the second of the lecture series entitled "An Introduction to
the Continuous Wavelet Transform" at 4:00 p.m. in N638 Ross
Building.
Monday, February 19, 1996
- Visit: Prof. David Brillinger visits from Berkeley as
an appraiser for the OCGS review of the graduate programmes.
He will stay two days. See message
for further information or contact the Director of the Graduate
Programme,
David.Promislow@mathstat.yorku.ca.
Tuesday, February 20, 1996
- Department meeting: *** postponed possibly to March 5 ***.
See message.
Thursday, February 22, 1996
- Colloquium:Krzysztof Ciesielski (West Virginia University)
will give a talk on "Set Theoretic Analysis" at 4:00 p.m. in N203 Ross.
See abstract.
Cookies and coffee will be available from 3:30 in the lounge.
Friday, February 23, 1996
- Topology Seminar:
Xiao-Qiang Zhao will continue a series of
talks on "The Theory of Monotone Dynamical Systems," at 12:30 p.m.
in N638 Ross. See abstract
- Algebra Seminar:Professor Trueman Machenry will speak
on "A subgroup of the group of multiplicative arithmetic
functions," at 2:30 pm in N638 Ross. See
abstract.
- Analysis Seminar:Zhiping He of York University will give the
third (also the last) of the first series of lectures entitled "An
Introduction to the Continuous Wavelet Transform" at 4:00 p.m.
in N638 Ross Building.
Monday, February 26, 1996
- First day of Reading Week
Friday, March 1, 1996
Monday, March 4, 1996
- Statistics Seminar: Dr. Yodit Seifu will give a talk
entitled
"Local Lyapunov Exponents: Results and Applications,"
at 11:30 in N 638 Ross. (Dr. Seifu is a candidate for a
position in Statistics)
See abstract.
Tuesday, March 5, 1996
- Probability Seminar:
John Verzani (York University) will give a talk entitled
"Some Path Properties of the Planar Brownian Snake,"
at 1:30 p.m. in N638 Ross.
See abstract.
- Physics and Astronomy Seminar:
E. J. Janse van Rensburg
will give a talk entitled
" Monte Carlo Study of the Theta Point in Collapsing Trees"
at 3:30 in Room 317, Petrie.
(Refreshments will be served in Petrie 315 at 3:00 p.m.)
See abstract.
- Department meeting: will be held in S173 Ross from
4 pm to 6 pm. See
agenda and
minutes.
Wednesday, March 6, 1996
- Graduate appraisal: First day of a two day visit by
Professor Angelo Mingarelli of Carleton.
See announcement.
- Statistics Seminar: Mr. Peter Xue-Kun Song will give a talk
entitled
"Stationary Time-Series Models With Exponential Dispersion Model Margins,"
at 11:30 in N 638 Ross. (Mr. Song is a candidate for a
position in Statistics)
See abstract.
Friday, March 8, 1996
- Statistics Seminar: Dr. Michael Xiaopeng Dong will give a talk
entitled
"Estimating Density Functions: A Constrained
Maximum Likelihood Approach,"
at 11:30 in N 638 Ross. (Dr. Dong is a candidate for a
position in Statistics).
See abstract
- Algebra Seminar:
Professor Trueman Machenry will give a second talk entitled
"On a subgroup of the group of multiplicative
arithmetic functions,"
at 1:45 pm in N638Ross.
See abstract.
- History and Philosophy of Math. and Math. Education:
Professor V. Tikhomirov,
Moscow State University & Mathematical College of Moscow
Independent University,
will give a talk on
"The History of Functional Analysis" at 2:45 pm in N638 Ross.
See abstract.
- Dissertation subject oral:
Ying Zhang
will speak on
"Resonance problems for semilinear pseudo-differential operators,"
at 3 p.m. in Room 1154 Vari Hall.
Saturday, March 9, 1996
- Classes:
Last day to withdraw from Term W courses without receiving a grade.
Monday, March 11, 1996
- Statistics Seminar: Dr. Joan Xiaoquiong Hu will give a talk
entitled
"Analyses of Data from Post-Market Studies,"
at 11:30 in N 638 Ross. (Dr. Hu is a candidate for a
position in Statistics).
See abstract.
Friday, March 15, 1996
- Graduate Seminar:
Tilaye Teshager will give a talk entitled "An application of the
Principle of Maximum Entropy to the study of an M/M/1 queueing system."
at 11:15 a.m. in N638 Ross.
- Topology Semianar:
Xiao-Qiang Zhao will continue a series of
talks on "The Theory of Monotone Dynamical Systems," at 12:30 p.m.
in N638 Ross. See abstract
- Social:
The departmental graduate student association will host a social in the
departmental lounge from 1 p.m. All are invited.
- Graduate Seminar:
Constanze Pumpluen will give a talk entitled
"Complete Valuation Rings and P-adic Numbers" at 3:30 in 1154 Vari.
Masters Mathematics students are expected to attend this talk.
Tuesday, March 19, 1996
- Probability Seminar
Vladimair S. Korolyuk (Institute of Mathematics, National Academy of
Science of Ukraine, Kiev) will speak on "Ruin problems: Explicit and
asymptotic approaches" at 1:30 p.m. in N638 Ross.
See abstract.
Wednesday, March 20, 1996
- Statistics Seminar:
Dr. Kalyanee Viraswami of the Department of Statistics at Oxford University
will speak on Higher order results on some robust test statistics under model
misspecification at 11:30 a.m. in N638 Ross.
Thursday, March 21, 1996
Probability Seminar:
Neil O'Connell (BRIMS, Hewlett Packard Labs, Bristol, U.K.)
will speak on "Large deviations for sample paths indexed by the half line"
at 11:30 a.m. in N638 Ross.
See abstract.
Friday, March 22, 1996
- Topology Seminar:
Vl. Todorov of the University of Sofia, Bulgaria will speak on
transitive shift operators at 12:30 p.m. in N638 Ross. Abstract to follow.
- SEMINAR IN HISTORY &PHILOSOPHY OF MATH AND MATH EDUCATION:
Professor Guenther Frei of the Department of Mathematics and
Statistics, University of Laval will give a talk on
"EULER'S CONVENIENT NUMBERS" at 2:45 p.m. in N638 Ross.
See abstract.
- Graduate Seminar:
Eng Ong will speak on "MODELS IN MULTIPLE DECREMENT THEORY"
at 3:30 p.m. in 1154 Vari Hall.
- Graduate Seminar
Constanze Pumpluen will give a talk entitled "HENSEL'S LEMMA"
at 4 p.m. in Ross S175.
- Departmental Colloquium:
Prof. Rongqing Jia of the Department of Mathematical Sciences,
University of Alberta will speak on
"Approximation and Smoothness Properties of Wavelets"
in N638 at 4:00 p.m.
Refreshments will be served in N620 Ross at 3:30 p.m.
See abstract.
Friday, March 29, 1996
- Party: The YGMA hosts a social in the lounge N620 between
1 p.m. and 5 p.m..
- Topology Seminar:
Jianping Zhu will speak on "+Ramsey filters and \alpha_i spaces"
at 12:30 p.m. in N638 Ross.
See abstract.
- Topology Seminar: Prof. Franya Franek (McMaster)
will speak on
"Structural Properties Of Universal Minimal Dynamical Systems For
Discrete Semigroups - the last chapter".
at 1:30 p.m. in N638 Ross.
See abstract.
- Seminar in History & Philosophy of Mathematics & Math. Educ:
Hardy Grant (Emeritus Professor, York University) will give a talk on
"SOME THOUGHTS ON THE HISTORY OF BEAUTY IN MATHEMATICS" at 2:45 p.m.
in N638 Ross.
- Graduate Seminar:
Maurice Tsang will give a talk entitled "Stationary Population Model"
at 3:00 p.m. in 1154 Vari Hall. Masters students are expected to
attend the talk. See abstract way below.
Thursday, April 4, 1996
- First day of Passover
No examinations will be scheduled on the evenings prior to, nor during this
day. However, classes will be held, and university offices will be open
normal hours.
Friday, April 5, 1996
- Second day of Passover
- Good Friday: University closed.
Wednesday, April 10, 1996
- Seventh day of Passover
No examinations will be scheduled on the evenings prior to,
nor during this day.
However, classes will be held, and university offices
will be open normal hours.
Thursday, April 11, 1996
- Eighth day of Passover
No examinations will be scheduled on the evenings prior to,
nor during this day.
However, classes will be held, and university offices
will be open normal hours.
- Colloquium:
Professor Donald St. P. Richards, of the Department of Statistics at the
University of Virginia will speak on "Algebraic Methods towards
Higher-Order Probability Inequalities" at 4 p.m. in N638 Ross.
See abstract.
Friday, April 12, 1996
- Classes: Terms Y, W classes end.
- Statistics Seminar:
Professor Ka-Sing Lau of the Department of Mathematics and Statistics
at the University of Pittsburgh will speak on
"The Discount Sum of Bernoulli Random Variables"
at 10:30 a.m. in N638 Ross.
See abstract.
- Graduate Seminar:
Kam Cheung Ko will give a talk entitled "SOME MATHEMATICAL MODELS FOR THE
UNCERTAINTY INDUCED BY FUZZINESS" 12:00 p.m. in S176 Ross.
See abstract.
- Social: The YGMA hosts an end of term party in the lounge N620,
from 1 p.m. to 4 p.m.. All are welcome to attend.
-
Seminar in History & Philosophy of Mathematics and Mathematics
Education
Professor Trueman MacHenry (York University) will speak on
"COUNTING REVISITED" at 2:45 p.m. in N638 Ross.
Saturday, April 13, 1996
TORONTO SPRING MEETING
A Weekend Meeting on Category Theory and Its Applications
YORK UNIVERSITY, North York (Metropolitan Toronto)
Saturday/Sunday, April 13/14, 1996
Announcement and Information
Monday, April 15, 1996
- Examinations: First day. (Last day to submit term work).
Tuesday, April 16, 1996
- M.A. Thesis Defense
G. Mihai Iancu will speak on
"GLOBAL SOLUTIONS OF SEMILINEAR EVOLUTION EQUATIONS IN HILBERT SPACES"
at 3:00 p.m. in 1158 Vari Hall.
Thursday, April 18, 1996
- Category Theory Seminar:
Professor Maria Cristina Pedicchio (University of Trieste, Italy) will give a
lecture on "Grids and topological systems" at 2:30pm in N638Ross.
- Category Theory Seminar:
Walter Tholen will give a talk on "Separated and connected maps"
at 3:30pm in N638Ross.
- Category Theory Seminar:
Dr. Dorette Pronk (Rijksuniversitet Utrecht and Dalhousie University)
will speak on "Etendues and Stacks" at 4:30 in N638Ross.
Tuesday, April 23, 1996
- Probability Seminar:
George L. O'Brien will speak on "Loosening Tightness" at 1:30 p.m. in N638
Ross.
Abstract:
Conditions are given which are equivalent to uniform tightness in the weak
convergence context and exponential tightness in the large deviation
context. It is shown both by general considerations and by virtue of
examples that these conditions can be much easier to verify than tightness
itself. In one example, Donsker's funtional central limit theorem in
D[0,1] is shown to be a simple consequence of the corresponding theorem
for C[0,1] and that in particular, no detailed consideration of compact
sets in D[0,1] is required. In a second example, a relatively simple
tightness argument is provided for a large deviation principle for a class
of processes including all spectrally positive stable processes.
- Council Meeting:
There will be a meeting of departmental council at 4:00 p.m. in S173R.
Please submit any other agenda
items as soon as possible.
Agenda:
Thursday, April 25, 1996
- Graduate Seminar:
Qunsheng Zhou will give a talk entitled "CONSISTENT ESTIMATES OF SUPER
IMPOSED EXPONENTIAL SIGNALS WHEN SOME OBSERVATIONS ARE MISSING" at
10:00 a.m. in N638 Ross.
This is a seminar requirement for the M.A. degree of the speaker.
All other M.A. Statistics students are required to attend.
Abstract:
Methods are proposed for estimating the parameters of undamped exponential
signals when observations are missing. Some consistency results have been
established. The finite sample behavior of the proposed methods have been
studied by Monte Carlo simulation.
- Graduate Seminar:
Guojie Liu will give a talk entitled "Sinusoidal Signals With Random
Amplitude: Least-Square Estimators and Their Statistical Analysis"
at 11:00 a.m. in 638 Ross.
This is a seminar requirement for the M.A. degree of the speaker.
All other M.A. Statistics students are required to attend.
Abstract:
The asymptotic properties of constrained and unconstrained least-squares
estimate of the parameters of a random amplitude sinusoid are analyzed. An
explicit formula for the asymptotic covariance matrix of the estimation errors
is derived for both the constrained and unconstrained estimators. Accuarcy
aspects are investigated with the following main results. For a certain
weighing matrix, which is shown to be the same for the constrained
and unconstrained methods, the estimation errors acheive their lower bounds. It
is proven that in the optimal case, the constrained method always outperforms
the unconstrained method. It is also proven that the accuarcy of the optimal
estimators improve as the number of least-squares equations increases.
- Category Theory Seminar:
Professor Christopher Gilmour (University of Cape Town) will give a talk on
"Realcompactifications, Lindelo"fications and cozero bases" at 2:30 p.m. in
N638 Ross.
Abstract:
Every Lindelo"f topological space is realcompact and it is well
known that the converse is false. However in the setting of frames, ie.
lattice theoretic generalisations of topologies, there is good evidence
that, at least in some respects, the two concepts
should coincide. We discuss various notions of realcompactness for frames
and the relationships between them. Of particular importance is the role
of the cozero elements. The appropriate setting for spaces is provided by
the Alexandroff spaces and, for frames, their counterpart the regular
sigma-frames. We will indicate some new proofs in frames which extend
classical topological results that relate compactifications,
realcompactifications, pseudocompactness and properties of cozero-bases.
- Category Theory Seminar:
Walter Tholen will speak on "Separated and connected maps" at 3:30 p.m.
in N638 Ross. This talk is rescheduled from last week.
Friday, April 26, 1996
- The Seventh Annual York-McMaster Probability Day
See program.
- Statistics Seminar:
Dr. R. Moeanaddin, of Shiraz University, will speak on
"Conditional Mean and Conditional Variance of Bilinear Models"
at 11:30 a.m. in N638 Ross.
Abstract:
In this article we consider the problem of evaluating the conditional mean
and the conditional variance of some classes of bilinear time series models.
We show that for some simple models the non-parametric estimates of the
conditional mean and the conditional variance can be used as a graphical
tool for distinguishing ARMA and bilinear model and identifying the adequate
bilinear model. It may be mentioned that in the context of linear model (AR,
MA, ARMA) the conditional mean shows a straight line pattern.
\[ E( Z(t) | Z(t-j) = x ) = a + b x, \hspace{0.5in} j=1,2, \ldots. \]
Some theoretical results are also included.
Monday, April 29, 1996
- Graduate Seminar:
Yun Yi will give a talk entitled "A GENERAL AKAIKE-TYPE CRITERION FOR MODEL
SELECTION IN ROBUST REGRESSION" at 9:00 a.m. in N638 Ross.
Abstract:
Akaike's procedure (1970) for selecting a model minimises an estimate of the
expected squared error in predicting new, independent observations. This
selection criterion was designed for models fitted by least squares. A
different model-fitting technique, such as least absolute deviation regression,
requires an appropriate model selection procedure. This paper presents a
general Akaike-type criterion applicable to a wide variety of loss functions
for model fitting. It requires only that the function be convex with a unique
minimum, and twice differentiable in expectation. Simulations show that the
estimators proposed here well approximate their respective prediction errors.
Tuesday, April 30, 1996
Wednesday, May 1, 1996
- Graduate Seminar:
Tilaye Teshager will give a talk entitled "GRADUATION (THE REVISION OF
ESTIMATES)" at 2:30 p.m. in N638 Ross. This talk has been rescheduled
from April 26.
Abstract:
Graduation is an exercise in statistical estimation, specifically in a
situation where a sequence of estimates is to be produced, such as mortality
rates by age. The graduation process must take into account the presumed
relationship that exists among the elements in the sequence under
consideration, and called this prior opinion about the data.
Thursday, May 2, 1996
- Graduate Seminar:
Wenbiao Zhang will give a talk entitled "Analysis of Ordinal Paired
Comparison Data" at 11:00 a.m. in N638 Ross.
Abstract:
The method of paired comparisons is used primarily in cases when the
objects to be compared can be judged only subjectively: that is to say,
when it is impossible or impracticable to make relevant measurements in
order to decide which of two objects is preferable.
Two types of models are dicussed for paired comparisons of several
treatments using ordinal scales such as
(A < < < B, A < < B, A < B, A=B, A > B, A > > B,
A > > > B), where A < < < B denotes strong preference
for treatments B over treatment A, A < < B denotes moderate preference
for B, A < B denotes weak preference for B, A=B denotes no preference, and
so forth. For the Binary scale ( A, B, A > B), special cases of the models
using logit transforms simplify to the Bradley-Terry model.
When the same raters compare each pair of treatments, one can allow
within-rater dependence by fitting the models with constrained maximum
likelihood.
Tuesday, May 7, 1996
- Meeting: There will be a meeting of the Pure Math section
at 10:00 a.m. in N638 Ross.
Agenda: Hiring (due to retirement there is a possibility of hiring)
Wednesday, May 8, 1996
- Analysis Seminar: Professor Mioljub Niki\'c (University of
Belgrade) will speak on "Certain classes of analytic functions and some
inequalities in the algebra C(T)" at 3:00 p.m. in 1154 Vari Hall.
Abstract:
The object is to generalize certain classes of the holomorphic functions in
the algebra C(T) concerning the Frechet derivative and to prove some
inequalities in this algebra.
Friday, May 10, 1996
- Graduate Seminar: Maurice Tsang will give a talk entitled
"Measurement of Equity" at 10:00 a.m in N638 Ross.
Abstract:
When insurers set rates, they have to classify risks according to certain
criteria. However, when only certain factors are considered and others
ignored, inequity may result. With ideas used by economists to measure income
inequality, Promislow proposed a method of measuring this enquity. The concept
of riskiness is introduced to explain it and the method is illustrated in a
simple situation.
- Graduate Seminar:
Christopher Benson will give a talk entitled "THE KOLMOGOROV-SMIRNOV
STATISTICS" at 11:00 a.m. in N638 Ross. Rescheduled from Thursday.
Abstract:
The Kolmogorov-Smirnov statistics can be used for testing the fit of a model.
These will be discussed, and an outline of their derviation will be given
using the empirical distribution, the Brownian bridge, the Glivenko-Cantelli
theorem, and the central limit and weak convergence theorems for random
functions.
Wednesday, May 22, 1996
- Probability Seminar:
Vladimir V. Korolyuk (Landford International Ltd.) will speak on
"Diffusion approximation of Markov systems and networks"
(joint work with Nicholas Derzko, University of Toronto)
at 11:00 a.m. in N638 Ross.
Lunch to follow.
Friday, May 24, 1996
- Analysis Seminar:
Professor Sid Scull will give the first of a series of two or three lectures
entitled "Spectra of Localization Operators" at 4:00 p.m. in N638 Ross.
Friday, May 31, 1996
- Analysis Seminar:
Professor Sid Scull will continue a series of lectures entitled
"Spectra of Localization Operators" at 4:00 p.m. in N638 Ross.
Friday, June 7, 1996
- Analysis Seminar:
Professor Sid Scull will continue a series of lectures entitled "Spectra
of Localization Operators" at 4:00 p.m. in N638 Ross.
Friday, June 14, 1996
- Probability/Statistics/Combinatorics Seminar:
Mike Mosca (Oxford University): "Expected number of colours in a
proper colouring of a graph". 10:30 a.m. in N638 Ross (lunch to follow).
Abstract:
For a graph $G$ on $n$ vertices, let $\Omega$ be the set of all proper
$n$-colourings of the vertices of $G$ (ie. using up to $n$ available
colours, and such that no adjacent vertices are the same colour).
Define $\mu(G)$ to be the average number of colours used in a proper
$n$-colouring of $G$. It is $\#P$-$hard$ to compute $\mu(G)$ exactly, thus
it is believed that there is no efficient algorithm for doing so for a
general graph $G$. However fast randomised approximation may be possible.
One method that often works in approximating $\#P$-hard quantities is
approximately uniform random generation.
A rapidly mixing Markov chain on the set $\Omega$ can be used to do
this.
Some bounds on $\mu(G)$ would be useful, and in attempting to find some,
the effect of adding and deleting edges is investigated.
Wednesday, June 19, 1996
- Graduate Student Seminar:
Clara Auyeung will give a talk entitled
"THE BROUWER DEGREE IN FINITE DIMENSIONAL SPACES"
at 2:30 p.m. in N638 Ross.
Abstract:
The seminar is an introduction to the Brouwer degree. The Brouwer degree is a
map from a specific set and function to an integer value. Its main purpose is
to aid in finding the zeros of a function. In this seminar, basic properties
of the degree will be derived from its fundamental axioms. The computation of
the degree will be briefly discussed.
Thursday, June 20, 1996
- Graduate Seminar:
Ya-Ping Chen will give a talk entitled
"ON THE MEAN LENGTH OF THE ENTROPIC ALGORITHM OF PATTERN-RECOGNITION"
at 9:30 a.m. in N638 Ross.
Abstract:
Problem of recognition occur in practically every field of human activity,
e.g., in medical diagnosis, in chemical analysis, in the recognition of a
failure in a complicated mechanism or and entity in archaeology, in
classification problems, and so on. The entropic algorithm of recognition
is a strategy of recognition. The entropic algorithm of L.N. Landa for
pattern-recognition is improved for the case when the auxiliary experiments
are incompletely determined by the main global experiment. The
minimization of the mean length of the final graph is analysed.
Friday, June 21, 1996
- Probability Seminar:
George O'Brien (York University) will speak on "Large deviations for
spectrally positive stable and related processes" at 10:30 a.m. in N638
Ross.
- Analysis Seminar:
Professor Sid Scull will continue a series of lectures entitled "Spectra
of Localization Operators" at 4:00 p.m. in N638 Ross.
Tuesday, June 25, 1996
- Probability Seminar:
Tom Salisbury (York University) will speak on "Conditioned superprocesses"
at 10:30 a.m. in N638 Ross.
- Applied Mathematics Seminar:
Dr. Jurgen Burzlaff, Dublin City University will speak on
"Existence theorems for vortex-vortex scattering"
at 3:00 p.m. in N638 Ross
Abstract:
We formulate a Cauchy problem, first for the nonlinear partial
differential equations of the Abelian Higgs model, and then for those of
the CP^2 model. The problems describe the scattering of two vortices. It is
shown that the problems have a unique finite-energy solution. The symmetry
of the solutions is studied. It is shown that a local analytic solution
exists near the origin, and that the vortices scatter at right angle.
Wednesday, June 26, 1996
- Graduate Seminar:
A. Jabashvili will give a talk entitled "THE NET DIAGRAM OPTIMIZATION"
at 12:00 p.m. in N638 Ross.
Abstract:
A problem of reduction of task accomplishment mean time is considered (the
task is paralleled along branches). Organizing the iterational process of
finding the optimal regime it is necessary to have some mode of evaluation
of task accomplishment mean time. One method of this is also considered.
Friday, June 28, 1996
- Graduate Seminar:
Sanjana Kapoor will give a talk entitled "REACHING A VERDICT BY WEIGHTING
EVIDENCE" at 10:00 a.m. in N638 Ross.
Abstract:
Discussion of Mathematical model describing the process of reaching a
verdict by probabilistically weighting the available evidence. The classical
decision making rules from decision theory proposed by Hooper, Dempster,
Bayes, and Jeffrey prove to be special cases of such a weighting process.
The evidence induced by a fuzzy set will be discussed.
Wednesday, July 3, 1996
- Graduate Seminar:
Ying Liu will give a talk entitled "ASYMPTOTICS AND LIKELIHOOD INFERENCE"
at 2:00 p.m. in N638 Ross.
Abstract:
For the purpose of making inferences for a one-dimensional interest
parameter, the directed likelihood can be adjusted so that the resulting
modified directed likelihood is under ordinary repeated sampling
approximately standard normal with error of third order, conditional on a
suitable ancillary statistic and hence unconditionally. In general,
suitable specification of the ancillary statistic may be difficult. Two
stable and invariant adjusted directed likelihoods were introduced by
Barndorff-Nielsen & Chamberlin (1994), and a modified profile likelihood
was introduced by Fraser & Reid (1995) using ancillary direction
techniques, which are similar to the modified directed likelihood defined
by Barndorff-Nielsen (1983), but do not require the specification of the
ancillary statistic.
Thursday, July 4, 1996
- Graduate Seminar:
Yun Yi will give a talk entitled "THIRD ORDER ASYMPTOTIC MODEL:
EXPONENTIAL AND LOCATION TYPE APPROXIMATION"
at 10:00 a.m. in N638 Ross.
Abstract:
Recent methods of asymptotic inference including ancillarity and subsequent
marginalization or conditioning lead almost universally to the very simple
case of a real variable and a real parameter model, to a distribution on
the real line with a single parameter. This paper develops some basic
asymptotic theory for such a simple statistical model. For this special
case we can determine canonical versions of the best approximation at a
data point, by an exponential or location type model. We describe a step
by step reduction procedure for reducing an initial model to the canonical
exponential or location type model. The canonical model in each case has just
three parameters, one of which identifies the location or exponential type.
Tuesday, July 9, 1996
- Colloquium:
Prof. Gabor Tardos
(Mathematical Research Institute, Hungarian Academy of Sciences)
will talk on "A survey of interactive proofs in complexity theory"
at 1:30 p.m. in room N638 Ross.
All welcome: this talk will be of a general nature, for non-experts.
Thursday, July 11, 1996
Friday, July 12, 1996
- Analysis Seminar:
Dr. Qihong Fan of Peking University and McMaster University will give a
lecture entitled "Symbolic Calculus on the Affine Group "ax+b" at 4:00 p.m.
in N638 Ross Building.
Monday, July 15, 1996
Wednesday, July 24, 1996
- Graduate Seminar:
Kam Cheung Ko will give a talk entitled "METHODS FOR UNCONSTRAINED
OPTIMIZATION PROBLEMS" 12:00 p.m. in N638 Ross.
Abstract:
The basic idea for the optimization is to find the minimum of a function.
The optimization techniques are important not only because they can be
applied to sove many problems of considerable practical importance but also
because they provide the numerical analyst with a tool for solving a wide
range of nonlinear problems. In addition, they form an important
complement to analysis of convexity, calculus of variations, optimal
control and function approximation. The linear search methods, steepest
descent and least squares are basic optimization techniques which will be
discussed.
Friday, July 26, 1996
- Graduate Seminar:
Kim-Quang Tran will speak on "Operads in some Symmetric Monoidal Categories"
at 10:30 a.m. in N638 Ross.
This is a survey paper requirement for M.A. degree.
Abstract:
In some particular categories, operads are described in explicit formulas.
Operads are also viewed as monoids in symmetric monoidal categories and then
the general formulas for operads are derived from this point.
Monday, July 29, 1996
- Graduate Seminar:
Ioulia Ioffe will give a talk entitled "NUMERICAL SOLUTION OF AN INTEGRAL
EQUATION IN FINANCE" at 9:30 a.m. in N638 Ross.
Abstract:
An intergral equation describing a model in finance will be presented.
Existence of a numerical solution will be shown. Then a numerical algorithm
will be presented and the order of the resulting error will be determined.
Also some numerical results will be demonstrated.
- Graduate Seminar:
Yudong Zhang will give a talk entitled
"NUMERICAL SOLUTION FOR CONSTRAINED MINIMIZATION OF OPTIMAL CONTROL PROBLEMS
USING LAGRANGE MULTIPLIERS AND PIECEWISE OPTIMIZATION"
at 12:00 p.m. in N638 Ross.
- Graduate Seminar:
Sanjana Kapoor will give a talk entitled "METHOD FOR CONSTRAINED OPTIMISATION
PROBLEM" at 2:30 p.m. in N638 Ross.
Abstract:
The conditions for a local optimum are related to the saddle property of the
Lagrangain function. Formal Lagrangian Multiplier principal is used to find
minimum of f(x) subject to the constraints gi(x)=0. Discuss the finite
difference version of Newton method called Quasi Newton method to solve the
system of Nonlinear equations resulting from Lagrangian multiplier principal.
Tuesday, July 30, 1996
- Graduate Seminar:
Ya-Ping Chen will give a talk entitled "TESTS BASED ON THE
GAUSS-NEWTON REGRESSION" at 1:30 p.m. in N638 Ross.
Abstract:
There are several tests based on the Gauss-Newton Regression (GNR)
that can be highly useful in applied econometric work. These are:
- tests for the equality of two (or more)sets of regression parameters;
- nonnested hypothesis tests, in which a regression model is tested
against the evidence provided by one or more nonnested alternative models;
- tests based on comparing two sets of estimates, where generall one set
is consistent under weaker conditions than the other; and
- tests for heteroskedasticity of known form.
As a result the standard F test for (1) is commonly referred to by economists
as the Chow test. J and P tests are asymptotically equivalent instruments
under hypothesis for nonnested hypothesis tests, (2). And a class of tests,
which we called Durbin-Wu-Hausman, or DWH, tests, that can be used for (3).
In my seminar talk, I will focus on testing nonnested regression model
by using J-test.
- Graduate Seminar:
Ying Liu will give a talk entitled "SIMPLE MODIFICATIONS FOR SIGNED ROOTS
OF LIKELIHOOD RATIO STATISTICS" at 3:00 p.m. in N638 Ross.
Abstract:
For the purpose of making inferences for a one-dimensional interest
parameter, the directed likelihood can be adjusted so that the resulting
modified directed likelihood is under ordinary repeated sampling
approximately standard normal with error of third order, conditional on a
suitable ancillary statistic A and hence unconditionally. In general,
suitable specification of the ancillary statistic A may be difficult. An
alternative modification of the directed likelihood is given by DiCiccio
and Martin in 1993 using Bayesian point of view, that is available without
knowledge of A and the standard normal approximation to its distribution
has error of second order.
Wednesday, July 31, 1996
- Graduate Seminar:
Avtandil Jabashvili will give a talk entitled "SOME ASPECTS OF ELLIPTIC
BOUNDARY PROBLEMS" at 12:00 p.m. in N638 Ross.
Abstract:
The subject of this seminar are essential aspects of the theory of boundary
value problems for pseudo differential equations. The starting object is the
class G of operators introduced in Boutet de Monvel[1] containing besides
the classical elliptic boundary problems their parametrices, too. Thus all the
assertions about elliptic operators in the class G are valid especially for
classical elliptic boundary problems. An essential point of view in the whole
theory is the symbolic calculus for interior and boundary symbols. G is
closed under composition and other natural operations which are compatible with
corresponding operations on symbolic level.
[1] Boutet de Monvel, L., Comportement d'un operateur pseudo differential
sur une variete a bord J. Anal. Math. 17 (1966) 241-304.
- Statistical Consulting Service Special Colloquium:
Kai-Tai Fang (Hong Kong Baptist University and Chinese Academy of Sciences)
will speak on "SOME Q-Q PROBABILITY PLOTS FOR TEST MULTIVARIATE NORMALITY
SPHERICAL AND ELLIPTICAL SYMMETRY" at 2 p.m. in 3005 Vari Hall.
Abstract :
Graphical techniques have been widely used in various branches
of statistics. Among various kinds of graphical techniques the
normal probability plot has a long history and is one of the most popular
methods. In this talk we shall propose some Q-Q plots for multivariate normal
distribution as well as for testing spherical and elliptical symmetry.