11h30-12h30 Wednesday, August 25
Chairman : Gilbert Saporta Room: Paul Painlevé
Devroye, School of Computer Science, McGill University,
"Complexity questions in non-uniform random variate generation"
will discuss the sub-discipline of random variate generation, starting
from the pioneering fifties, and ending with the challenges that lie
11h-12h Friday, August 27
Chairman : Stan Azen Room: Paul Painlevé
Edler, Division of Biostatistics, German Cancer Research
"Computational Statistics Solutions for Modern Biomedical Research: A Challenge and Chance for Both"
talk will address how computational statistics, computing power, data
analysis and data interpretation has invaded in a very short time many
cutting edge research areas of modern biomedicine and biomedical
research. Which are consequences for both sides when they want to be
partners in successful projects? The challenging areas are not at all
restricted to the genomic and post genomic molecular data but reach far
more out into all fields of diagnosis, prognosis and outcome prediction
of diseases and their treatment. This includes complex designs as well
as the efficient use of image data. The lecture will demonstrate
successes fields, elaborate challenges and also point out computing
needs to communicate computing.
9h30-10h30 Monday, August 23
Chairman : Patrick Groenen Room: Paul Painlevé
Hand Statistics Section, Imperial College, London
"The laws of Coincidence"
events often lie at the roots of discoveries in science or actions in
other domains. Familiar examples are the discovery of pulsars,
the identification of the initial signs of an epidemic, and the
detection of faults. In general, they are events which are seen
as so unexpected or improbable that one suspects there must be some
underlying cause. One particular kind of anomalous event is the
coincidence, defined by Diaconis and Mosteller (1989) a `a surprising
concurrence of events, perceived as meaningfully related, with no
apparent causal connection'.
The search for explanations - for the causal connections - has led to
superstitions, beliefs in gods, fate, synchronicity, and other
representations of the notion that there are unseen forces guiding our
destiny. These might be characterised as `prescientific
explanations' as they rely on forces which do not have the observed
empirical infrastructure of physical forces, such as gravity or
magnetism, as well as social and behavioural forces such as Adam
Smith's invisible hand.
While one can never prove that there are no such forces (one can prove
a presence, but never an absence), closer examination shows that in
many cases one should in fact expect apparent `coincidences' to
occur. This explanation can be conveniently couched in terms of a
number of laws. Since these laws are concerned with what are
perceived as unusual data configurations and events, they can be seen
as serving a role complementary to familiar statistical laws covering
mass phenomena, such as the law of large numbers, and principles which
derive from this, such as the laws of statistical mechanics and
thermodynamics in physics and many laws in economics and the social
sciences. These laws describing the origin of coincidences include:the
law of total probability, the law of truly large numbers, the law of
near enough, the law of search, the law of the lever, the law ofthe
tortoise, and the law of selection.
These laws are defined and illustrated with a series of examples.
The implications for anomaly detection methods are discussed.