Digital Burgess

From
Date
Dear all:

Someone mentioned in a earlier posting they were interested to hear what
occurred at the Digital Burgess meeting held in Banff, Canada Aug 29 - Sept
1.  What follows is in no way an "official" summary, just the impressions of
one palaeontologist who was lucky enough to be able to attend.

Complete details about the conference are at:
http://www.biota.org/conf97/index.html

The stated aims of the meeting (according to the programme) were to "provide
an interdisciplinary forum for the exchange of ideas between natural
scientists, computer scientists, and artists" including "demonstrations of
digital tools that model organisms and living systems."  This means that the
meeting sat at an intersection between palaeobiology, computer science,
social science and the arts, and the attendees included representatives of
all those fields - the sort of people an academic palaeontologist like me
does not normally get to meet.  The role of the Burgess Shale was really
that of an icon or mascot for the meeting.  Sponsorship came from, among
others, Silicon Graphics, British Telecom, Sun Microsystems and Intel, so
clearly there is industry interest in the link between evolution and computing.

The meeting opened with a keynote session open to the general public.
Presentations included the mesmerising NHK film/computer graphic documentary
on the Burgess animals; Randle Robertson of the Burgess Shale Foundation who
hope to open a visitor centre in Field; Paul Johnston from the Royal Tyrrell
Museum who described the display they are designing and discussed aspects of
reconstructing extinct organisms for the public; and Des Collins on the
importance of the Burgess Shale.

Day 2 saw about 35 of us make the non-trivial hike to Walcott's quarry on
the Burgess Shale, led by Des Collins, and in the company of a Discovery
Channel camera crew.

The morning of Day 3 was Palaeo Morning:  Bill Riedel introduced the
proceedings; Stefan Bengtson discussed issues surrounding the use of today's
life as key to understanding past life; Roy Plotnick showed how computer
models can help in understanding the response of ecosystems to perturbation;
Bruce Runnegar introduced the systematic method and discussed virtual
reconstructions of fossil organisms; Richard Gordon illustrated
differentiation waves on axolotl embryos, and showed how they may relate to
the generation of Bauplans.

The afternoon of Day 3 was Artificial Evolution Afternoon:  Tom Ray (ATR)
showed us Tierra which explores the dynamics of evolution and creation of
diversity among digital "organisms" which reproduce with mutation; Larry
Yaeger (Apple) discussed what it means to be "alive" with illustration from
his "PolyWorld" program in which virtual creatures with artificial neural
net brains reproduce sexually and compete for food; Karl Sims (Genetic Arts,
Inc.) of "Panspermia" (the computer graphic movie, not the theory) fame
showed us "evolved" (through crossover and mutation) art and some very funny
movies of virtual creatures which exhibit learned behaviours including
swimming and walking.  He also had his creatures compete for a "desirable
object" in a memorable sequence of 1-on-1 bouts.  Demetri Terzopoulos (U of
Toronto, Intel) showed us his extremely life-like virtual fish with their
own cognitive processes and learned swimming and schooling behaviours.

Day 4 started with P Prusinkiewicz (U of Calgary) illustrating modelling of
living and extinct plants (including Cooksonia, Lepidodendron) using the
method known as L-systems; Christian Jacob (U of Calgary) modelled plant
recolonisation of cleared land and showed how genetic algorithms can be used
to model plant evolution; Ricardo Colasanti (Cyberlife) showed how a
surprisingly simple cellular automaton can model plant growth and response
to light and nutrient conditions; Chris Winter and Paul Marrow (British
Telecom) spoke about how this high tech company believes that understanding
evolution and origins of diversity may be important in optimising software
efficiency; Rajarshi Das (Los Alamos Nat. Lab.) discussed the way in which
sophisticated computation can emerge in simple dynamic systems and the
implications for understanding emergent collective behaviour.

On the afternoon of Day 4, Steve Grand (Cyberlife, and "Creatures" creator)
discussed virtual pets and gave more thoughts on what "life" is; Bruce Damer
(DigitalSpace Corp.) discussed cyberspace as a real place.  We ended with
three artists: Steven Rook showed us the results of "aesthetic selection" of
genetic (i.e. "bred") artwork; Darrel Anderson talked about his algorithmic
art and "GroBot", his 3D drawing and programming environment for children;
Joel Hagen discussed past and present perceptions of extraterrestrial life
and showed us his museum of "alien fossils".

The meeting closed with another evening keynote which was open to the local
public, at which Sims, Rook, Yaeger and Ray reprised their performances.

OK, so not everything was fine and dandy.  Interdisciplinary interaction,
especially among people with such diverse backgrounds and agendas, is
unlikely to run completely smoothly, and it did not here.  There were a few
misunderstandings caused during the open discussions by misconceptions,
preconceptions, and misuse of terminology, and the discussion sessions often
did not get anywhere.  Also it seemed to me that the "Stately and
Progressive Unfolding of the Story of Life from Slime to Humans" image of
evolution, as opposed to the "Many-Branched Tree" image that most of us hold
dear, coloured the perceptions of many of the non-palaeontologists, and of
course their picture of the fossil record was skewed by the one example they
were being shown - the Burgess Shale.  Having said that, all non-palaeos I
spoke to were interested in the subject and wanted to learn more (but maybe
they were just a captive audience).

Also there was a definite problem in that some of the talks and discussion
sessions were summarised for the Conference web site, but this summarisation
was done by none specialists.  Although I am sure the summarisers did their
best, it does mean that what is recorded is a vastly oversimplified version
of what was actually said, and in some cases has resulted in statements
being attributed to people who do not remember making them (and some of the
statements do not make sense in the English language).

That said, I found these 4 days a very rewarding experience.  I have long
thought that computer modelling (possibly including "evolutionary
computation") could be a very useful tool in several areas of paleobiology,
and some of the presentations at this meeting have helped me to focus some
of my own ideas.  I understand preparations are underway for a second
conference in San Francisco in October 1998 - if it does not clash with GSA
it may well be worth a few more palaeontologists going and taking a look.

Congrats to conference organiser Bruce Damer and the staff of the Banff
Centre for the Arts.

BTW, explanations of L-systems, cellular automata, genetic algorithms can be
found on the web, often by following links from the Digital Burgess
Homepage.  Alternatively, there is a book "Artificial Life" by Steven Levy
which is a good way into the field.



------------------------------------------------------------
Tim McCormick
Department of Palaeontology, The Natural History Museum,
Cromwell Road, London SW7 5BD, United Kingdom

E-mail: T.McCormick@nhm.ac.uk

Dialling from within the UK:
	Tel: 0171 938 9289	Fax: 0171 938 9277

Dialling from outwith the UK:
	Tel: 44 171 938 9289	Fax: 44 171 938 9277
------------------------------------------------------------