The Golem
Project
By - Automatic Design and Manufacture of
Robotic Lifeforms Hod
Lipson and Jordan
B. Pollack, at the DEMO
Lab CS Dept., Brandeis University
Evolution of machines
The
field of Artificial Life examines "life as it could
be" based on understanding the principles and
simulating the mechanisms of real biological forms. Just
as airplanes use the same principles as birds, but have
fixed wings, artificial lifeforms may share the same
principles, but not the same implementation in
chemistry. Every feature of living systems seems
wondrous until it is understood: Stored energy,
autonomous movement, and even animal communication are
no longer miracles, as they are replicated in toys using
batteries, motors, and computer chips.
Complex
biological forms reproduce by taking advantage of an
arbitrarily complex set of auto-catalyzing chemical
reactions. Biological life is in control of its own
means of reproduction, and this autonomy of design and
manufacture is a key element which has not yet been
understood or reproduced artificially. To this date,
robots - a form of artificial life - are still designed
laboriously and constructed by teams of human engineers
at great cost. Few robots are available because these
costs must be absorbed through mass production that is
justified only for toys, weapons, and industrial systems
like automatic teller machines.
In
the Golem project (Genetically Organized Lifelike
Electro Mechanics) we conducted a set of experiments in
which simple electro-mechanical systems evolved from
scratch to yield physical locomoting machines. Like
biological lifeforms whose structure and function
exploit the behaviors afforded by their own chemical and
mechanical medium, our evolved creatures take advantage
of the nature of their own medium - thermoplastic,
motors, and artificial neurons. We thus achieve autonomy
of design and construction using evolution in a limited
universe physical simulation, coupled to off-the-shelf
rapid manufacturing technology. This is the first time
robots have been robotically designed and robotically
fabricated.
What would you like to learn more
about?
The Golem@Home Project
The Golem@Home Project is a
screensaver aimed at harnessing idle CPU power across
the Internet to perform massively distributed
evolutionary computation. By installing this software
you are volunteering your computer to participate in
this distributed AI experiment. While you participate,
any creatures born (evolved) on your computer are
copyrighted to you. At any point you may uninstall this
program and thus discontinue your participation.
The Golem program is a screen saver.
Once you complete the installation, click the right
mouse button on your desktop to activate desktop
properties dialog, and select "Golem" from the
list under the Screen Saver tab. Thereafter, whenever
Golem is activated it will continue evolving bodies and
brains of electromechanical robots, and animate some on
the screen. Occasionally (say, once in a week or two),
if a network connection is available, one or few evolved
creatures might migrate from your computer to another
Golem screen saver that happens to be active on the net,
and some new ones might migrate into your Golem program.
For security considerations Golem will exchange only
data with other programs, and will not receive or
transmit executable files. Furthermore, for privacy
considerations Golem will only transmit data associated
with the Golem project. In any case, you may use the
advanced settings communication tab to disable incoming
communications or all communications entirely.
From
time to time, Golem will check if a version update is
available. If so, a new version will be downloaded
automatically directly from the DEMO Lab at Brandeis
University. For security reasons, no other source will
be used. Again, automatic version update can be disabled
at the advanced settings communication tab. If you
disable this option, please check this site from time to
time to see if any major changes occur, or ask us to
notify you by email.
Creatures born on your computer are
copyrighted to you according to the contact information
you provide at installation or at the advanced settings
identification tab. They will carry your ownership
information during their entire life span, even when
they leave your computer. If communication is enabled,
this information will be accessible to others. You may
change this information or enter false identification at
any time.
To minimize human intervention in this
experiment, you have little control over the evolution
of these creatures. They will evolve autonomously. At
most, you may have indirect control by changing
population size, wiping or reverting an entire
population, or changing the landscape in which they
evolve. You may try to use the control you have to make
the creatures evolving on your computer better and more
robust, so that they survive when they migrate to other
machines. This is a form of human machine co-evolution.
We reserve the right to publish any
statistics obtained in this experiment.
If you have any questions about this
study or your rights, please contact
lipson@cs.brandeis.edu.
LiveTruss
This
is a standalone version of the evolution program that
was used to generate the results shown in the paper. We
provide this program so that you can reproduce the
results and experiment with various parameters. The
default parameter settings should generate interesting
results without any tweaking, so try them first: just
download the program, run it and select
"Evolve!" from the Evolve menu. Leave the
program running overnight on a fast computer or for a
day ot two on a slow computer and you should see some
interesting creatures appear. Minimizing the window
makes the program run faster. You can select properties
from the Evolve menu to see how well the population is
doing. You can select "stop evolving" and
"continue evolving" at any time. When
simulation is stopped, flip through the population using
"Previous" and "Next", and observe
the performance using "Simulate", If anything
particularly amazing comes up, save the creature and
email it to us! (or make a video of it, using the
"Start Recording" option).
Copyright (c) 2000
Lipson & Pollack
lipson@cs.brandeis.edu
Further reading
H. Lipson and J. B. Pollack (2000),
"Automatic
design and Manufacture of Robotic Lifeforms",
Nature, to appear.
C. Adami (1998), Introduction to
Artificial Life, Springer Verlag
Husbands P., Meyer J. A. (1998),
Evolutionary Robotics, Springer Verlag
S. Nolfi, D. Floreano (2000), The
Biology, Intelligence, and Technology of Self-Organizing
Machines, MIT Press
