Copy Link
Add to Bookmark
Report
Alife Digest Number 028
ALIFE LIST: Artificial Life Research List Number 28 Friday, July 6th 1990
ARTIFICIAL LIFE RESEARCH ELECTRONIC MAILING LIST
Maintained by the Indiana University Artificial Life Research Group
Contents:
Genetic Programming - New TR
Two Contests
re: Failing the Turing Test
Incorrect Phone Number for GA Short Course
Using GA's to configure and train Neural Networks.
----------------------------------------------------------------------
Date: Fri, 29 Jun 1990 14:59:18 PDT
From: John Koza <koza@Sunburn.Stanford.EDU>
Subject: Genetic Programming - New TR
A new technical report entitled "Genetic Programming: A Paradigm for
Genetically Breeding Populations of Computer Programs to Solve Problems" is
now available as Stanford University Computer Science Department technical
report no. STAN-CS-90-1314.
ABSTRACT: Many seemingly different problems in artificial intelligence,
symbolic processing, and machine learning can be viewed as requiring
discovery of a computer program that produces some desired output for
particular inputs. When viewed in this way, the process of solving these
problems becomes equivalent to searching a space of possible computer
programs for a most fit individual computer program. The new "genetic
programming" paradigm described in this report provides a way to search for
this most fit individual computer program. In this new "genetic programming"
paradigm, populations of computer programs are genetically bred using the
Darwinian principle of survival of the fittest and using a genetic crossover
(recombination) operator appropriate for genetically mating computer
programs. In this report, the process of formulating and solving problems
using this new paradigm is illustrated using examples from various areas.
Examples come from the areas of machine learning of a function; planning;
sequence induction; symbolic function identificiation (including symbolic
regression, empirical discovery, "data to function" symbolic integration,
"data to function" symbolic differentiation); solving equations (including
differential equations, integral equations, and functional equations)'
concept formation; automatica programming; pattern recognition; time-optimal
control; playing differential pursuer-evader games; neural network design;
and finding a game-playing strategy for a game in extensive form.
AVAILABILITY: (1) A limited number of copies of this report can be obtained
from the author FREE between now and August 31, 1990, by writing John Koza,
Post Office Box K, Los Altos Hills, CA 94023.
(2) Copies may be obtained for $15 from Taleen Nazarian, Computer Science
Department, Margarget Jacks Hall, Stanford University, Stanford, CA 94023
USA.
------------------------------
Date: Mon, 02 Jul 90 08:11:11 CDT
From: "Dave Goldberg (dgoldber@ua1vm.ua.edu)" <DGOLDBER@UA1VM.ua.edu>
Subject: Two Contests
The International Society for Genetic Algorithms is pleased to announce
two graphic design contests. Contestants are invited to submit
draft design ideas for (1) a society/conference logo and (2)
a conference poster. The society name is the International Society for
Genetic Algorithms (ISGA), and the conference name is the International
Conference on Genetic Algorithms (ICGA). The 1991 conference will be
held in San Diego (La Jolla) at the University of California at San Diego
between July 14-17, 1991. Submissions will be judged by the conference
committee, and winners will receive a free conference registration and
admission to the tutorial sessions. Submissions receiving honorable
mention will receive a free copy of the conference proceedings.
The logo design should be appropriate for use on letterhead, calls for papers
and other literature. The design may include ICGA and ISGA separately or
together.
The poster graphic should be appropriate for use on 11x17 posters, but it
also should be reducable to 8.5x11 for use on the proceedings cover
and t-shirts. Last year's graphic (see 89 conference proceedings) has
an automatic entry into this year's contest.
The contest is open until August 31, 1990. Please send submissions to
me at the following address:
Department of General Engineering
117 Transportation Building
104 South Mathews Avenue
Urbana, IL 61801-2996
I can answer questions by email (dgoldber@ua1vm.ua.edu). I look forward
to seeing your submission. Dave Goldberg
------------------------------
Date: Mon, 2 Jul 90 10:16:26 EDT
From: "David M. Chess" <CHESS@IBM.COM>
Subject: re: Failing the Turing Test
Robert Davidge <davidge@computing-science.aberdeen.ac.uk>:
> The Turing Test could only be applied to the
> principals of AL if both Natural Life and AL exhibited through the medium
> of earth based carbon chemistry or a computer terminal. This is clearly
> not possible. Therefore the Turing Test as applied to AI can not be
> applied to AL.
This is quite true as far as it goes. But I think something like (or
at least reminiscent of) the Turing Test could be adapted for use
in AL. I'm not sure exactly how useful it would be, but I will
put forward a few thoughts on the subject, just in case.
The underlying form of the Turing Test is "Allow a person to
interact with each of two systems through a filter that allows
only certain sorts of information to pass; see if the person
can determine which system is X, and which not-X". For the
traditional Turing Test (which isn't in fact quite what
Turing described, but that's OK), the filter allows only written
communication to pass, and X is "human". So the person in the
test has to figure out, by interacting only through written
communication, which system is a human being.
What might the filter and the X be for an AL Turing Test?
The most likely candidate for "X" is "a system rich in
carbon-based life". That is, the person in the test is
allowed to interact in some restricted way with two
systems: one a "real world" system rich with carbon-based
life, and the other an AL system of some kind. If the
person can't tell which is which, the AL system has
passed the test (for whatever that's worth). To make
the test fairly difficult, the person should probably
be a biologist of some stripe!
What should the filter be? The reason for the written-word
filter in the AI Turing Test is presumably that there is
some consensus that written communication allows broad scope
for the exhibiting of intelligence, without requiring many
intelligence-irrelevant properties of the AI system. Do
we have a similar consensus for AL? I suspect not yet,
at least none that anyone has thought hard about. I
will toss out a couple of suggestions.
One possibility would be to have the systems in the
test involve one (or a small number of) living/AL
entities, in a reasonably simple environment. The
filter between the person and the systems would allow
the person to receive information like "there is an
entity at point x,y", "the entity at point x,y is
low on energy", "the entity at point x,y is making
loud sounds of type t". The person might also
be able to interact in the other direction by giving
commands like "insert a barrier at point x,y",
"insert an item of 'food' at point x,y".
Another, more "macro", test would involve an entire
ecosystem. The filter might give the person information
like "there are K entities of type n in area A", or
"half of the entities in area A are moving rapidly
towards area C". I'm not as sure what the commands
in the other direction would be like. "Deposit N units
of type f food in area A", perhaps.
In both possiblities, the idea is that the filter allows
the person to see only those characteristics of the
system that are important to the question of liveness
(the person would not, for example, ever be able to
determine the chemical make-up of any entity in either
system), and to change the systems only in terms of
qualities that are relevant to liveness. Because we
don't have (yet?) a "common-sense" consensus to match
the one about the written word in the AI Turing Test
(and even that consensus has been challenged several
times), an AL Turing Test may not be as compelling as
the AI equivalent, but even if no particular test is
ever settled on, trying to find one might stimulate
interesting thoughts/discourse...
DC
------------------------------
Date: Tue, 03 Jul 90 05:49:58 CDT
From: "Dave Goldberg (dgoldber@ua1vm.ua.edu)" <DGOLDBER@UA1VM.ua.edu>
Subject: Incorrect Phone Number for GA Short Course
For those of you seeking information regarding the five-day
short course entitled "Genetic Algorithms in Search,
Optimization, and Machine Learning" to be presented at Stanford
University's Western Institute in Computer Science on August 6-10,
the wrong phone number was given previously. Contact Joleen Barnhill,
Western Institute in Computer Science, PO Box 1238, Magalia, CA 95954,
(916)873-0575.
The course, presented by John Koza and myself, includes in-depth
coverage of GA mechanics, theory and application in search, optimization,
and machine learning. Students will be encouraged to solve their own
problems in hands-on computer workshops monitored by the course
instructors. New material on Walsh functions, Boltzmann tournament
selection, Koza's genetic programming, messy genetic algorithms (mGAs),
and the theory of real-coded GAs and virtual alphabets will be presented
in a classroom setting for the first time. I hope to see some of you
there. Dave Goldberg
------------------------------
Date: Fri, 6 Jul 90 14:31:23 BST
From: Geoff Ballinger <tesun2!stu1gb@relay.eu.net>
Subject: Using GA's to configure and train Neural Networks.
I am currently working on a system which will use a genetic algorithm
to produce a configured and trained neural network given an example dataset.
Has anyone on the list done any work in this area or does anyone know of any
such work. I am already aware of the various papers in ICGA 89 and Montana and
Davis's 1989 paper. Thanks in advance,
Geoff.
(stu1gb@thorn-emi-crl.co.uk)
------------------------------
End of ALife Digest
********************************
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=---=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
= Artificial Life Distribution List =
= =
= All submissions for distribution to: alife@iuvax.cs.indiana.edu =
= All list subscriber additions, deletions, or administrative details to: =
= alife-request@iuvax.cs.indiana.edu =
= All software, tech reports to Alife depository through =
= anonymous ftp at iuvax.cs.indiana.edu in ~ftp/pub/alife =
= =
= List maintainers: Elisabeth Freeman, Eric Freeman, Marek Lugowski =
= Artificial Life Research Group, Indiana University =
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=---=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
End of Alife Digest
********************************
