Can some alife practioner tell me the practical application cases of Alife?

Can some alife practioner tell me the practical application cases of Alife?

Post by Hu Jianju » Thu, 22 Mar 2001 00:54:26



I am very interested in the practical application of ALife paradigm.  I am
wondering if anyone can list some important application cases ALife has
achieved so far.

www link Reference is also appreciated.

Hu Jianjun
Michigan State University

 
 
 

Can some alife practioner tell me the practical application cases of Alife?

Post by George Maydwe » Fri, 23 Mar 2001 08:52:01




Quote:

>I am very interested in the practical application of ALife paradigm.  I am
>wondering if anyone can list some important application cases ALife has
>achieved so far.

>www link Reference is also appreciated.

I'll list two practical applications: screen savers and games.

ALife and cellular automata in particular can be used to make visually
interesting screen savers:
   www.collidoscope.com

Supposedly (from searching google for "sim city cellular automata",
the "principles" of cellular automata were used to create Sim City and
Sim Ant. This page:
   http://simcity.ea.com/us/guide/
at Electronic Arts may or may not tell more. I cancelled a dismally
slow page download.

Regards,
George Maydwell

 
 
 

Can some alife practioner tell me the practical application cases of Alife?

Post by Tim Tyle » Sat, 24 Mar 2001 01:36:50



:>I am very interested in the practical application of ALife paradigm.  I am
:>wondering if anyone can list some important application cases ALife has
:>achieved so far.

: I'll list two practical applications: screen savers and games.

See also: the movies.

Quite a few "artificial living organisms" in things like Star Trek.
--
__________
 |im |yler  Try my latest game - it rockz - http://rockz.co.uk/

 
 
 

Can some alife practioner tell me the practical application cases of Alife?

Post by Randy Crawfor » Sat, 24 Mar 2001 13:40:30


Ah.  A chance to challenge the status quo.

I'm not an alife practitioner, but I have a degree in biology, CS and
and working on a PhD in AI.  My opinion?  Alife has evolved into a form
of computer-based entertainment that has little focus and even less purpose.
It's composed of biologically-motivated computer models for people who
aren't sufficiently interested in studying biological systems to put in the
effort necessary to model them meaningfully.

I looked seriously into artifical life several years ago and found that as
an intellectual pursuit, there were lots of geeks (like myself) having fun
with expressing the philosophical underpinnings that compose the field.  But
from what I saw, nobody was doing anything that could be considered research --
that is, nothing that was rigorous, goal-directed, replicable, and intended
to lay a foundation for future work.

So I'd suggest that if you choose to get involved in alife, you'll need to
1) do some hard work on realistically modelling biological systems, or
2) plan on having fun but expect your work to be dismissed by professionals.

    Randy


> I am very interested in the practical application of ALife paradigm.  I am
> wondering if anyone can list some important application cases ALife has
> achieved so far.

> www link Reference is also appreciated.

> Hu Jianjun
> Michigan State University

--
Randy Crawford

http://www.engin.umich.edu/labs/cpc
 
 
 

Can some alife practioner tell me the practical application cases of Alife?

Post by Erik Max Franci » Sat, 24 Mar 2001 13:58:25



> I looked seriously into artifical life several years ago and found
> that as
> an intellectual pursuit, there were lots of geeks (like myself) having
> fun
> with expressing the philosophical underpinnings that compose the
> field.  But
> from what I saw, nobody was doing anything that could be considered
> research --
> that is, nothing that was rigorous, goal-directed, replicable, and
> intended
> to lay a foundation for future work.

I take it you've only read popularization, not actual technical papers.

--

 __ San Jose, CA, US / 37 20 N 121 53 W / ICQ16063900 / &tSftDotIotE
/  \ We are victims of our own design
\__/ Oleta Adams
    Interstelen / http://www.interstelen.com/
 A multiplayer, strategic, turn-based Web game on an interstellar scale.

 
 
 

Can some alife practioner tell me the practical application cases of Alife?

Post by mathieu capcarrer » Sun, 25 Mar 2001 00:36:35



> So I'd suggest that if you choose to get involved in alife, you'll need to
> 1) do some hard work on realistically modelling biological systems, or
> 2) plan on having fun but expect your work to be dismissed by professionals.

I think the problem is taht you take ALife to be a Modelling of
living systems. I do not say this is not, or at least this was not
at all the aim at the beginning, but it has now taken a completely
different direction, which is fun I grant you that, but which is also
research. It would be a bit like repproaching that AI system are as far
as you can be from intelligence. The aims, research etc... of AI
now has not much to do with intelligence, but it still reasearch.

I agree that some research in the ALife field, like evolutionary
algorithms, is a bit  : I did this and that and it worked. This is
quite normal if you think it's quite young and you need a breadth
of experience  to begin theorizing.

Now I think Alife is clearly: bio-inspired computing: i.e., most (not all)
of the work is vaguely, one way or another, by living systems, and
aims at solving _Computing_ problems.

Mathieu Capcarrere

 
 
 

Can some alife practioner tell me the practical application cases of Alife?

Post by Tim Tyle » Sun, 25 Mar 2001 04:13:48


: Now I think Alife is clearly: bio-inspired computing: i.e., most (not all)
: of the work is vaguely, one way or another, by living systems, and
: aims at solving _Computing_ problems.

Much as I love that subject, GAs seem to me to be one area among many.

AFAICS, much of the alife that pays for itself is in the entertainment field.

http://www.creatures.co.uk/ would be one example.
--
__________

 
 
 

Can some alife practioner tell me the practical application cases of Alife?

Post by >ma » Sun, 25 Mar 2001 13:36:19


Like most other technologies, alife perhaps began/begins in
a tinker-toy fashion to see what it "is" and along the way,
people ponder "what it might be".  In the early days, video
games were considered a passing fad by corporations which
resulted in the game market crash of '83, however true
to form, early pioneers who believed in microprocessor-based
real-time graphics and saw it's potential, without perhaps
even realizing it's full potential... without such pioneering,
we would not likely have desktop computers with graphical
user interfaces as "the norm", much less many of the
related graphics applications based thereon.

Since corporations typically only invest in something that
already has a proven market demand, sheer forces of
greed, and they profit on the backs of the labor of many
who remain unknown but take credit for the work
"after the fact".  This leaves those "experimenters" to
themselves, and guess what ? The cheapest route to
getting most technologies off the ground is typically
via the consumer, and consumers like entertainment
above most else when they have cash to burn.  It is
only during economic hard times that this fact dwindles.

So what do experimenters do who want to explore
new technologies such as alife ?  They write games,
screen-savers, hack toys and make little robots, or
find a new way to solve otherwise tedius or boring
problems with some new methodology.

Alife is not a fad, it is an infant technology which
has an extremely expansive future ahead.  Where
it will lead beyond critters for films and other entertainment,
games etc is yet to be seen.  What is practical ?  It has
become defined not in terms of being frugal as it
once perhaps was, but in terms of "pay back"... RTO,
return on investment, perhaps.   So long as alife has
growing application in the entertainment industry,
regardless of form, it is likely to be considered practical
If you think of it in terms of being applicative to a
common problem and frugal, then perhaps the applications
are sparse, but I'm sure they exist.  Part of the problem
is defining artificial life.

I will avoid diving into that discussion but if one restricts
the term "artificial life" to mimicking biological systems,
then the degree of that mimickery must be ascertained,
and perhaps to what purpose.

Some people may consider the application of expert
systems to be a form of artificial life since it mimics
decision making processes.  Some may consider
some fuzzy logic applications to be sufficient, as
in anti-lock braking systems or crane controller
which are designed to load freight cars onto ships
in a very human-like way, attempting to prevent
swing and sway and to load the package as
quickly onto a ship as possible.

If we dive into neural nets and consider the
various ways it is used, such as voice
recognition, does that count ?  If not, why not ?

Going another step and considering autonomous
systems, the new ump*-million-neuron
cat project being done in Japan will likely
be hailed as an alife milestone, but is it because
there is a threshold of system complexity that
is involved ? Because certainly there do not
seem to be any extraordinarily new technologies
in use there, i.e. while amazing and interesting,
it still rests on the foundations of many basic
underlying systems, or so it seems from what
I've read about it.  Some may say this is
superfluous.... well, not all people will agree
on many things.

Alife has many applications in day to day things,
eg. perhaps wheelchairs which can be instructed
to take it's passenger from one location to another
and it finds it's way, avoiding pedestrians and
obstacles along the way, using alife heuristics.
Is it only alife if it is a robot with no passenger ?
I personally don't think so.

The term "robot", in fact, is grossly abused,
such as in the popular "Robot Wars" television
shows which are really just radio controlled
vehicles as far as I've seen so far.  Does this
lessen the glimmer of the efforts taken for
such shows ?  One, although not a battle
type of robot show, had some little vehicles
which danced in unison, climbed a rope
and, as I recall, tried to solve a maze problem.
Those verge more nearly upon the realm of
what I personally consider to be robotics.
I do not yet see the Honda "robot" doing
anything other than walking yet, and
that seems only a few feet from the reports
I've seen...  does that qualify it as a robot
or artificial life ?  You can see the problem...
definitions, acceptable terminologies etc. I seem
to get into this a lot in a number of newsgroups.

Corporations are indeed slowly beginning
to see the benefits of artificial life, but so
far focusing on data mining, forecasting and
other systems, some of which are considered
autonomous systems... the term "autonomous
agent" being quite popular to throw around
today.  Where do individual components
end and globally acceptable "artificial life"
begin ?  This is the question.  How complex
does a system have to be to qualify universally
as artificial life ?

The "Turing" test, which seems more to be
a "chat program" test as I've seen it to date,
is presumed to be one measure of the intelligence
of a system, as I understand it.  But, one winner
called "Alice", while very entertaining, is extremely
hard-coded and seems only to be a more expanded
version of a distant cousin "Eliza".  Many argue
about the importance of linguistics as a qualifier
to artificial life.

Personally I see it a bit more loosely defined, and
as time goes forward, the expectations of artificial
life will increase proportionally with capabilities.
This is just the way we are as consumers, I believe.
In the early and mid-80's, very crude dots on the
screen were considered "computer graphics" and
today you would be hard-pressed to find a
"graphics guy" to accept such systems as a
"real" computer graphics system...  expectations
increased with capabilities.

Artificial life therefore should not be taken lightly
just because some people believe that entertainment
related applications "do not qualify" for their
snobbish views... .people like that tried very
hard to quash early personal computers and
early video game consoles with such snobbery.
The technologies live on and grow.

Early experimenters may not have PhD's but they
have more prowess than most PhD's because they
have the insight to "go for it", to dig, to explore,
in spite of ridicule and near-sighted speculators.

I'm not sure if any of this answered your question,
but hopefully it is a bit of a kick-start to thinking
about it. :)

Mike


> I am very interested in the practical application of ALife paradigm.  I am
> wondering if anyone can list some important application cases ALife has
> achieved so far.

> www link Reference is also appreciated.

> Hu Jianjun
> Michigan State University

 
 
 

Can some alife practioner tell me the practical application cases of Alife?

Post by Randy Crawfor » Mon, 26 Mar 2001 18:49:30


I don't disagree with what Mathieu says.  I do see alife as biologically inspired,
but from the alife writings I've seen (Alife I, II, III, IV, and another book or two),
almost all the efforts appear to 1) attempt to adopt biologically-inspired means
to achieve some end, and 2) the end is seemingly less important than the means.

I liken alife (as I've seen it) to neural nets, expert systems, genetic algorithms,
and sundry other AI technologies whose application may fairly be characterized
as a hammer in search of a nail.  The fact that all of these techniques have been
applied to tasks such as the traveling salesman problem is testament to the oft-silly
priorities to which fans of a given technology will put them.  Of course there *has*
been good work in each of these fields, but it's often been hard to find amid the
seemingly directionless research that characterizes a discipline that's seemingly at
cross purposes.

A good example of a bio-inspired field that has lacked focus is genetic algorithms.
A great deal of work and publication in the field has explored their application on
problems for which they are clearly a poor fit, and IMO, a reluctance to reduce and
analyze the method for its underlying principles and their strengths and weaknesses.
This may be changing, however, if Michael Vose's book "The Simple Genetic Algorithm"
is any indication of impending trends.  Perhaps the same can be said for alife?

In part, I'm coming at this from the perspective of someone who left the AI field
in the mid-90s only to return recently.  It's clear to me that much of AI has moved
toward a more rigorous foundation, more like OR or applied math than the heuristic-
based approaches common in the early 90s and before.  I wonder, can the same be said
for alife?  Since alife efforts have traditionally been means-oriented and not goal-
directed, is it even possible for alife to adopt a more rigorous basis?  What would
it look like?  How would it stand apart of more traditional AI, decision-theoretic,
or OR optimization technologies?  Simply by having biological-analogs?

Unlike Tim Tyler's perspective, I don't include neural nets or even genetic algorithms
as forms of alife.  When a discipline grows sufficiently to define itself, it takes
on its own identity, and it will no longer be identified as principally part of another
endeavor, e.g. alife.  Alife has yet to do this, IMO, except as a conglomeration of
various hybrid and derived techniques from AI and mathematics that have only one thing
in common -- an biologically-analogous component that often has little to do with an
organized effort to rigorously assess a technology or achieve consistent and thematic
goals.  IMO, too much of alife can be characterized as "let's construct this system,
perturb it, and see what happens".  It's hard to build upon goals as fuzzy as that.

Perhaps the field has moved in a purposeful way while I was `away'.  If so, I'd be very
interested if someone could point me to a URL or a few papers that would help me to put
this objection to rest.  I suspect a large fraction of computer scientists other than
myself have already dismissed alife as `fanciful dalliances' and would also be interested
to hear that the discipline has matured since last they paid it any attention.

    Randy



> > So I'd suggest that if you choose to get involved in alife, you'll need to
> > 1) do some hard work on realistically modelling biological systems, or
> > 2) plan on having fun but expect your work to be dismissed by professionals.

> I think the problem is taht you take ALife to be a Modelling of
> living systems. I do not say this is not, or at least this was not
> at all the aim at the beginning, but it has now taken a completely
> different direction, which is fun I grant you that, but which is also
> research. It would be a bit like repproaching that AI system are as far
> as you can be from intelligence. The aims, research etc... of AI
> now has not much to do with intelligence, but it still reasearch.

> I agree that some research in the ALife field, like evolutionary
> algorithms, is a bit  : I did this and that and it worked. This is
> quite normal if you think it's quite young and you need a breadth
> of experience  to begin theorizing.

> Now I think Alife is clearly: bio-inspired computing: i.e., most (not all)
> of the work is vaguely, one way or another, by living systems, and
> aims at solving _Computing_ problems.

> Mathieu Capcarrere

--
Randy Crawford

http://www.engin.umich.edu/labs/cpc
 
 
 

Can some alife practioner tell me the practical application cases of Alife?

Post by Kent Paul Dol » Sat, 31 Mar 2001 18:27:24



Randy asks what practical use alife has enjoyed.

Well, we don't have a wonderfully consistent definition of alife, but
going forward based on the rather cheesy one: "computer imitation of
living systems", this Goodle search:

http://www.veryComputer.com/+simulation

which finds about 120,000 hits, will give you plenty of chances to find
imitations of living systems put to such practical uses as modeling
done in advance of policy changes in wildlife management approaches, to
assure that what is going to be done at least gives the impression that
it will have the intended results, and not some wildly different
results not at all the intention but inherent in the numbers.

I am myself in receipt (I helped with some coding issues) from the
author of a simulation of a stretch of the Cloquet River and the
crayfish population in it, which is used by the author in education to
teach students to estimate the influence of various parameters on the
success of first year crayfish on settling the parts of the habitat
where their predecessors have had success in previous seasons.  If the
author Steve Morse's request to the MIT team is honored, this should
eventually show up as cdad.slogo somewhere on the site rooted at:

http://www.veryComputer.com/

The famous SugarScape project, a dumbed-down demo of which is currently
available via the same site, uses artificial ants to model rather deep
questions in interactions of economic agents.  To learn more:

http://www.veryComputer.com/

There is a research area called "Ant Colony Optimization" which uses
the successful pheremone-trail-following techniques of foraging ants to
solve combinatorially complex problems in a variety of Operations
Research fields and similar areas with some successes.

    [This is in my mind somehow very tied into the advanced math topic
    "differential geometry", involving characterizing large geometric
    entities based only on local information, because ant colonies also
    achieve global results working only from very local information.]

See for example:

http://www.veryComputer.com/~mdorigo/ACO/ACO.html

So, I have given you several very practical uses to which alife
mechanisms are currently being put: wildlife management, education,
economics research, and providing heuristic approximate solutions to
combinatorially complex mathematical problems.

Looking up from a narrow biological focus to consider more
possibilities for alife, I'm sure you can find dozens more without
further help.

As someone along the chain of responses noted, in general so far we
computists are using alife tools to solve computing problems, not
limiting alife applications to biological problems [which are granted
to be very difficult for those without deep training in the biological
sciences].

In parallel with the usual way that mathematics progresses, we will
create the abstract, general purpose tools of alife, for our own
amu*t if you wish to characterize it thus.  It is then up to you
the biologist to bring them to bear in places where your domain
knowledge is critical to success, if you wish to see them put to use
there.

That urge to work with the abstractions we know best just comes out of
who we are.

My own several alife links might or might not be of interest, but for
the sake of completeness:

http://www.veryComputer.com/

Cheers!

xanthian.
--

--
Posted from smtp.well.com [208.178.101.27]
via Mailgate.ORG Server - http://www.veryComputer.com/

 
 
 

1. The ALife online and ALife.org website merge

It's not completely official by now, but you might see some changes in
the ALife internet landscape soon. After the upgrading of ALife online
(http://alife.santafe.edu) started slacking, and my ALife.org
(http://alife.org) started growing and getting its first hundreds of
visitors, I informed with Christopher Langton whether a merge of the two
websites could be possible.

I now received an e-mail from Chris in which he states that "[somebody]
was supposed to be maintaining alife online, but has done very little,
as far as I can tell - recently anyway - I would be happy if you wanted
to take it over and/or merge it with the alife.org"

"Either someone with the time to maintain the alife online site should
take it over, or it should just quietly fade away..." says it all.

Expect to see a slow change in both web sites. (information from
alife.santafe.edu will move into alife.org and ALife.org will get its
new name, ALife Online...)

Some of the most important e-mails yet need to be sent, but I'll be glad
to receive a couple of thousands more visitors to the ALife.org
website...

All comments are of course welcome

Anthony Liekens

http://alife.org

2. Space and Return pressed together on button causes crash

3. Alife Database (Alife Section) Updated

4. slippm won't dial

5. Is there any alife on comp.ai.alife?

6. BTL - the successor [i]

7. ALife Applications for the PC

8. GeForce 2 and SWX

9. Application Of ALife

10. ALife Application Newsletter

11. A practical case

12. CFP:The Practical Application of Intelligent Agents and Multi Agents (PAAM98)