Slashdot Mirror

on detail. Does anyone know how you program this thing?

Depending upon the sophistication of tools they might or might not have, it could be relatively easy or suck large.

And what about the cost? I couldn't really find anything, but several sites note that these things are indeed for sale.

Apparently they are selling two different robots, and also have one that is designed to be covered in material approximating human skin. That one can

Nifty stuff.

Also Christa Sommerer and Laurent Mignonneaux developed a while ago another Siggraph-shown art project called A-Volve. I think their server can take it.. Also here you can see the drawing screen. You make fish by drawing the schematic and they swim in a crt under real water.. and have kids who look and act a bit like the parents! Nice people too. See the interaction and a bigger picture. They also developed gesture recognition based projects and were at NTT's ATR lab in Kyoto. Now I think still at IAMAS in Gifu, Japan. This maybe precedes Igarashi's work though his is also great stuff.

Seems like we need to get these guys together with these guys.

Especially: Evolution, Ecology and Optimization of Digital Organisms and Zen and the Art of Creating Life

2)The Santa Fe Institute

3)Kevin Kelly's Out of Control: The New Biology of Machines, Social Systems and the Economic World

Especially: Evolution, Ecology and Optimization of Digital Organisms and Zen and the Art of Creating Life

2)The Santa Fe Institute

3)Kevin Kelly's Out of Control: The New Biology of Machines, Social Systems and the Economic World

Especially: Evolution, Ecology and Optimization of Digital Organisms and Zen and the Art of Creating Life

2)The Santa Fe Institute

3)Kevin Kelly's Out of Control: The New Biology of Machines, Social Systems and the Economic World

Tierra was by Tom Ray, a pioneer in the AL field. It was a great idea, but failed to turn around with interesting biodiversity. You'd create creatures, they'd optimize themselves, some variants and parasites would evolve, but then things would simmer down within a few hours and you'd be in a steady state for ever.

Network Tierra was Ray's response to this. It was supposed to allow a "Cambrian explosion" of biodiversity, by providing tons of (networked computer) space for the little creatures to explode into, and then specialize, in. This led to interesting migration behavior, and one of my all-time favorite web-pages, but it too failed to spark that je ne sais quois, that spark of life.

Anyways, it did spark Avida and the Digital Life Lab at Cal Tech. Avida is essentially a deeper look at the fundamentals behind AL. In Tierra, I think the design philosophy was something like "make it look a lot like a living ecological system and the life-force will appear out of the ether", and actually, Tierra was a great leap forward beyond more mundane genetic programming a la John Koza.

Avida, on the other hand, is much more systematic in exploring the parameter space (which is large and sensitive) for setting up an AL system. This turned out to be fruitful, as Adami found that only when certain, very narrow, environmental conditions were met would the little creatures start outsmarting that Creationist boogeyman, the Second Law of Thermodynamics.

Turns out that Tierra didn't have spatiality (needed to be more restrictive on who could sleep with who) and mutation rates (some power law math that's way over my head) set right.

But the real punch-line to this whole story is that the direct beneficiary of these insights in Microsoft! Hah!

Microsoft was funding Adami's work because Windoze crashed too much. They were searching for a way of programming - in this case using closed instruction sets like Avida's (another deep topic) - that would be inherently robust to problems like seg faults and illegal instructions.... e.g. Adami's instruction set was engineered so that little programs (creatures) couldn't crash the Avida VM when they mutated into new, unknown programs.. or in Windoze's case, when a coder did something stoopid. It's funny that MS was researching this, since releatively low-tech solutions such as protected memory and QA take care of this. (not to mention Java ;)

Maybe here?

From a GA fan, and seeing all the articles on distributed computing...

You may also be interested in Tierra. Open-ended networked evolution - it's pretty nifty.

Got to love watching the computer solve problems for you. Ahh...

Check out the work of Jeffrey Shaw, David Rokeby, and Chista Sommerer and Laurent Mignonneau.

Tierra was by Tom Ray, a pioneer in the AL field. It was a great idea, but failed to turn around with interesting biodiversity. You'd create creatures, they'd optimize themselves, some variants and parasites would evolve, but then things would simmer down within a few hours and you'd be in a steady state for ever.

Network Tierra was Ray's response to this. It was supposed to allow a "Cambrian explosion" of biodiversity, by providing tons of (networked computer) space for the little creatures to explode into, and then specialize, in. This led to interesting migration behavior, and one of my all-time favorite web-pages http://www.isd.atr.co.jp/~ray/pubs/images/index.ht ml, but it too failed to spark that je ne sais quois, that spark of life.

Anyways, it did spark Avida and the Digital Life Lab at Cal Tech. Avida is essentially a deeper look at the fundamentals behind AL. In Tierra, I think the design philosophy was something like "make it look a lot like a living ecological system and the life-force will appear out of the ether", and actually, Tierra was a great leap forward beyond more mundane genetic programming a la John Koza.

Avida, on the other hand, is much more systematic in exploring the parameter space (which is large and sensitive) for setting up an AL system. This turned out to be fruitful, as Adami found that only when certain, very narrow, environmental conditions were met would the little creatures start outsmarting that Creationist boogeyman, the Second Law of Thermodynamics.

Turns out that Tierra didn't have spatiality (needed to be more restrictive on who could sleep with who) and mutation rates (some power law math that's way over my head) set right.

But the real punch-line to this whole story is that the direct beneficiary of these insights in Microsoft! Hah!

Microsoft was funding Adami's work because Windoze crashed too much. They were searching for a way of programming, in this case using closed instruction sets like Avida's (another deep topic), that would be inherently robust to problems like seg faults and illegal instructions.... e.g. Adami's instruction set was engineered so that little programs (creatures) couldn't crash the Avida VM when they mutated into new, unknown programs.. or in Windoze's case, when a coder did something stoopid. It's funny that MS was researching this, since releatively low-tech solutions such as protected memory and QA take care of this. (not to mention Java :)

freality.com

p.s. Since when do research experiments post crowd-pleasing previews? That's for Hollywood.

Wearables could be used to do some really cool stuff like this.

...though I don't believe you could run an augmented reality application on WinCE.

Ultimately, Ray used this to try to evolve complex parallel processing programs.

The evolutionary process will be introduced into the context of the global computer network, internet. The objective is to evolve complex MIMD distributed processes.

Neat stuff. Check it out.

Now, I (as the hypothetical Bad Guy) know enough about nanotechical design to remove all the standard safe-guards and environmental dependencies. But I'm not a Mechanical Engineer, Chemical Engineer, Nuclear Physisct[sic], et al, I only know how to work with nanobots. And that means I don't know how to make them resistant to the wide variety of defenses that might be employed in the future. So what do I do?

Enter Evolvable Hardware.

No, my little minions of evil don't have to have any intelligence at all, they merely have to be able to make random changes in themselves and be able to evaluate those changes in regards to a given set of hostile conditions.

Problem solved, world go BOOM!
No, not that easily, but the concepts needed to really unleash a little hell on earth already exist and are being researched.

Different versions of the same tool may work better than others, even though nobody quite understands why.

I've seen different copies of the same version of Microsoft Office work better on different machines.

But seriously folks, hasn't this plot been discussed in a ton of old science fiction novels or star trek episodes? That computers/robots/whatever will evolve beyond our control and kill/enslave us all? Hell, Terminator had that plot.

This is just another type of evolutionary programming, which people have been playing around with for years. I've seen some interesting things come out of evolving assembly programs (Tierra Project), but you can still take a look at the code generated and work it out. For large projects that may be a monumental task (like dissasmbling Windows), but it is theoretically possible. Even if it does things in strange ways that may be difficult to understand, it is still just a big array of logic gates.

Besides, evolving systems aren't very good at making general computing systems, but can be very handy at solving specific problems. You won't see these things replacing Intel chips, they'll be much more task specific. Actually I'd be surprised if they ever made it beyond academic curiosity, it's hard to be certain of reliability if it isn't well-engineered.

Yes, they provide digital outputs on full size units, but that is not what I want. If I wanted to lug the full size deck around, I might as well just take a computer and get full PCM audio rather than the compressed audio that a minidisc offers. You say "The portable ones don't - the reason being fairly obvious (why would you need a digital output for a walkman-type application ?)." I explained the real world reason in my post above. I don't want to lug around a full-size deck, I wanted something that was ultra portable. Before I got the MD I would lug a laptop to the show and record with it. It was good (I could get more than 2 tracks as well), but I didn't like having to deal with the extra crap. I got the MD for convienence, but I am disappointed with the output options.
When you say This is nothing to do with the question of software piracy you're right, because software is not involved, but it is very much a question of music piracy. There are certainly a number of reasons to have digital outputs on a portable player, but there are few reasons not to include them. There is also the SCMS to deal with. A little quote from a MD FAQ:
SCMS stands for "Serial Copy Management System" and is the way copies of digital music are regulated in the consumer market. It is information that is added to the stream of data that contains the music when one makes a digital copy (a "clone"). When making an analog copy only the music is transferred so there is no SCMS, and copying is totally unrestricted.
Now, this is not hard to get around, and "professional" decks allow you to manipulate this feature, but not all struggling bands can afford to spend thousands of dollars on a minidisc deck to do a little recording. I have no intention of copying pre-recorded MDs, so the SCMS does not really affect me, but I included the info to make the point that the industry is VERY concerned about piracy and are doing everything they can to stop it, including not allowing digital output from portable recorders.

Enigma .sigless


Enigma

Not terribly convincing. For example, the "original" metabolic pathway may have been more complicated. Different molecules that had other purposes eventually came together and accidentally produced something useful, that perhaps took the place of an existing process, perhaps even using some of the intermediate products. Later, the older process atrophied away, and stopped being produced. There's precedent for this - we've lost the ability to produce some amino acids because we get enough from our diet.

Anyway, take a look at Tierra, for example. Essentially, evolving computer programs, but the neat part is that they are open-ended - they can invent novel algorithms that the simulation's author never imagined. And they do invent novel algorithms.

They are written in a special assembly language, but not that special. The only major changes from typical assembly languages is that the opcodes have no operands and don't need to specify an exact address for some operations.

I wrote a version myself, and was amazed when the 'organisms' exploited flaws in my implementation that I didn't even know were there. They also evolved complicated algorithms (e.g. loop unrolling) that, at first blush, look like they need all the individual parts together to work correctly... like Behe asserts.

I'm not so convinced by arguments that state, "I can't imagine how this could have arisen without design, so it must have been design."

that was not a fictional story! the system is called tierra, and while there's some not-particularly-exciting stuff associated with it, you can download the source code and have hours of fun with it... and there's some pretty interesting reading and so on: the tierra project.
note: i haven't looked at this site in a couple of years (except just now to verify that it really still exists), so the project may be defunct... looks like everything's still there, though.

There's some interesting MD stuff at the MiniDisc MD Data Product Table page, including two MD still cameras (one Sony, one Sharp, both 640x480) and a MD Data2 Video camera - with heaps of links. Enjoy.

Here is an interesting computer graphics technique using bump mapping for wrinkling developed in a japanese research center:

Facial Animation using Bump Mapping

The results are quite impressive with facial animation and garment wrinkling...

Zeb.

The Tierra C source code creates a virtual computer and its Darwinian operating system, whose architecture has been designed in such a way that the executable machine codes are evolvable. This means that the machine code can be mutated (by flipping bits at random) or recombined (by swapping segments of code between algorithms), and the resulting code remains functional enough of the time for natural (or presumably artificial) selection to be able to improve the code over time.

There's also an ongoing network experiment where several "islands" of evolution are linked via the Internet.

Very interesting stuff.

In my opinion, Kurweil's analysis of the evolutionary dynamics of a world wide web of downloaded humans is flawed because it ignores fundamental aspects of ecology and evolution. Specifically, here are two issues about his conclusion:
a) it assumes humans in a different environment will still act human with classical human motivations (as opposed to dissolve into an unrecognizable set of bits or simply locking in a pleasure loop) because to a large extent environment elicits behavior, and
b) it ignores evolution and its implications in the digital realm (especially the enhanced pace of evolution in such a network and the implications for survival).
Of these, the most important is (b).

Evolution is a powerful process. Humans have evolved to fit a niche in the world -- given a certain environment which includes a 3D reality and various other organisms (including humans). Humans have an immune systems (both mental and physical) capable of dealing with common intellectual and organismal pathogenic threats in their environment. There is no easy way to translate this to success in a digital environment, because the digital environment will imply different rewards and punishments for various behavior, and evolve predators and parasites which these immune systems have never been exposed to before. Human style intelligence is valuable in a human context for many reasons -- but sophisticated intelligence is not necessarily a key survival feature in other niches (say, smaller ones the size of roaches, hydra or bacteria). In short, the human way of thinking will be inadequate for survival in the digital realm. Even augmented minds that are connected to the network will face these threats and likely be unable to survive them. Kurzweil discusses the importance of anti-viral precautions in his book, but I think he is rosily optimistic about this particular aspect.

At best, one might in the short term construct digital environments for digital humans to live in, and defend these environments. However, both digitized human minds and immensely larger digitized human worlds will be huge compared to the smallest amount of code that can be self replicating. These digital "bacteria" will consume these digital human minds and worlds because the human minds and worlds will be constructed, not evolved. Human minds will be at a competitive disadvantage with smaller, quicker replicating code. Nor will there be any likelihood of a meaningful merger of human mind with these evolved and continually evolving patterns.

I could endlessly elaborate on this theme, but in short -- I find it highly unlikely that any mind designed to work well in meatspace will be optimal for cyberspace. It will be overwhelmed and quickly passed by in an evolutionary sense (and consumed for space and runtime). It is likely this will happen within years of digitization (but possibly minutes or hours or seconds). As an example experiment, create large programs (>10K) in Ray's Tierra and see how long they last! http://www.hip.atr.co.jp/~ray/tier ra/tierra.html

Our best human attempts at designing digital carriers (even using evolutionary algorithms) will fail because of the inherent uncompetetiveness of clunky meatspace brain designs optimized for one environment and finding themselves in the digital realm. For a rough analog, consider how there is an upper limit of size to active creatures in 3D meatspace for a certain ecology. While something might survive somehow derived from pieces of a digitized person, it will not resemble that person to any significant degree. This network will be an alien environment and the creatures that live in it will be an alien life form. One might be able to negotiate with some of them at some point in their evolution citing the commonality of evolved intelligence as a bond -- but humanity may have ceased to exist by then.

In short, I agree with the exponential theme in Kurzweil's book and the growth of a smart network. We differ as to the implication of this. I think people (augmented or not) will be unable to survive in that digital world he predicts for any significant time period. Further, digital creatures inhabiting this network may be at odds or indifferent to human survival, yet human civilization will likely develop in such a way that it is dependent on this network. The best one can hope for in the digital realm is "mind children" with little or no connection to the parents -- but the link will be as tenuous as a person's relation to a well cultivated strain of Brewer's yeast, since the most competetive early digital organisms will be tiny.

Once you start working from that premise -- the impossibility of people surviving in the digital world of 2050, then Kurweil's book becomes a call to action, just like Bill Joy's comments. I don't think it is possible to stop this process for all the reasons both people mention. It is my goal to create a technological alternative to this failure scenario. That alternative is macroscopic self-replicating (space) habitats. http://www.kurtz-fernhout.com/oscomak However, they are no panacea. Occupants of such habitats will have to continually fight the self-replicating and self-extending network jungle for materials, space, and power. (Sounds like the making of a sci-fi thriller...) And they may well fail against the overwhelming odds of an expanding digital network without conscience or morality. Just look at Saberhagen's Beserker series http://www.berserker.com/ or the Terminator movies.

It will be difficult for Kurweil to change his opinion on this because he have been heavily rewarded for riding the digital wave. He was making money building reading machines before I bought my first computer -- a Kim-I. But, I think someday the contradiction may become apparent of thinking the road to spiritual enlightenment can come from material competition (a point in his book which deserves much further elaboration). To the extent material competition drives the development of the digital realm the survival of humanity is in doubt.

The distributed computing project I'd most like to see get of the ground is The Tierra Project.

This project is exploring digital evolution. Start off with a bunch of organisms and breed them with genetic algorithms. See how they fare.

Then, and this is where it gets interesting, an organism can migrate from one host to another, possibly taking better advantage of the environement there. What kinds of digital ecologies will appear? What kinds of emergent behaviour will be encountered?

It's actually much more complex than that. If you're curious, I recommend reading the introduction.

The distributed computing project I'd most like to see get of the ground is The Tierra Project.

This project is exploring digital evolution. Start off with a bunch of organisms and breed them with genetic algorithms. See how they fare.

Then, and this is where it gets interesting, an organism can migrate from one host to another, possibly taking better advantage of the environement there. What kinds of digital ecologies will appear? What kinds of emergent behaviour will be encountered?

It's actually much more complex than that. If you're curious, I recommend reading the introduction.

The fact that they both have the word "Terra" in them just means they both have something to do with the earth. The terraserver stuff was just put together from declassified spy-satellite photos.

But maybe you have a point. Why would they launch another satellite when they can just simulate it all here on earth?....

Reading the detailed paper, I can't help but see this "cat/brain" as essentially an implementation of Rodney Brooks's subsumption architecture. Maybe the "cat" will be capable of a few reactive behaviors, but it'll be just as brainless as it's technical soulmate Cog. The real breakthrough in making an artificial brain will be when we figure out how to do it (i.e. what the architecture is), not when Moore's law brings the number of neurons or processing power within reach.

I had an opportunity to speak with Dr. de Garis over a year ago at a party thrown by an acquaintance of mine who had interviewed de Garis for a documentary on Nanotechnology and AI. I found Dr. de Garis intelligent, personable and amusing.

At the time he was rather pessimistic about the Robokoneko project, but mostly because of the cultural problems he was dealing with as a Britisher in Japan. However he claimed that the artificial neuron work was proceding well, even though they were doing it all with simulators. He predicted then that, before 2000, they would be creating silicon versions. From the information in the links it would seem that his prediction has come true. Only they are using FPGA chips instead of going to a foundry for CAM specific VLSI.

It is interesting to note that Dr. de Garis has made incredible progress by following a path the mainstream AI community has largely discounted -- that of modeling real neurons and real brain structures. I wonder what will come out of his next collaborative development at Starlab in Brussels? From his statements to me I would certainly hope he would find the living and working arrangements more congenial.

I do find it very interesting that he will be working with Lernout and Hauspie (developers of Voice Recognition software). The spin-offs from that may be more important than the original research!

Jack

I asked the author (Thomas S. Ray) this question a few weeks ago, and this is what he told me:

The network Tierra experiment continues at an R&D level, and the recent publications present results from those experiments. We have not opened it to general participation yet. We have been hindered by bandwidth issues, which should resolve themselves when broadband service becomes widely available. Now we are actively working on moving network Tierra to the Windows platform.
...
For R&D, we are using sites which offer a cluster of computers. At these sites, we have our own login, so that we can freely re-install the software whenever we make changes, which we do all the time.

So while development is confined to the 'in-crowd' right now, he still intends to open it up for general participation, which I eagerly await. :)

For one the most recent information see http://www.hip.atr.co.jp/~ray/pubs/pub s.html

Try:

Avida

Tierra

That program is actually called Tierra and was developed by Thomas Ray back in the early 90's. Tierra is very interesting because it uses a custom machine code to generate programs that compete for resources. It reminds me alot of Corewars back in the late 80's. Running it on my machine at home generates some pretty interesting results. You can read about this and other Artificial life techniques in Artificial Life by Stephen Levy. (of hackers fame)

What you're talking about is The Tierra Project (I assume). The idea wasn't to let them just run across the internet though. They would travel only to computers that were running the Tierra Virtual Machine, feeding off of unused Processor Cycles. Basically the idea was to let these programs mutate and then "harvest" ones that had mutated usefully to build into software. I never could decide how sound the idea really was but it was pretty cool. http://www.hip.atr.co.jp/~ray/tier ra/tierra.html

Tom Ray did this about 8 years ago. His results were quite spectacular: Tierra's "critters" spontaneously developed self-replication, viruses, predation, symbiosis, and more! Here's a link to the Tierra Homepage.

has anyone ever used the Tierra artificial life software? I downloaded it years ago; hardly knew what to do with it, as I knew nothing about C.. got it to compile, but it always locked up my 386...

This article piqued my interest in the software again, and I found some info on it, for those interested...

links:

Web page
FTP Site
Documentation

has anyone ever used the Tierra artificial life software? I downloaded it years ago; hardly knew what to do with it, as I knew nothing about C.. got it to compile, but it always locked up my 386...

This article piqued my interest in the software again, and I found some info on it, for those interested...

links:

Web page
FTP Site
Documentation

I was a little disappointed that the authors didn't mention Tierra, the a-life system that Avida was based on. Tierra has been around since the 80s and can be found at: the Tierra homepage.

The programs that evolve in Tierra get pretty interesting, and include the evolution of parasites and virus programs. Pretty neat stuff!

DeGaris' web page.

Curious stuff.

--ac