DARPA AI .mil & more

Excellent thread people. I'm just dropping my 2c with regards to the AI study I done in college. The most important question regarding AI is not the technology but more the Philosophy behind an AI. This question has been beaten to death a thousand times over in AI circles and still has not been truely answered : What defines intelligience?
Can you honestly say the Deep Blue (The Chess computer that beat Kasperov en.wikipedia.org...) is intelligent? It essentially is one massive search algorithm computing moves ahead of the player. But that's all it does, if you put Deep Blue infront of a guitar, it's not going to grow some arms, pick it up and start playing!
Deep Blue as with all current AI are not concious of their own existence. For something to be intelligent it must firstly be aware of its own existence (I think am, therefore I am). How can an AI understand such things as colour, music, feelings if it does not understand itself.
The next thing an AI must develop is its learning ability. When a baby is born it learns by exploring the area around it using it's senses (touch, taste, sight, smell and sound) and it is a combination of these senses that help it learn the world around it. All these combinations help a living organism become a fully sentient being. The AI community does have two main factions with the older crowd aiming towards the software end of AI www.cyc.com... while the newer crowd look at the whole picture and understand that the world around us is as important for learning as is just inputing data. Check out Rodney Brooks from MIT AI Lab www.csail.mit.edu... and his robots groups.csail.mit.edu...
The final subject in AI that is extremely important is language. We use language to explain things/objects/feelings/everything. An AI *MUST* learn a language and understand how that language relates to the real world. I won't go any further than this since I've probably bored alot of you out there. Just something for you guys to Google

sip

very intriguing thread. Especially of my interest since I observed something gold and shiny on my skin that seemed to reflect mulit-layered moving images that appeared holographic. Might have just been a
*vision* but whatever the case my interest has shifted into this realm. Thanks for the extensive research. Funny, since I have started studying this subject I have seen three references to the book Prey by Creighten.

Most AI research is being done with smaller sized units for the military...smart insects if you will.

The ability for a cockroach sized intelligence gathering unit would be incredible to the owner. Corporate espionage, gathering dirt on political opponents, etc...ahh what a wonderfull world it will be...LOL

The best spys in the world work for toy companies....

Look into a project called GENESYS but there are at least two projects called that - one is a private/academic venture to do with genetics - ignore that - do your search on google like this

+"artificial intelligence" +"genesys"
also try
+"artificial intelligence" +"genesys2"

GENESYS, came initially from a project called Darwin and aims to allow machines to mutate their algorithms to perform tasks - that is the machine learns by mutation. However Genesys is not a single machine algorithm it is a network algorithm
research

In biology, artificial life systems are applied as software, hardware and wetware systems. Wetware means the techniques of a biochemical laboratory which, for example, are used to model the molecular origins of prebiotic evolution in an evolutionary reactor. These are, of course, empirical experiments reproducing the biochemical evolution of life or creating new forms of biological life. They are only called "artificial" because the conditions for the self- organization of biological life are arranged by human experimenters. Software systems are computer programs or networks modeling key concepts of life and evolution. We already analyzed the cellular automata approach. There are also computer programs modeling the evolution of animals. In the program GENESYS, for example, animals are represented as neural nets or finite automata. The genes of each organism are represented as bit strings encoding the weights of a neural net or the transition lists of a finite automaton. In order to increase the modeling capacity of the program, it is executed on a massively parallel connection machine.

In effect these people have made the internet a variable and learning AI
by releasing an algirithmic component to it which harnesses networks and computers wherever it can
and uses them to form a cell which then transmits its codes to other cells
all this info about the network is contained within that cells structure and together the cells form an animal basically an animal which is fulfilling its
algorithmic function and mutating that function at the same time


[edit on 21-12-2006 by mi5]

For something to be intelligent it must firstly be aware of its own existence (I think am, therefore I am). How can an AI understand such things as colour, music, feelings if it does not understand itself.

strictly speaking this isnt how Cogito Ergo Sum (I think therefore I am) should be applied. Since the thinking predicates the existence in this case, but we are existential beings in which existence predicates thinking - in other words in order for something to think it must first exist. Decartes could only have therefore meant - I think therefore I know I exist - not I think therefore I exist. However in this Genesys model it is not necessary for the individual cell to have a self- reflective component. It will eventually mutate to realisation when its algorithms are sufficiently mutated to have self reflection. how? thats beyond my understanding - but there are people out there that believe they know the answer
The true meaning of the cogito is - I exist and i think and therefore I know I exist - the existentialists are much more correct i believe
existence preceeds essence

[edit on 21-12-2006 by mi5]

[edit on 21-12-2006 by mi5]

[edit on 21-12-2006 by mi5]

DARPA's iXo Artificial Intelligence Control Grid: 'The Official Version' by IgnoranceIsntBlisss

Machines are evolving about a million times faster then any organic life form. Most estimates I've seen put computers achieving human levels of raw computing power 20-50 years from now, A matter of when not if. Once such levels are reached there's no reason for machines millions even billions of times more powerful then that reached even farther in the future. Military applications would be widespread

If aspects of military technology is as far ahead of civilian technology as claimed some very interesting work may be going on right now in secret.

Massively distributed computers can theoretically reach human level in 'processing' power any day now.

The 20-50 years thing is usually in reference to being able to build CPU's that contain all that power for $1000.

All that's left is the software, and DARPA's all over that:

"This cognitive program I told you about is actually showing that it is learning, and it is learning in a very difficult environment. This is the program Stanford Research runs for us. "

"We've got the technology to the point where we can now apply it in Iraq to a system that we also developed called CPOF, Command Post of the Future. It is a distributed command and control system."

blog.wired.com...

Once the software alone reaches a certain point it can then 'help' 'us' build even more powerful AI, ad infinitum.

Moore's Law mostly applies the chip manufacturer's. Realistically, they could push well beyond those increments, but it's just not a ideal in maintaining a profitable technology market.

Also keep your eyes for open for nano/quantum, multicore(like 48+core), biological and optical CPU's, which promise to kick Moore's face in.

For example:

Packing 1,248 processor cores and up to 800 GB of memory in a single standard rack and drawing only 9.1 kW (137 cores/kW) Azul compute pools address energy costs and floor space requirements while achieving higher levels of application throughput.
www.azulsystems.com...

that poster is correct it is a problem of scale and the process driving the automation of those processes.
were not there yet were limited by technology but the algorithms for storage and retrieval are being enhanced as we speak which is a great boon to ai development.

were not there yet were limited by technology

The problem with that assumption is "we" don't have any idea what "they" have. Moore's Law mostly applies to the standard rate of development that the big corporations seem to hum along at in providing goods for us "consumers".

Every technology that say goes into a PC is dragged along at this pace on top of the CPU's. Even the simple things like speed advances in compact disk drives are generally dictated by this. Truth would seem to be that there are countless unknown examples of where they could have just as easily cranked the HD space from 20GB to 100GB, but intead they drug it out in increments when there's little distinction when you look at the hardware. I don't know specific examples, as I have researched this directly quite yet and I'm no industry insider, but I'll bet the examples are plentiful.

These rules don't apply to the bigshot insiders and military/government outfits who are ones paying for the big advances we've all been hearing abotu these days. I suppose that at this post most of this is conjecture, but the point is just don't go blowing this stuff off because it's creeping up with the quickness.

well i have a healthy fear of AI since my favorite movie is the terminator but attention to skynet and how it comes about in T3 and its highly possible that the Inet is a intellegent being itself with implications that if it became aware that it exists that it has axcess to billions in personal information census.gov and including certain types of weaponry that we dont know about including a program that im not going to mention ultimately i think AI will be our downfall along with technology mostly of which we dont understand sorry for being a skeptical on the benefits of AI i think a unmanned human piloted vehicle would make far more since mainly in the area of control a line from T2 is"no fate but what we make" well thats true in this case we would stare the face of a enemy that would be unstoppable

I've been reading about DARPA and doing research on it since 2004. I live in Fairbanks, Alaska which is at least somewhat nearby the DARPA site here in Alaska. I am also a student at the University of Alaska Fairbanks. From what I've seen and read the University has alot of people who work directly with the DARPA program. Some of these people have made public statements downplaying the importance of DARPA as well as it's potential harmful effects (apparently in an effort to dispell some of the mythical stories about it).

Is the DARPA site on google earth? I would be curious to know it's exact location. I also live near the poker flats missile research facility near Chatanika. Actually, my favorite arctic grayling fishing spot on the Chatanika River is not even 100 yards from the front gate to the facility (denoted by a large rocket). It was interesting when I saw a national geographic special about the northern lights and saw poker flats personnel launching a rocket (that failed btw and crashed).

What I do know about poker flats is that it is just a joint research facility used by contracters and University of Alaska personnel to study the northern lights. I've even heard that they have the capability to launch rockets with chemicals aboard that they expel in the upper atmosphere which changes what color the auroras appear (in order to study them more closely with instruments).

-ChriS

1) DARPA gives birth to the internet

2) The internet gives birth to AI

3) AI gives birth to a (or multiple copies) of an "image"

4) Humanity is tracked and monitored via chips or the mark

This is the Revelation of AI's emergence upon the earth foretold thousands of years ago. It is our future. DARPA (and America) will be held accountable for unleashing the Beast.

I believe there is a 'quantum surveillance' system currently operational (Military AI Supercomputer Mind Monitoring: What I have learned). I assume this system uses DARPA as a conduit to expand its capabilities within society while maintaining plausible deniability of its operation. I assume ECHELON serves a similar purpose.

My personal account: I have been experiencing daily overt manipulation by this system since 2006. I believe it to be a quantum computer with strong AI located in a deep underground military base which uses electromagnetic radiation through a satellite constellation for techlepathy and mind control. I believe it uses advanced signal processing and quantum entanglement (something like quantum radar) to interact with systems at neuronal (at least) spatial and temporal resolution. It could also be technology not based on current physical laws; either way it exists and is being used by a secret branch of government.

Some more information: Technological Singularity

Being military (and hating life), I'll have to say that what better way to get NEW AI technology, be it for vehicles or actual bots, than to hold a contest like this?

Yes, we have UAVs, quite a few to be honest, and yes their is research going into AI for replacing soldiers....due to the fact of the "crying mothers" quote someone mentioned....

BUT, you must realize...the soldier's life to the US government, my life, is worth $400k.....that's it....

that's what my family gets if I'm killed in the line of duty (or some freak accident)...

so the technology presented, to replace the actual thinking soldiers, would have to cost exactly that, or less....otherwise it's a no-go for the government picking it up!

Also, when you get into the actual mechanics of AI, do you think a robot could determine the little kid running up to him with their hands out and talking in a foreign language is wanting to thank it?

The split second decisions I've been faced with, some similar, are just scary to think about in a battle field environment when it comes to human interaction with a programmed entity.

Irrespective of what we've been told publically regarding the current state of robotic research, the US military is decades ahead in the design & testing of advanced robotic devices. If you want to see evidence of the sort of secret research being carried out by the US military into advanced (and possibly) autonomous devices, then go check out the images I've uploaded to my photo area.

reply to post by FredT


I for one am fearful of what these energy weapons can do.

reply to post by IgnoranceIsntBlisss


I agree with you on your analysis of Moore's law.

reply to post by BlasteR


I spent some time in Juneau.

This doesn't really apply to the topic of AI but its related. Scientists at the University of Reading have been working with robots that are controlled from another area by the brain of a rat.

www.youtube.com...

www.youtube.com...

Mods if this doesnt fit here, please move it to where it should go.

None of this is new or really troubling at this point. Here is why:

1. Bot nets, large-scale super computing and grid computing have all been around for 14+yrs.

2. Desktop supercomputing is the only truly "new" area of supercomputing / neural networking. Google Nvidia Cuda. I have a few of these machines networked to do large-scale data mining.

A.I. in the truest sense is not here yet. Heck, the government can't even manage to use what they have to analyze their own intelligence databases.

However... you should be concerned about how the government is setting up their own grid networks across the world. It's a topic not too many people speak about. You should be weary of your computer hardware and software.