00:00
00:00
Rabid-Echidna
The spectrum always seems to shift back to the left. What a terrible stroke of bad luck, and things were just starting to go right.

Age 34, Male

I am the walrus

UCSB

Santa Barbara, CA

Joined on 9/10/03

Level:
43
Exp Points:
19,620 / 20,530
Exp Rank:
961
Vote Power:
8.36 votes
Art Scouts
1
Rank:
Brig. General
Global Rank:
464
Blams:
15,587
Saves:
5,357
B/P Bonus:
44%
Whistle:
Deity
Medals:
72

Constructing Our Own Matrix

Posted by Rabid-Echidna - December 7th, 2011


Posting my final essay, since DAVID requested that I do so.

-----

The Wachowski Brothers' film The Matrix takes an optimistic view of human ingenuity if nothing else, taking place in a future setting where AI has advanced to the point where it's overtaken the planet and reduced the human race to an efficient power source. Realistically this is a highly implausible scenario that only makes sense when extreme artistic liberties are taken when defining artificial intelligence, but the way in which the actual Matrix operates could theoretically be possible for the most part. Graphics technology is already at the point where a simulated environment running at sixty frames per second could easily be mistaken as reality at first glance. The computational requirements for producing the world are understood well enough that they can't be considered the difficult part of replicating The Matrix, the trouble comes when dealing with the metaphysical issues of creating a seamless human interface. The cybernetic implants seen in the film look like a mix between Star Trek's Borg and an H.R. Giger painting, but the concept of having what's essentially a big metal USB cord jammed directly into the back of our skulls is a bit more brutish than what might really be necessary for complete immersion in the virtual world. We already have prosthetic limbs which are capable of receiving the electric signals intended for the real arm, so the development of this technology might lead to the possibility of controlling an entirely prosthetic, simulated human being. Millions of people already immerse themselves in video games as a kind of fantastic alter-ego, and if we eventually figure out a way to completely close the gap between reality and simulation there's little doubt that we'll create with it our own personalized version of the Matrix which better suits our own purposes.

A test devised by Alan Turing is a good basis for skepticism when considering the plausibility of AI evolving and enslaving humanity. The artificial intelligence we have today is no doubt impressive, but it's not something which could actually be considered real thought. The machines in The Matrix have somehow designed and built their own massive supercomputer to run the program of the Matrix indefinitely along with towering harvester bots, sentry drones with tentacles, and all kinds of other high tech stuff that would never be able to spawn from artificial intelligence. One of the most convincing attempts at passing the Turing Test to date was an instant messenger chat bot named SmarterChild. Users could send SmarterChild messages and it would respond to basic conversation in a relatively coherent manner for the most part. It was also a great example of the limitations of the technology, as any sort of extended conversation would eventually cause it to start repeating phrases and spouting indecipherable gibberish. The limitation of creating anything like the AI present in the film is that by definition computers are never going to be able to think. They are at their most basic level a machine which is only capable of input-output functions, and all the attempts to program a robot to converse with a human are only following an established set of parameters. The concept of a system of artificial intelligence moving beyond a role of data retrieval is antithetical to purpose to begin with.

This limitation does nothing to hinder the integration of human beings into a simulation resembling The Matrix, it simply means that it will be of our own volition rather than something forced upon us by our fictional robot overlords. The ability to render something closely resembling the real world has already been possible for about a decade, at least in terms of pre-rendered graphics. A movie like Cars 2 is procuced entirely from computations worked out at Pixar's massive render farm, but the usage of new special effects clearly demonstrates how close we've come to erasing the line between artifice and reality even if it does require a 200 million dollar budget. The more modest home computer is much less powerful, but can still utilize the same effects on more restrained settings and produce a real-time environment which is close to indistinguishable from reality in its own way. Graphics card technology generally progresses along with updates to Microsoft's DirectX programming interface, and each update tends to bring with it a few new tricks
which can be utilized in graphics engines to make the simulated world seem more realistic.

Demonstrations of the new graphics engines make it clear how important lighting and shadows are when attempting to successfully imitate real life. Some of the issues with lighting have been worked out for the most part with the advent of a process called ray tracing, which applies a computational model for the source of light and then fires clusters of light rays which can then interact with surfaces and transparent materials accordingly. This is closer to the way real light actually behaves, and produces a more convincing final effect especially when mixed with ambient shadow technology. One of the main theoretical issues with trying to successfully model reality is that you have to deal with a seemingly infinite amount of data. The way to render a chair in a video game is to take a series of polygons, stick them together in a way which resembles a chair, them map a chair texture to that model. This is entirely different from what actually constitutes a chair in real life, where one would have to account for the individual pieces of wood, followed by the chemical bonds in that wood, the interactions between all the individual atoms in that chair, and so on for as far down as our current understanding of molecular physics can take us. The most recent line of graphics cards also have their own special fix for the issue of modeling reality with a process called dynamic tesselation. This at least deals with the issue of modeling something out of polygons by taking those polygons and breaking them down further to better adhere to the ideal shape of the object. Attempts to make this chair real could then be furthered by applying high resolution textures to the chair to make it as close to real as the human eye can detect, and then apply a new physics technology called PhysX to it to imitate real wood splintering.

This is all just based on the technology that's already available, and what makes these techniques special is that they're all limitless as far as computational boundaries are concerned. Ray tracing can be done with any amount of light beams and will yield better results the more rays are sent out, tesselation can divide the polygons over and over to produce a more detailed model each time, and PhysX can offer varying degrees of realistic physics depending on how much
computational power the computer can realistically handle. There comes a point of diminishing returns as far as how convincing a scenario appears to the naked eye, and any Matrix, either built by man or machine, would logically choose this point to stop adding complexity beyond what can actually be percieved. The infinite nature of the Universe can be ignored for the most part as far as a realistic simulation goes, since every new rendering trick carries with it a point where we could say that no visible distinction exists between the real thing and the simulation. At that point we can simply label it as "highest setting" and move on to the most efficient possible usage of graphics cards to process things like realistic water physics, anti-aliasing, Screen Space Ambient Occlusion, and depth-of-field. Assuming a reasonable price for a home desktop computer today is around $1000, we're at the point where it's feasible that anyone with that kind of expendable income can build their own system which will come close to utilizing the previously listed effects to their full potential. The most graphically intensive game on the market is arguably Crysis 2, which runs off a new graphical engine known as CryEngine 3 and includes just about every known special effect to date. As was the case with the first Crysis, at the time of its release there wasn't actually the technology available to play it at anything near highest settings, but if Moore's Law continues to prove true it should be possible in about two years for any common person to turn them all to the highest available value for the standard price of a home computer and have their own Matrix-grade synthetic world which they can navigate in real-time.

The problem with making a direct connection to The Matrix is the issue of of the interaction with this newly created world. Video games are a far shot from being able to actually transport a human being into them, which is a feat that will most likely remain impossible on account of an inherent problem with the idea of transferring a human conscsiousness from the brain to an electronic device. It's similar to the problem raised with the idea of a teleportation chamber. You could theoretically transport all the atoms which constitute a person from one place to another through the wonders of some yet-undiscovered technology, but it's hard to say that what comes out
the other side is really going to be the same person. Even in a world where neuroscience had advanced to the point where we could fully understand the processes which power the human brain, there would be no way to actually take that consciousness out of the person and move it into a virtual world since that consciousness is directly connected to the brain matter. Even if the process was somehow successful, the result would rightly be classified as an act of killing the subject and then uploading an exact replica program of their personality to the virtual world.

It's for this reason that The Matrix might be conceptually impossible, but it's still easy enough to see how we might get close when considering interface technology. Though video games are a good parallel to draw in terms of the visual representation of the world, they are entirely different as far as a realistic representation of reality goes. They focus more on completing arbitrary objectives than attempting to recreate real life, since our input devices don't allow for any sort of real physical presence in the virtual worlds. We only control the characters through predefined commands such as "run forward" or "examine object," whereas the ability for actors in the film to directly interface with the Matrix and download information directly into their brains demonstrates a kind of linkage with technology that isn't within the realms of what should ever be possible. This limitation only applies to direct, two-way interaction between the human mind and technology, and doesn't rule out the possibility that we could still issue commands by thought to an artificial representation of ourselves within the virtual world. Our of our voluntary bodily functions could be seen as another set of predefined commands as well, and recent advances in prosthetic limb technology and demonstrations of signal capturing show that a direct interaction with the virtual world is at least possible even if full projection isn't.

Scientists working with the limbs have effectively worked to decode the signals that the brain sends to the muscles and apply them to robotic constructs instead with great success, but perhaps more interesting is that researchers at Duke University are also working on technology which would allow the transfer of sensory information back into the brain such as feel and
temperature. This means that we could also recieve tactile feedback from the actions our characters performed in these virtual worlds, and feel a direct connection with the artifice even if we're still well aware that we're sitting on a couch at home. The final barrier would seemingly be a way of capturing the motion signals sent by the brain before they can make it to the body and redirecting them directly as inputs to the virtual character, which would work out the problem of having to move through space in a way which corresponds to the character. Ideally we could make it so we told our arm to move, and our characters arm moved in the simulation while we remained perfectly stationary in the real world.

Ultimately, we're well on our way to completing the goal of producing our own Matrix, and a good deal of early video games could be considered early prototypes for the final model. The imitation of image is already well enough understood and feasible with current technology, though there are few games which can actually take full advantage of the limits of our graphical technology. The Matrix as a film could essentially be summarized as a bunch of robots forcing everyone in the world to spend all their time playing one giant MMORPG in the same vein as World of Warcraft, only the interaction between the representative body and the human is absolute and the world is modeled around human civilization before the great collapse instead of one in which goblins run around throwing fireballs at one another. If computational advances continue at their current rate as they have over the past half a century and the barriers between interaction with a human and a computer can be pushed to their theoretical limit, we could end up with some sort of at-will Matrix of our own, entered by attaching a device which redirects and recodes our own attempts to speak and move through space as their corresponding actions in the simulation. After that it would simply be a matter of running a video feed from the computer to some sort of wrap-around visor with a monitor on the inside, and we could explore and look through our entirely artificial worlds at will while sitting completely motionless and unresponsive in our homes.


Comments

Comments ain't a thing here.