What If Gamers Taught AI / Jade Helm Program Psychology?

a reply to: elliotmtl

You seem to forget.... Rule #1... If a man can make it, a man can break it.... 1st post....

a reply to: FissionSurplus

The more I think about about the more it makes sense.

When I was a kid I was playing flight 'sims' like MIG on Sega and enjoyed RC cars. Now the two have merged and the end result was UAVs.

However the Kinect system cannot spy on you if you disconnect it-the same applies to smart TV's-Samsung was caught in the act numerous times and that could've been avoided if people weren't so eager to connect every device they own to the internet.

a reply to: Aazadan

" Going against an AI in a brand new game, you'll search for ways to repeat behavior, and then abuse it. Going against humans is something else entirely because humans aren't logical, and unlike the computer they are not compelled to play the game to it's conclusion. "

This comment reminded me of a youtube video I saw a while back. If you google AI plays video games Im sure you will stumble upon it. A "AI" system was developed and it learns by playing video games. It started out by playing simple atari games like pong and moved on to some more advanced games. The interesting thing was that it used the games exploits to its advantage. Glitches and weak points, it did not know the weak points right away but it had to learn them just as we would.

Your forgetting that a AI at the magnitude the OP is talking about would be 1000's if not millions of times smarter then the average mad scientist. The AI is just as aware if not more aware of it self and surroundings then a human is, who's to say your so called "illogical" decisions cant be predicted? Although this thread is not really about humans vs AI just think about the being (the AI) that can sort threw all those emails, videos, threads, websites, game chat ect.... Its reading 1000's of gigabytes worth of information in a second, not only that but its aware so it can use that information. Its estimated that a human being can only read about 100000 books in its life time, this AI can do it in seconds if not a single second all while executing its other processes.

", assuming it could write an indepth enough report what it would find is that before I take any action "
there is no assuming its a computer AI it can

"I'm thinking about what area is most defended. In this case I think about 400 turns out."
The AI is thinking the whole game threw before the game even started

Dont think of it as a computer, think of it as a electronic God. Its far more powerful, and smarter then a single human being.

Name one video game where children are used as cover and enemy combatants.

Just one.

No seriously.

Just one.

a reply to: JamesCarrot

I'm a programmer so I really disagree with a lot of what you're saying. I understand how the machine works, and the limitations of programming.

When it comes to intelligence it depends on what you're talking about. Computers can learn information faster than people, but they're about the same speed at recalling information, and they are horrible at applying information, a human who has read 15,000 books can easily outsmart a computer that has read 1,000,000 books.

And no, it is not an electronic God. A computer will never consistently beat a human in any game where hidden information is a significant component. The closest you'll ever get is a game like poker where a computer will win 55% or so, and thus win the long game but still face a 45% chance of losing at any specific point.

a reply to: Aazadan

This Quantum computing not your regular intel i7 CPU. Its AI artificial intelligence, you know like the movie terminator or the matrix. Just because you have programing skills does not mean you know full scope of technology involved in jade helm. Do you know what the poster is even talking about?

Hears a thread about it
www.abovetopsecret.com...

A quantum computer is capable of processing more information than there are atoms in the universe in a second.
This is what quantum computing is capable of, there is no arguing its a fact. The human brain is not capable of producing a tenth (in its whole life time) compared to what this ai can do in a second.

We are not talking about regular computers, we are talking about quantum AI computers. A living thinking being.

Jade Helm demands more Hot Pockets and Mountain Dew.

a reply to: FissionSurplus

OK I went over to youtube to watch some Jade helm videos and what the...........Something very sketchy is going on and I had my cable go two nights ago for no reason given like a storm or tech issue. The youtube videos seem to be censored or something.....Freezing up and sound removed.



This might actually be happening and it is kinda freaky to say the least.

If this is true, why hasn't the army cannon rushed ISIS' expansions?

a reply to: blueman12

"You need more Gold"...zub zub

a reply to: JamesCarrot

Are you aware of how quantum computers work? The idea that quantum computers represent a million fold increase (or more) in processing power is only partially true. In a quantum computer each bit represents all possible states, however when it's measured (data being determined/provided) each qubit must be in a finite state of 0 or 1. By using some advanced probability calculates a quantum computer determines which state is the most statistically likely, then the next most, and so on. At that point each output is tested in order of likelyhood for a success. This makes them very good at solving specific problems that involve information that either doesn't change or changes slowly. However, this system also makes them very inefficient at determining the states and best possible course of action when the data is rapidly changing.

When scientists refer to quantum artificial intelligence, what they are referring to is the algorithms that determine which machine states are most likely.

a reply to: Aazadan

"However, this system also makes them very inefficient at determining the states and best possible course of action when the data is rapidly changing."

Well they obviously found a way to get around this.
Straight out of the PDF

" Beyond Reactive Planning:
Self Adaptive Software and Self Modeling Software in Predictive Deliberation Management
The purpose of this paper is to examine an approach to planning which extends beyond the traditional reactive planning state space. We present the following hypothesis: predictive deliberation management using self adapting and se lf modeling software will be required to provide mission planning adjustments after the start of a mission. Self adaptive software evaluates its own behavior and changes behavior when the evaluation indicates that it is not accomplishing what the software is intended to do, or when better functionality or performance is possible1.
Self modeling systems construct their own abstractions as a basis of computational intelligence2. In order to provide a proper process context for the evolution of software toward this level of autonomy, and in alignment with the proposed planning maturity models3, we put forth a concept of a NCW C2 Software Maturity Model. This new C2 software maturity model will take software beyond the service oriented paradigm into a new era of software designing its own replacements or modifications in order to satisfy new command and control requirements."

Great post.
some one teach this jade computer tic tac toe.
It can NOT win!

a reply to: JamesCarrot

The pdf you are referring to is this
www.dtic.mil...

It has nothing to do with quantum computing. You don't need a quantum computer to create a program capable of modifying itself. There are several programming languages which allow for the program to rewrite it's code as it's running but these systems aren't very efficient because computers cannot create their own goals, they have to be given specific parameters before they can attempt to optimize. For example is it solving for efficiency, cost, durability, or something else. You can't simply tell it to provide the best solution.

What that pdf primarily speaks of is the ability for a machine to rate it's own ability to solve a problem, and determine if it can do so within the constraints required, and if not, what it requires in order to do so.

originally posted by: buddha
Great post.
some one teach this jade computer tic tac toe.
It can NOT win!

Tic tac toe is a solved game, and it's one a computer plays very well. If played optimally the whole game is nothing more than who makes a mistake and doesn't make the correct move, if everyone plays optimally every game results in a draw. It's one of the easiest games to solve actually.

Here is a picture that solves it.
upload.wikimedia.org...

a reply to: FissionSurplus

All they'd learn is that no one works together and most people would rather troll the other players than play the game properly. It is a game, after all, so what should we expect? But definitely no secret military tactics are being learned from gamers. The only valuable thing gamers (or people in general for that matter) are capable of teaching AI is what not to do.

Bah, ran out of edit time
news.sciencemag.org...

This is an example of about where this sort of AI is at right now. Notice that it's not using a quantum computer. The program is able to nearly perfectly play a simple variation of poker where most information is public.

The types of predictive AI the military wants, are probably possible but they're a few mathematical breakthroughs away.

Oddly enough, the main breakthrough they need is on the P vs NP problem. If you don't know what that is, you can read about it here but the main idea is that it's a problem that's difficult to compute but easy to verify (traveling salesman for example)
www.claymath.org...

A quantum computer can perform the example task easily, because a quantum computer can brute force it as only a small number of permutations work out to true, and all are equally valid. However, this problem can ideally be solved through a formula instead, because brute forcing things is very inefficient, even with quantum computing.

Computer games are too unrealistic in my opinion to reflect how we would react in a situation, even with no respawns. Starting the game again is a type of respawn in itself. The only true model would be to say "no respawns, no playing again, in fact no console/pc". The extreme would be to actually wire the game to a suicide machine, your character dies - you do. I'm sure very few would be gunning it across battlefields or pulling stunts through hails of virtual artillery knowing they're going to snuff it. No magic bandages to restore missing limbs or being resuscitated by team mates and then seconds later engaging in ninja style hand to hand combat. Most people would probably hide and hope to be the last survivor.

Just ignore it. "A.I." is inevitable and the human species' contribution to 'life' will be contained therein. Flash and blood humans are far too vulnerable to live forever. I mean, we can barely travel in space without being poisoned to death from radiation!

Either way, nothing can be done to stop the march towards artificial intelligence. There are no 'John Connors' in the real world.