Posted on 06/29/2016 9:30:29 PM PDT by Vince Ferrer
Artificial intelligence (AI) developed by a University of Cincinnati doctoral graduate was recently assessed by subject-matter expert and retired United States Air Force Colonel Gene Lee - who holds extensive aerial combat experience as an instructor and Air Battle Manager with considerable fighter aircraft expertise - in a high-fidelity air combat simulator.
The artificial intelligence, dubbed ALPHA, was the victor in that simulated scenario, and according to Lee, is "the most aggressive, responsive, dynamic and credible AI I've seen to date."
The application is specifically designed for use with Unmanned Combat Aerial Vehicles (UCAVs) in simulated air-combat missions for research purposes.
Retired United States Air Force Colonel Gene Lee, in a flight simulator, takes part in simulated air combat versus artificial intelligence technology developed by a team comprised of industry, US Air Force and University of Cincinnati representatives. CREDIT Lisa Ventre, University of Cincinnati; Distribution A: Approved for public release; distribution unlimited. 88ABW Cleared 05/02/2016; 88ABW-2016-2270
Journal of Defense Management - Genetic Fuzzy based Artificial Intelligence for Unmanned Combat Aerial Vehicle Control in Simulated Air Combat Missions
High pressure and fast pace: An AI Sparring Partner
ALPHA is currently viewed as a research tool for manned and unmanned teaming in a simulation environment. In its earliest iterations, ALPHA consistently outperformed a baseline computer program previously used by the Air Force Research Lab for research. In other words, it defeated other AI opponents.
In fact, it was only after early iterations of ALPHA bested other computer program opponents that Lee then took to manual controls against a more mature version of ALPHA last October. Not only was Lee not able to score a kill against ALPHA after repeated attempts, he was shot out of the air every time during protracted engagements in the simulator.
Since that first human vs. ALPHA encounter in the simulator, this AI has repeatedly bested other experts as well, and is even able to win out against these human experts when its (the ALPHA-controlled) aircraft are deliberately handicapped in terms of speed, turning, missile capability and sensors.
Lee, who has been flying in simulators against AI opponents since the early 1980s, said of that first encounter against ALPHA, "I was surprised at how aware and reactive it was. It seemed to be aware of my intentions and reacting instantly to my changes in flight and my missile deployment. It knew how to defeat the shot I was taking. It moved instantly between defensive and offensive actions as needed."
He added that with most AIs, "an experienced pilot can beat up on it (the AI) if you know what you're doing. Sure, you might have gotten shot down once in a while by an AI program when you, as a pilot, were trying something new, but, until now, an AI opponent simply could not keep up with anything like the real pressure and pace of combat-like scenarios."
But, now, it's been Lee, who has trained with thousands of U.S. Air Force pilots, flown in several fighter aircraft and graduated from the U.S. Fighter Weapons School (the equivalent of earning an advanced degree in air combat tactics and strategy), as well as other pilots who have been feeling pressured by ALPHA.
And, anymore, when Lee flies against ALPHA in hours-long sessions that mimic real missions, "I go home feeling washed out. I'm tired, drained and mentally exhausted. This may be artificial intelligence, but it represents a real challenge."
An AI Wingman: How an AI Combat Role Might Develop
Explained Ernest, "ALPHA is already a deadly opponent to face in these simulated environments. The goal is to continue developing ALPHA, to push and extend its capabilities, and perform additional testing against other trained pilots. Fidelity also needs to be increased, which will come in the form of even more realistic aerodynamic and sensor models. ALPHA is fully able to accommodate these additions, and we at Psibernetix look forward to continuing development."
In the long term, teaming artificial intelligence with U.S. air capabilities will represent a revolutionary leap. Air combat as it is performed today by human pilots is a highly dynamic application of aerospace physics, skill, art, and intuition to maneuver a fighter aircraft and missiles against adversaries, all moving at very high speeds. After all, today's fighters close in on each other at speeds in excess of 1,500 miles per hour while flying at altitudes above 40,000 feet. Microseconds matter, and the cost for a mistake is very high.
Eventually, ALPHA aims to lessen the likelihood of mistakes since its operations already occur significantly faster than do those of other language-based consumer product programming. In fact, ALPHA can take in the entirety of sensor data, organize it, create a complete mapping of a combat scenario and make or change combat decisions for a flight of four fighter aircraft in less than a millisecond. Basically, the AI is so fast that it could consider and coordinate the best tactical plan and precise responses, within a dynamic environment, over 250 times faster than ALPHA's human opponents could blink.
Since that first human vs. ALPHA encounter in the simulator, this AI has repeatedly bested other experts as well, and is even able to win out against these human experts when its (the ALPHA-controlled) aircraft are deliberately handicapped in terms of speed, turning, missile capability and sensors.
Lee, who has been flying in simulators against AI opponents since the early 1980s, said of that first encounter against ALPHA, "I was surprised at how aware and reactive it was. It seemed to be aware of my intentions and reacting instantly to my changes in flight and my missile deployment. It knew how to defeat the shot I was taking. It moved instantly between defensive and offensive actions as needed."
He added that with most AIs, "an experienced pilot can beat up on it (the AI) if you know what you're doing. Sure, you might have gotten shot down once in a while by an AI program when you, as a pilot, were trying something new, but, until now, an AI opponent simply could not keep up with anything like the real pressure and pace of combat-like scenarios."
But, now, it's been Lee, who has trained with thousands of U.S. Air Force pilots, flown in several fighter aircraft and graduated from the U.S. Fighter Weapons School (the equivalent of earning an advanced degree in air combat tactics and strategy), as well as other pilots who have been feeling pressured by ALPHA.
And, anymore, when Lee flies against ALPHA in hours-long sessions that mimic real missions, "I go home feeling washed out. I'm tired, drained and mentally exhausted. This may be artificial intelligence, but it represents a real challenge."
An AI Wingman: How an AI Combat Role Might Develop
Explained Ernest, "ALPHA is already a deadly opponent to face in these simulated environments. The goal is to continue developing ALPHA, to push and extend its capabilities, and perform additional testing against other trained pilots. Fidelity also needs to be increased, which will come in the form of even more realistic aerodynamic and sensor models. ALPHA is fully able to accommodate these additions, and we at Psibernetix look forward to continuing development."
In the long term, teaming artificial intelligence with U.S. air capabilities will represent a revolutionary leap. Air combat as it is performed today by human pilots is a highly dynamic application of aerospace physics, skill, art, and intuition to maneuver a fighter aircraft and missiles against adversaries, all moving at very high speeds. After all, today's fighters close in on each other at speeds in excess of 1,500 miles per hour while flying at altitudes above 40,000 feet. Microseconds matter, and the cost for a mistake is very high.
Eventually, ALPHA aims to lessen the likelihood of mistakes since its operations already occur significantly faster than do those of other language-based consumer product programming. In fact, ALPHA can take in the entirety of sensor data, organize it, create a complete mapping of a combat scenario and make or change combat decisions for a flight of four fighter aircraft in less than a millisecond. Basically, the AI is so fast that it could consider and coordinate the best tactical plan and precise responses, within a dynamic environment, over 250 times faster than ALPHA's human opponents could blink. That receiver - rather than standing still on the line of scrimmage before the play trying to remember all of the different specific statistics and what they mean individually and combined to how he should change his performance - would just consider the cornerback as 'really good.'
The cornerback's historic capability wouldn't be the only variable. Specifically, his relative height and relative speed should likely be considered as well. So, the receiver's control decision might be as fast and simple as: 'This cornerback is really good, a lot taller than me, but I am faster.'
At the very basic level, that's the concept involved in terms of the distributed computing power that's the foundation of a Genetic Fuzzy Tree system wherein, otherwise, scenarios/decision making would require too high a number of rules if done by a single controller.
Added Ernest, "Only considering the relevant variables for each sub-decision is key for us to complete complex tasks as humans. So, it makes sense to have the AI do the same thing."
In this case, the programming involved breaking up the complex challenges and problems represented in aerial fighter deployment into many sub-decisions, thereby significantly reducing the required "space" or burden for good solutions. The branches or sub divisions of this decision-making tree consists of high-level tactics, firing, evasion and defensiveness.
That's the "tree" part of the term "Genetic Fuzzy Tree" system.
Programming that is language based, genetic and generational
Most AI programming uses numeric-based control and provides very precise parameters for operations. In other words, there's not a lot of leeway for any improvement or contextual decision making on the part of the programming.
The AI algorithms that Ernest and his team ultimately developed are language based, with if/then scenarios and rules able to encompass hundreds to thousands of variables. This language-based control or fuzzy logic, while much less about complex mathematics, can be verified and validated.
Another benefit of this linguistic control is the ease in which expert knowledge can be imparted to the system. For instance, Lee worked with Psibernetix to provide tactical and maneuverability advice which was directly plugged in to ALPHA. (That "plugging in" occurs via inputs into a fuzzy logic controller. Those inputs consist of defined terms, e.g., close vs. far in distance to a target; if/then rules related to the terms; and inputs of other rules or specifications.)
Finally, the ALPHA programming is generational. It can be improved from one generation to the next, from one version to the next. In fact, the current version of ALPHA is only that - the current version. Subsequent versions are expected to perform significantly better.
Again, from UC's Cohen, "In a lot of ways, it's no different than when air combat began in World War I. At first, there were a whole bunch of pilots. Those who survived to the end of the war were the aces. Only in this case, we're talking about code."
AI can do deception?
Now I am scared
Unfortunately for AI, all it takes to whip its ass is for a human to walk over to the wall and unplug it. LOL!
I have beaten some of the most advanced AIs the gaming world has constructed, even ones that “cheat.” Yeah, I “died” a lot, but there is always a flaw you can exploit in the algorithm - even the “learning” ones (you teach them bad tactics then shift).
They may get there, but as for now put a 14 yr. old kid who has been gaming since age 6 - as opposed to some old guy like me - against the AI and the kid will shred it.
As you stated, you died a lot. What if each death cost you a 30 million aircraft? Can you afford to train it in bad tactics, and then switch? And what if that technique can only work at most one time, as the data can be transmitted back to HQ in order to train surviving aircraft to recognize that technique?
well, let’s see. computer vs. human: computer decides. It’s all software.
“ALPHA is already a deadly opponent to face in these simulated environments. The goal is to continue developing ALPHA, to push and extend its capabilities, and perform additional testing against other trained pilots. Fidelity also needs to be increased, which will come in the form of even more realistic aerodynamic and sensor models. ALPHA is fully able to accommodate these additions, and we at Psibernetix ...
Gee... that name sounds familiar.
*I* “died” a lot. Like I said, I’m an old guy pushing 50. But still managed to win.
Watched one of my nephews shred the thing that took me two weeks to beat. He crushed it the third time he tried (i.e., same day). Wasn’t even close.
As to your question, the answer is “yes” if you have a sufficiently strong economy and culture to absorb the losses until you go for the “kill.”
Will it necessarily work? Maybe not. Hope we don’t have to find out. Actually hope the AIs are on OUR side!
Ender’s Game.
http://www.imdb.com/title/tt1731141/?ref_=nv_sr_1
Don't be their target/objective.
bkmk
Unless it gets a chance to distribute copies of itself over the Internet first...
I think short answer is “retired” general officer, a chosen SME in a watered down industry that is about money not real results. Thats my personal opinion based on zero knowledge of the capabilities of artificial intelligence. If AI is indeed that advanced our development of ways to counter such with digital ordnance versus explosive ordnance would seem paramount. Very interesting read.....
You sir, are technically correct!
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.