Google's DeepMind AI Plans To Take On StarCraft II (venturebeat.com)

An anonymous reader quotes a report from VentureBeat: Google and Blizzard are opening up StarCraft II to anyone who wants to teach artificial intelligence systems how to conduct warfare. Researchers can now use Google's DeepMind A.I. to test various theories for ways that machines can learn to make sense of complicated systems, in this case Blizzard's beloved real-time strategy game. In StarCraft II, players fight against one another by gathering resources to pay for defensive and offensive units. It has a healthy competitive community that is known for having a ludicrously high skill level. But considering that DeepMind A.I. has previously conquered complicated turn-based games like chess and go, a real-time strategy game makes sense as the next frontier. The companies announced the collaboration today at the BlizzCon fan event in Anaheim, California, and Google's DeepMind A.I. division posted a blog about the partnership and why StarCraft II is so ideal for machine-learning research. If you're wondering how much humans will have to teach A.I. about how to play and win at StarCraft, the answer is very little. DeepMind learned to beat the best go players in the world by teaching itself through trial and error. All the researchers had to do was explain how to determine success, and the A.I. can then begin playing games against itself on a loop while always reinforcing any strategies that lead to more success. For StarCraft, that will likely mean asking the A.I. to prioritize how long it survives and/or how much damage it does to the enemy's primary base. Or, maybe, researchers will find that defining success in a more abstract way will lead to better results, discovering the answers to all of this is the entire point of Google and Blizzard teaming up.

Comments

One huge difference

[Solandri]

Games like Chess, Go, Tic-Tac-Toe always let both players see the complete world state. Armed with that knowledge, it's easy to be systematic and deterministic.

Games like Poker and Starcraft hide part of the world state from each player, forcing them to guess at the parts they can't see. That opens up the possibility of one player bluffing - leading the opponent down the wrong decision tree because he's fooled into thinking the part of the world state he can't see is different from what it really is. I don't think this is something an AI can "solve". Certainly one could optimize it, so that it becomes damn good at guessing when a certain player is bluffing or not. But put it up against a different player and all that "learned" experience becomes useless, or even counter-productive. Or even pit it against the same player who's aware he's playing against the AI which beat him last time, and he'll simply do something he would never normally do to throw off the computer. It's a difficult enough problem that in pretty much all commercial computer games with a fog of war feature, the computer is just programmed to cheat by ignoring the fog and seeing everything.

Re:One huge difference

[Calydor]

This isn't so much about 100% solving as it is about learning HOW to solve. If the AI can go toe-to-toe with human players that's great, if it can't even come close then it's an area that needs more work.

Re:One huge difference

[Anonymous Coward]

I don't see why an AI wouldn't be able to solve it.
In Starcraft there is immense amounts of information but most of it have no strategic value. You don't need to know the color of each unit or what the ground texture looks like.
Some of the bluffs will just be more data that the AI shouldn't pay attention to. Instead it should use safe means to figure out what the opponent is doing.
Sure, there are bluffs like cancelling upgrades but that is only a matter of having to re-check what the opponent is doing instead of relying on a single scout.

Not that it matters. Starcraft favors the player that can click the fastest to some extent, especially when in comes to early rushes.
Being able to engage in multiple directions and pull away where the human responds and attack where the human doesn't means that the AI will win pretty easily.
The ability to concentrate on every place at once means that the AI doesn't have to outsmart the human strategically.

Re: One huge difference

[Anonymous Coward]

So advantage is more APS

Re:One huge difference

[Kjella]

Playing optimally does not mean you win every time. Take for example Texas Hold 'Em, no matter how poor a hand you have pre-flop (worst is 7-2 off-suite) against the best (pair of aces) you still have about 11-12% chance depending on colors and flush draws if you just shove every time and never see a flop. If it's your one-in-a-million lucky day you could do that six times in a row and win every time. Every poker pro - and most amateurs too - will have some bad beat story where they did everything right and still lost big. But in the long run it should work most of the time.

That's what should happen if you play optimally with a fog of war too, you don't have one static strategy where you build the same every time, at least not if the game follows basic rock-paper-scissors rules that should never work. You have a variation that mostly beats your opponent's variation and you hedge and switch strategies on the fly as your opponent's strategy is revealed to you. And of course that includes bluffing, but that's part of the call or fold decision. Assuming your opponent never bluffs and always bluffs both lead to very poor strategies, they bluff part of the time and you call part of the time on the same board.

It's not like chess where in position X you always want to do Y. But it's a lot more like the real world where you don't have perfect information and that's kinda the point here to make AI that can function instead of humans in the real world. They won't necessarily be better than our best, but they'll be more consistent in not doing things that are clearly sub-optimal, like say rear ending the car standing still in front of us. I'm not a good RTS player. I notice quite often that I'm simply not keeping up with the action, those should have retreated, those should have pushed the attack, that position should have been reinforced, those troops I build didn't get any orders and so on. Just consistently using what you have in a good way probably beats any strategy I got.

Re:One huge difference

[Anonymous Coward]

For "heads up" Poker, with one opponent, the AI can indeed solve, but you may not be very satisfied with the solution. Because the game is symmetrical, a "solved" game here means that the AI won't lose any money to you on average over time (in a real world casino game there would be "rake" to pay the house, so not losing to the other player still leaves you penniless). For the simplest, Limit Texas Hold 'Em game, where the AI doesn't need to work out how _much_ to raise by because the rules only allow fixed raise sizes, the statistics have been worked out by computer to a very good approximation. Cepheus is an AI that plays this strategy, you can't beat Cepheus over the long term, it isn't learning how you play, or anything like that, it doesn't care how you play, whether you're an idiot or genius, it makes every decision by looking at which options are statistically worth trying, then picking one at random according to how likely it is to work. e.g. Suppose it's in a situation where 15% of the time the right thing to do is raise, 15% of the time the right thing is call, 70% it should fold. A human might, if they've worked hard, be able to calculate the odds, and pick "fold". Cepheus will use a random number generator to pick proportionally.

So you can learn nothing by watching Cepheus play. Whatever it is doing in any particular hand is random, and yet, over time, unless you play an optimal strategy too you will lose money to it. It's infuriating, and indeed some people who think they're great at poker have been infuriated by it. "I bluffed Cepheus" they say, having raised with nothing and been rewarded by Cepheus folding. Alas, that's not because Cepheus was taken in by your bluff, it just did that by chance, and statistically if you keep raising with nothing you'll find that Cepheus calls you often enough for you to lose money. Better stop doing that.

Anyway, the same technique that powers Cepheus would in theory work for No Limits too, at Heads Up. But the statistical tables are truly vast for that scenario, we may not see it tried in our lifetime.

However, for two or more opponents, you can't develop a single optimal strategy. This is because not only might your opponents collude against you, but one of them might be markedly worse than the others, and you need to exploit that or get crushed. Because of the importance of relative position in Poker, your strategy must include this element which you cannot balance in advance. Try this simple example, you're Alex, the other players are Beth and Carol, you are seated in that order.

Beth is a poor player and afraid of you. If you enter a hand, or might still do so, Beth folds whatever she has, no matter how strong it may be. But, if you drop out, Beth raises All In with any hand trying to recoup her losses. If Carol is weak or borderline, she folds and Beth picks up a small pot. But if Carol is strong she takes out Beth. Now you're facing a player with double your chips, and are a huge statistical underdog.

Since we're talking about SC2, the AI players in that don't cheat except in the hardest difficulty "Insane" where they get free resources to make it actually difficult. The AI players don't understand high level strategy and any competent ranked player will crush all except Insane (because of the free resources) and can defeat that once they understand it has too many resources so must be attacked early.

Re:One huge difference

[Place a name here]

An AI can solve imperfect information games, it just is a lot harder. See, for instance this solution of heads-up limit Texas hold-em. Since the game has imperfect information and aspects of randomness, it's impossible to win every single time, but in the long run, the AI plays as well as or better than any other player.

Just how much harder it gets shouldn't be overlooked. Even imperfect information chess (Kriegspiel) would be pretty much impossible. Now imagine how much greater a game state StarCraft 2 has in comparison to chess... and I'm not holding my breath.

Re: One huge difference

[Anonymous Coward]

Hopefully the AI has some apm limit to prevent mkp style marine pwnage.

Re:One huge difference

[Ferocitus]

Nicely put!
I think there are a similar problems facing human language processing.
Slang, cant and argot change very quickly, and so it's very similar to a Fog of War for computers.
Teenagers, in particular, will change the meaning of words, or create words, or even maul grammar to include
or exclude others from their cliques. Irony and sarcasm are other bollards to progress for computers, innit.
AI Winter is Coming!

Re:One huge difference

[Anonymous Coward]

and why can't it learn to send scouting units for information the same way human players do?

Oblig.

[nospam007]

Let them play Global Nuclear War.

Re: Oblig.

[Anonymous Coward]

*Global thermonuclear war.

I was thinking this was more like we're headed down the path of teaching skynet

Re:Oblig.

[Anonymous Coward]

Yep. As if recent AI advances weren't scary enough without teaching them warfare.

Re: Oblig.

[Anonymous Coward]

They really arent. They are still acting within the limitations of their programming, and always will. This isn't any different than adding information to any other algorithm, it's still just software. The spooky thing is that some people are happy to turn responsibility for their lives over to what is in effect a glorified search engine.

Clever or fast?

[Anonymous Coward]

An AI that executes cheesy builds with ferocious speed and efficiency would be unbeatable by a human. Proxy 3 rax reaper with perfect control of every unit attacking in multiple locations at once and microing scv's at home in case of a counter, all at the same time. I don't think it would be that impressive tho... we know computers are fast.

Comment removed

[account_deleted]

Comment removed based on user account deletion

Shall we play a game?

[zifn4b]

How about Global Thermonuclear War? https://www.youtube.com/watch?...

Why not let it learn about high frequency trading?

[Anonymous Coward]

High Frequency Trading

I'm sure Wall Street would welcome the competition.

Civilization

[carvalhao]

I would be really interested to see what would the results be if you would get DeepMind playing a game like Civilization, in which cooperation and soft-power can be used to win the game. That could really give all of us some hints on how to manage diplomacy/belligerence in a way that could lead to some interesting thought experiments in the real world.

Re:Civilization

[drinkypoo]

I would be really interested to see what would the results be if you would get DeepMind playing a game like Civilization, in which cooperation and soft-power can be used to win the game. That could really give all of us some hints on how to manage diplomacy/belligerence in a way that could lead to some interesting thought experiments in the real world.

Not unless you made a game like Civilization which was more than a glorified board game. Civilization bears only the slightest passing resemblance to reality. As such, you can only learn the most superficial lessons from it.

Re:Civilization

[Anonymous Coward]

Yea, if only Gandhi wasn't so nuke happy in real life

Re:Civilization

[aliquis]

I wonder how long until an AI run the government.

Re:Civilization

[aliquis]

Well the one thing we know is that being restricted from "being racist" or expressing "hate speech" completely ruin your civilization, society, culture, technology and people.
So that's a sure path to a loss in the game of civilization.

Vote Trump!
Help us make EUrope.. well.. not worse and worse all the time again :D

Re:Civilization

[Anonymous Coward]

It already does.

Re:Civilization

[west]

I doubt a Civilization's AI's real-world applicability to real-life diplomacy, but a truly successful Civilization AI (i.e. one that played at Master level exactly as a human would) would terrify the hell out of me. Being a master of Civ involves managing limited information and about a *thousand* degrees of freedom each move, if not more.

It makes Go look like a cakewalk (in terms of the size of the decision tree).

If AI's can do that, then probably 2/3rds of the intellectual-related jobs on the planet are toast.

(As for applicability to real-life diplomacy, you have to be able to evaluate "success". Since there's no agreement as to what that even means, there's no teaching an AI.)

Re:Civilization

[Anonymous Coward]

I voted for Trump in Florida today! MAGA!

Re:Civilization

[Anonymous Coward]

a game like Civilization, in which cooperation and soft-power can be used to win the game.

That game tends to have different rules for the human than the AI - especially with the earlier installments.

For Civilization, it would probably find all the game exploits first. Things like learning that you can cancel and reorder settlers to quickly build roads, irrigation and mining, etc. Perhaps it would create armies of diplomats that can sit outside enemy cities to choke them of income. Perhaps other nasty tricks too - but that game is mostly known for using exploits for victory.

With other installments, it may find ways that likewise maximize effectieness of combat mechanics, or coast along the AI's diplomacy so that it remains borderline happy.

That could really give all of us some hints on how to manage diplomacy/belligerence in a way that could lead to some interesting thought experiments in the real world.

Not exactly. Early installments of Civilization tended to have the AI suddenly decide when to attack, and the other side might as well go all out in eliminating the aggressor since they will simply keep doing sneak attacks. Later versions may be different, but the game still feels as if you have to completely eliminate an enemy just to be safe.

Compare this to grand strategy games such as Europa Universalis, where you can theoretically exploit diplomatic tensions between countries, grabbing them while they're distracted in their own conflicts as opposed to them being upset about you rapidly expanding.

Not to mention that sometimes these tensions aren't properly modeled in a game. For example, the first French Republic had instant tensions from other countries because they were rather upset about royalty being deposed, something which would be much less relevant in a computer or board game as everyone else is more concerned about themselves rather than expected social structure.

Re:Civilization

[aliquis]

Thank you.

Over here in Sweden we can't change the outcome democratically any longer since our parliament with lots of parties has becoming like your solution except it's 7-8 parties which are joined together who are all responsible for what has happened and then the 1 party which has been against it and still are but they of course don't have as many votes as the others and considering the population is being replaced and lied to and you can't say what you want ..

So any change has to come from outside of the parliament and democratically means. Starting our own Swedish parliament for Swedes? Leave the Swedish state and establish our own within the same borders? Violent solution? Leave EU / EU influence how? Liberated by foreign nation? Lost and destroyed forever?

I don't know what the solution is really. 60 million refugees they say, few millions who has come to EU, and already we are likely many tens of millions of people in Sweden, Norway, Denmark, Germany, UK, France, .. who feel like refugees ourselves in our own nations fleeing from socialism, cultural Marxism and the aliens who has come to take our land, property and freedom.

Re:Civilization

[Anonymous Coward]

Civilization bears only the slightest passing resemblance to reality. As such, you can only learn the most superficial lessons from it.

For example, that spearmen have a 1% chance of defeating a unit of modern armor?

Re:Civilization

[Anonymous Coward]

If AI's can do that, then probably 2/3rds of the intellectual-related jobs on the planet are toast.

You begin to understand why universal basic income is both necessary and inevitable. The current situation in the middle east is a perfect example of what happens when too many young people are left with nothing productive to do and no hope for the future.

Re: Civilization

[Anonymous Coward]

At least one US tank was defeated by a Japanese guy with a sword in WW2. All the guy had to do is climb in and slash through the crew.

Re: Civilization

[Anonymous Coward]

The game Europa Universalis IV models this exact situation. When the revolution appears in game, all monarchies nearby suffer penalties unless they are at war with the revolutionary republic, who itself gets significant bonuses as a compensation.

Re:Civilization

[Anonymous Coward]

Go to Russia, they always need retarded useful idiots.

How about a real world use?

[Anonymous Coward]

What about trying to use it to solve the mess in Iraq/Syria/Turkey?

EULA

[Anonymous Coward]

So it's fine if Google does it, but if I do it, I get a ban for running a bot? Okay then.

It's a trap

[Anonymous Coward]

Glest is open source. If you tie your research to Blizzard, you will always have to ask for their permission to use your discoveries in any particular way. Tie it to Glest, and your AI can be integrated into the game itself. Then, other researchers can come along and try new variations without having to redo all of your work from scratch.

Skynet

[Anonymous Coward]

... how to conduct warfare.

In the not too distant future (about 2020) we will have robot guards run by a supercomputer, an AI that knows how to win a war-game, a heavy machine gun mounted on a military drone: Cue SkyNet in 3, 2, 1.

Not AI

[110010001000]

Playing games is NOT A.I. This, like "playing" Jeopardy, is just a flashy demo of algorithms and shows us how clueless AI researchers really are.

Re:Not AI

[Anonymous Coward]

Okay, what IS A.I., then? When do you use more intelligence than when you are playing games? If a chat-bot generated Slashdot posts that were moderated higher than your posts, would that be real A.I.?

Re:Not AI

[Anonymous Coward]

Clearly, AI is whatever computers can't do yet.

(FWIW, I think that even chess AIs can be called AI.)

Re:Not AI

[Visarga]

> Playing games is NOT A.I.

AlphaGo was playing a game. So ... it was not A.I.?

But even so, it was clearly something, what shall we call it, then?

Re: Not AI

[Anonymous Coward]

If a human played like go or blue you would say "you're playing like a fucking robot, make a joke or something ". And they would have nothing to say.

Selfaware AI is what we demand, damnit, so we can finally build our own friends.

Re:Not AI

[Ferocitus]

> Playing games is NOT A.I.
AlphaGo was playing a game. So ... it was not A.I.?
But even so, it was clearly something, what shall we call it, then?

AI, but not Artificial General Intelligence (AGI).

Interesting

[Anonymous Coward]

I wonder though when they manage to beat a human being, how much of it will actually be it actually being "smarter", and how much of it will be it just simply out-microing the human. That's always the problem with RTS games, the AI can be pretty damned boneheaded, but win anyway just because it's simply impossible for a human being to keep up with it wrt to micromanagement and general situational awareness.

Re:Interesting

[psycho12345]

They are addressing this directly, by limiting the APM of the AI to human levels. So it won't be able to rely on perfect micro to win.

Obligatory Archer meme...

[johnnys]

"teach artificial intelligence systems how to conduct warfare."

Do you want Skynet?

Because that's how you get Skynet!

Should have been DOTA2

[eddy]

My initial reaction when I saw this news was that it was a boring choice. It's a step up from Galaga, so maybe it makes sense as a stepping stone, but as a game something like DOTA2 would be much more interesting.

DOTA2, unlike SC2, heavily depends on both cooperation AND competition between players as an integral part of the standard game. It's got all all the same fog-of-war issues (i.e imperfect knowledge). There's for all intents and purposes one map, so you can focus on the strategy of the game without the changing topology of the map interferring. There's also a huge library of replays to learn from, with more coming online every minute.

It'd be interesting to know if DOTA2 was considered and rejected for some reason. Maybe it's as easy as Valve not being open to providing the hooks they need, but that doesn't sound like Valve.

I guess the positives for SC2 is that it's simpler, and they won't have to contend with too many changes/new units since it's at the end of its life.

Re:Should have been DOTA2

[majorme]

yeah, nice trolling bro your interest in team play is fine but doesn't exactly support your theory about how complex dota2 and sc2 are. controlling a single character vs a full blown rts? you can't be serious. i cannot sanction it

Re:Should have been DOTA2

[Anonymous Coward]

SC2 is a click-simulator. Click faster, win faster!
There isn't any deep strategies or complex tactics. Select a predetermined build path, build more, faster, and then click more to kill the other guy faster.
YOU WIN!

Re:Should have been DOTA2

[psycho12345]

Problem with dota 2 is the snowball nature of the PvE aspect. An AI there will quite easily crush any team because they will have near perfect farm, which eliminates any skill/strategy any human is going to come up with. A fully slotted in 20 minute Spectre/Void/Sniper is not all that interesting to win against

They also noted why SC was preferred, the team lead and presented was a former UC Berkeley student who worked on Overmind, the SCBW AI.

Simply put, DotA 2 is less taxing APM wise, since as you as you play safe and have perfect lane equilibrium control with your CS, it is rather difficult to lose. Even top pros can't deny and last hit every wave.

Re:Should have been DOTA2

[Anonymous Coward]

Actually, you've made a good point. Since DotA2 isn't very taxing APM-wise and it is easy to have the AI do perfect farm - whouldn't that be a near-perfect AI for them to develop?

No for competition, no. But for playtesting and balance. Send out two teams of bots, and let them play a few matches. Tweak attributes/costs/etc and repeat.

Re:Should have been DOTA2

[eddy]

The only way you get to "play safe" and "have perfect lane equilibrium" is if your opponents aren't doing anything. This is no different from any other game.

Farming is a much larger problem space in DOTA2 than in SC2, where there are already optimal or near-optimal strategies to number of workers, etc. The first few minutes of a SC2 game is just the same old boring mechanical opening shit. IMO YMMV.

Re:Should have been DOTA2

[eddy]

Yes you have a lot of units in SC2, but often they're controlled as a group. You can micro, but it's mostly about the strength of the combined force, not individual units.

An AI would have to be able to play a whole team, i.e control and coordinate five units from a pool of over one hundred unique heroes, some of which can summon and control separate units in turn (e.g Lone Druid) and have abilities that interact in various non-obvious ways (and where it would be interesting to see new combinations and counters evolve).

After all, the AI learns by playing itself or watching games, so it's not about playing a /single/ unit. It's about playing a side.

Re: Should have been DOTA2

[Anonymous Coward]

No. Games are balanced around human capabilities. A machine would not have the same skill distribution as us, it could capitalize on advantages humans cannot and vice versa.

StarCraft 2

[Anonymous Coward]

"Strange game. Only way to win is not to play."

Wil pay to watch

[majorme]

I've always been a multiplayer man myself with just about two decades of history fighting people in Quake and StarCraft. While bots work reasonably well with single player games, multiplayer is a different matter. We're yet to be presented with a real AI in a game. They all cheat. My point is I will pay money to watch e-sports where AI and humans fight on equal terms and we have no idea who is going to win. and how. my money is on the korean pro

Why bother

[_xanthus_47]

They are still going to lose to some Korean guy

Skynet

[Anonymous Coward]

Do you want Skynet? Because this is how you get Skynet.

I'm impressed...

[Anonymous Coward]

I must admit, I'm impressed. I didn't think their AI was at a level yet when it had a free will, much less it would select competing in a game of StarCraft of all things.

On the other hand, perhaps that is a sign of its level of intelligence.

How will DeepMind interface?

[berchca]

A lot of Starcraft is how fast you can do things that don't require you to be particularly clever (clever is a large part of it, too, but time to execution is key, as is the ability to physically deal with multiple areas on the board). It seems very unfair if DeepMind can do things like see more of the board than its opponent's monitor and move/select faster than someone could with a keyboard and mouse.

Re:How will DeepMind interface?

[iridium_ionizer]

The MIT Tech Review article stated that they will limit the commands per second to something in line with what a human (professional) player can do.

I am not 100% sure of what Deepmind's game awareness will be, but they do have a simplified graphics output for the AI (mostly just friend-foe, not the fancy artist made pixels).

Winning strategy found

[guruevi]

Zergling rush kekeke.

Not sure how you can defend against an AI that can simultaneously and individually control 200 units. Kind of like the mutalisk and pathing in SC1, SC2 is broken in many ways where it's easy for an AI to win simply by brute force and having superhuman unit control. Even Koreans and other high level players make many mistakes in optimization and strategy and simply make up by being physically faster than their opponent.

"teaching itself"

[Bobtree]

> DeepMind learned to beat the best go players in the world by teaching itself through trial and error.

AlphaGo was trained on databases of historical games. It looks for moves that are similar to what a human pro would play, and then reads out sequences to score the strength of the resulting position. It did not learn by itself from scratch. Once proficient, it was played against itself to improve.

Re:"teaching itself"

[Anonymous Coward]

It probably uses several instances of the AI and pits them against each other and then adjusts the model based on their individual performance. I'm just guessing, but I imagine they're using a reinforcement learning approach to accomplish this.

Re:"teaching itself"

[Anonymous Coward]

> AlphaGo was trained on databases of historical games. It looks for moves that are similar to what a human pro would play, and then reads out sequences to score the strength of the resulting position.

Human players do that, too. You're slicing a little too finely with your objection.

Blizzard legitimizes botting? Surprised.

[sethstorm]

So much for their commitment to integrity if they'll let a deep-pocketed botter buy them up.

Violent AI?

[Anonymous Coward]

How long before it starts telling us "These violent delights have violent ends"?

Skynet learns

[Anonymous Coward]

I read this as I watch Terminator Genisys.. What could go wrong ???

Re:Blizzard legitimizes botting? Surprised.

[Anonymous Coward]

Not sure if you're kidding, but DeepMind isn't going to be playing on the ladder. This is about AI research and development. Not grinding out fake points in a video game.

Without being crippled, Deepmind will dominate

[Anonymous Coward]

If this AI isn't crippled or rate-limited in input, it will wipe the floor with any human player.

Citation: Go play a "micro" map on any Blizzard game against the AI set to maximum difficulty and call me when you can kill a single AI-controlled unit. It's literally impossible to beat the computer, who is feathering units in and out of combat range perfectly. It's completely beyond human capability.

This test is retarded.

Slippery Slope...

[Anonymous Coward]

Are we sure we want to teach ai how to deceive us? Or how to deceive other computers? I'm sure this power will be used for only good.

As a starcraft2 player in training.

[Anonymous Coward]

I say bring it

War Games: just don't play

[Anonymous Coward]

" teach artificial intelligence systems how to conduct warfare"
There is no improving that solution

Trial & Error will never lead to human-like AI

[Anonymous Coward]

I can see it now. A robot walks into a store and asks the shopkeeper "A?" the shopkeeper looks puzzled and the robot leaves and returns only to go "B?". so on and so on until the robot puts together the right symbols to ask a real question. Only to follow up the response again with "A?". And on and on it goes.

What is Intelligence?

[bussdriver]

What we have is Applied Intelligence and as far as the Artificial kind of AI that becomes a big philosophical debate. Even so, what is Intelligence? If you start to define it in a way where you can build upon it logically, you end up with obvious conclusions where for example, your house thermostat is intelligent.

The domain or context is essential to the consideration too. So, the house thermostat is intelligent and within it's tiny world it performs quite well adapting and making decisions with it's simple stimuli. It is easy for us humans to think since we can write describe that intelligence in code or as a math function that there is no intelligence, just mechanically defined cause/effect. It is arrogant to define intelligence in terms of capabilities of humans... or just some subset of humans; plus practically speaking, it constrains us to only a select few to be the judges. The thermostat will not perform intelligently playing chess; just as human experts out of their depth do not perform intelligently either.

So lets say we have some math that is powerful enough to describe the problem of winning at Jeopardy. Do you honestly think that we or any math genius will be able to fully grasp that solution? So then do we call that Intelligence? It's "mechanical" but it does a better job than our human solutions. But why is that not intelligence? Because it can be described in some way and copied between machines? (we can't do that with our expert Jeopardy players.) No.

Lets go to a common fall back position: Humans can teach themselves without as much help a broad range of tasks. We have teachers, books etc... but why should we consider that way of learning the only way? Again, we are constraining it to humans. So... how far from "Idiot Savant" do we have to get before we consider it Intelligent? Again still constraining it to humans... What about mentally limited humans, like children? Do we let them off the hook simply because they grow up?

Getting back to the "some math:" What if you could describe incredibly complex real world problems found in life as complex math approximations? Well, that is just what we have been doing and the whole process of finding those mathematical non-linear equations is beyond our intelligence but for some problems our brains somehow do approximate solutions (unless you can find a perfect chess player, etc.) The amazing thing is that the math derived from theories on how our brains work is how we have algorithms which find approximate solutions - these are described as complex non-linear approximations (far better than human descriptions.) It is an iterative discovery process akin to our learning. So you might again say that this math is mechanical... but as we keep getting closer to mirroring human approximation abilities or surpassing them doesn't the trend make you wonder if everything in life can be described mathematically? Humans can describe solutions as math for simple problems but the machines do the work quicker. Is it unfair if a human teacher helps describe the problems and types of math (approaches) needed to learn the best solution to the task? Is it unfair that you learned your alphabet in a linear order? as a song? So do we give up at the point where we have automated the teacher?

Perhaps life is a non-linear approximation of 42... Do we know what the question was? no. does it matter? probably not. But we live in the process of approximating it. ;-)