Slashdot Mirror
AIs vs Humans - Next Battle: Starcraft (businessinsider.com)
braindrainbahrain writes: Having conquered checkers, chess, and more recently Go, artificial intelligence research now looks at the next frontier: the popular real-time strategy game of StarCraft.
Blizzard Entertainment's president reached out to Google's DeepMind researchers last month, who are now describing StarCraft as "our likely next target". But many top StarCraft experts believe AIs will fail because "Unlike machines, humans are good at lying," reports the Wall Street Journal. An executive at the Korea e-Sports Association tells them "It's going to be hard for AI to bluff or to trick a human player."
One University of Alberta computer scientist David Churchill counters that âoeWhen the AI finds that the only way to win is to show strength, it will do that. If you want to call that bluffing, then the AI is capable of bluffing, but there's no machismo behind it." Unfortuantely, for five years Churchill has been running AI-vs-human StarCraft tournaments, and "So far, it hasn't even been close... Using a mouse and keyboard, the world's top players can issue 500 or more commands a minute," the Journal reports. But they add that now both Facebook and Microsoft are also working on small StarCraft AI projects.
Blizzard Entertainment's president reached out to Google's DeepMind researchers last month, who are now describing StarCraft as "our likely next target". But many top StarCraft experts believe AIs will fail because "Unlike machines, humans are good at lying," reports the Wall Street Journal. An executive at the Korea e-Sports Association tells them "It's going to be hard for AI to bluff or to trick a human player."
One University of Alberta computer scientist David Churchill counters that âoeWhen the AI finds that the only way to win is to show strength, it will do that. If you want to call that bluffing, then the AI is capable of bluffing, but there's no machismo behind it." Unfortuantely, for five years Churchill has been running AI-vs-human StarCraft tournaments, and "So far, it hasn't even been close... Using a mouse and keyboard, the world's top players can issue 500 or more commands a minute," the Journal reports. But they add that now both Facebook and Microsoft are also working on small StarCraft AI projects.
For instance, micro marine so they never stay closer than their maximum range... or any unit in fact. I'm also looking at tank and medivac drop... I would see a deadly combination here. But overall strategy, I don't think AI is ready to be human... yet.
Is it AI or A-with-a-hat-followed-by-oe-I?
With EditorDavid you can never be sure. At least manishs has the excuse that he's an H1B.
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
First off it's very easy to write an algorithm that lies and misinforms when optimal. Second, and this is a joke, have you ever seen a progress bar be accurate when downloading or installing something?
Just look at how passionate about chess so me of their kind is.
They should use something like Street Fighter II instead. 1 vs 1 and if you lose, it's because you messed up.
Like how they support "space."
Data couldnt beat Kolrami, so he forced him into what would have been an indefinite stalemate. Kolrami found this incredibly insulting and forfeit. Data won by having no ego. He busted him up.....
Good-bye
It seems like an AI would be really susceptible to being "trained" to react in a certain way by a player, who could then take advantage of that by sending up fake signals early and doing something that takes advantage of the anticipated AI response.
That may seem the same as "AI's cannot lie", but it's actually more about an AI being more susceptible to bluffs than a human player would be.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
Space is just corporate welfare that takes from the people to give to the corporations.
Because beating Starcraft will help supply people with clean water...not!
Blizzard is a dying company. HOTS is a huge flop. Overwatch will also be a flop. Starcraft is basically dead. Diablo has a following and Hearthstone is the only real hit they have. WoW is kept on life support by the fanboys. After the next expansion flops again it will finally die.
But you are part of the problem. So often the political posts bring just as many page views from disagreement as agreement.
They incorrectly view being pro-science as being anti-liberal arts.
Unfortuantely
But the US will just print more money to cover the debt, so saying we have a spending problem is disenginuous.
Disingenuous. Pelosi has used that word several times to describe the lie that we're spending too much money.
They should do this with Total Annihilation and see what kind of behavior the AI comes up with.
There was some fun, and devious, stuff in TA -- like using your robot troop transport to fly into an enemy base, kidnap the Commander, and then self destruct.
Tactics that weren't documented, and only emerged in the community over time as they were discovered.
Every games will be over in 4min as the computer just cheeses everyone.
ALS v starcraft?
But people need to know that they must construct additional pylons.
... SJWs ...
DRINK!
The AI emperor needs no clothes...
I'm a CS Masters student doing a thesis on a RTS AI. Computers can beat humans, we just haven't tossed enough CPU at it yet. RTS are exactly the same as checkers, chess, go, etc... except you have more pieces, more board positions, and more than one piece can be moved per turn. To reduce that into something computable, you need good abstractions. Once you have those the game becomes a tree search, same as all the board games. Google/IBM can bring enough computing resources to the table to win. There are some bumps in that: imperfect information, teams, etc... but they don't change the core algorithms.
Computing the entire game tree is too expensive. They'll probably do it at a unit/battle level, at a squad level, at a city level, and at a long term strategy level. Doing things at different levels greatly reduces the search space. From your training data you'll know how well you can expect the battle manager to handle an upcoming attack with an expected loss of XYZ at some probability, so the strategy component doesn't need to bother with all the minor details of how to fight it.
500 commands a minute? That's nothing. With the computing resources of a super computer, expect the AI to be able to issue an order to every individual unit every game turn. And yes, at the game engine level all real-time strategy games are actually turn based.
When you have the resources, a tree search over a game's state space with a little bit of memory (so the enemy can't get your units stuck in a circle) is effectively unbeatable.
The following is an AI unit positioning heuristic for determining which locations are good points for units to hangout (assuming the AI wants to mass units and not just individually control each one for perfectly executed, constant attacks across the entire border. Humans group units into manageable sets. An AI with a lot of processing power behind it doesn't need to do that.):
1) Does the enemy know we can build these units?
2) Has the enemy seen many of these types of units?
3) If 1 and 2 are false, then hold the units at a spot not visible to the enemy, assuming the unit isn't needed for something else.
4) If 1 or 2 are true, then put rand() percent of the units in an enemy visible location and the rest in a non-visible position, assuming the unit isn't needed for something else.
Where's my million dollar grant? You can do a similar thing to draw defending units away from a base. However this is too high level. Google will just train a massive neural net-based tree search on millions and millions of games with a large selection mini patterns. If it looks like this, then consider this best option that worked in the past if we can't come up with a better one within the calculation time. Really, that's a generic, high-level way of solving any problem. You just need enough CPU power to do it.
How about they work on writing an AI that can play a competent game of Civ V without cheating.
RTS is much less interesting since a big component of RTS is actions per minute/reflex based. Of course a computer is going to be better at that.
a significant limiting factor in playing starcraft is players ability to manage a large number of units at the same time. with a machine, speed is not an issue, so it will always win by that measure. chess and go didn't have a time component to them, so it was purely about strategy. the only way to make a battle of starcraft a fair fight over intelligence is to slow the game down to requiring both players to agree to move on to the next event cycle (aka "tick") in the game. it would be an absurdly slow game but intelligence isn't about how fast you can do something, it's if you can do it at all.
Anons need not reply. Questions end with a question mark.
Starcraft is not as calculation heavy as a game of chess, it is more of a test of human efficiency using a mouse and keyboard than anything else. Understanding that, a computer can do infinitely more commands per second than a human could ever do, so most traditional understanding of strategy would be thrown out the window.
Hopefully this demonstrates why a computer would simply trounce any human without even real AI. Strategies could simply be pre-scripted and unstoppable.
Sorry BoxeR, but you will be beaten.
How about beating humans at something like Cards Against Humanity? Or Werewolf?
File under 'M' for 'Manic ranting'
Maybe the AI could have a hard time tricking humans, but at the same time humans will have a hard time tricking the AI. Odds are that the AI will not be lured into traps if they appeal to humans.
One of the first times I played Command and Conquer with multiple players, I ended up winning by deception, which for sure wouldn't work on an AI. Most players were dead and we were only two left. He attacked me over and over and each time I sent forward my tanks and all tanks on both sides died and it ended up as a draw. I decided on a chance of strategy because my tanks would not overpower his tanks. I started building helicopters in the corner I knew he couldn't see and kept them hidden even during attacks. Them being secret was the major factor. I built tanks at the same time and more combats ended with a draw. At some point he attacked, I set all my tanks to attack move to a point behind his tanks and they would likely behave ok on their own after that. I then set all my helicopters to circle his attack and went directly to his base. He was busy with my tanks and that battle generated lots of "you are under attack" warnings, meaning he didn't realize what happened to his base until the tank battle was over. At that time he had no tank production facilities left.
My point here is that I intentionally attacked while I knew he would look at the tanks and I knew the attack warnings would drown in all the other combat messages. Even better, exploding buildings would be hidden by the sound of exploding tanks. Because the sounds were so important for the distraction, I don't think I would be able to fool an AI by doing the same thing.
Due to being a LAN party, the dead players had gathered around the still alive players and it sounded really funny when the other "team" realized what was going on. Luckily the people who watched what I did didn't talk about it at all. One started a little bit and I hinted "no talking about my surprise" and they shut up. Oddly enough despite the number of LAN parties I went to after that, I never encountered a C&C game where it really made sense to repeat that strategy. Most likely because I never encountered somebody who didn't try to change strategy when attacks kept on failing.
... is they can't issue 500 commands per minute (8.33 commands per second) like human players? Surely issuing large numbers of commands in a short space of time is not where AI falls short. If you preloaded a plan involving issuing 500 commands into an AI, it would execute them at the fastest speed the game code could keep up with, whatever that may be in Starcraft. Working out what commands they should be issuing is the problem.
Imagine how much money you're going to make when you short their stock. You'll be rich I say! Rich!!!
Simply not true. The level of SJW has always been the same. But with more assholes like you labeling anyone who calls an asshole as an SJW (even if, or especially if true), the instances of accusations of SJWs by assholes has increased greatly. Use of the word has increased greatly, but not the people who are incorrectly called that by assholes.
Learn to love Alaska
Blizzard is a dying company. HOTS is a huge flop. Overwatch will also be a flop. Starcraft is basically dead. Diablo has a following and Hearthstone is the only real hit they have. WoW is kept on life support by the fanboys. After the next expansion flops again it will finally die.
You'd best stick to gaming because with predictions like that your budding career as a financial analyst is likely to be undistinguished.
The API is for the Starcraft Broodwar. If anyone knows of an API for the more recent Starcraft II, please post.
Blizzard is a dying company. HOTS is a huge flop. Overwatch will also be a flop. Starcraft is basically dead. Diablo has a following and Hearthstone is the only real hit they have. WoW is kept on life support by the fanboys. After the next expansion flops again it will finally die.
You'd best stick to gaming because with predictions like that your budding career as a financial analyst is likely to be undistinguished.
While I don't completely agree with the guy, I don't think he's completely wrong either. I don't think HOTS will be successful and Overwatch is a big question mark. But in both case, I got the "Too little, too late" feeling.
But he's wrong about Diablo 3, with 30 million copies sold I can hardly call it a flop. Also, the WoW is heading to it's end. But I've no doubt that they could make WoW2 successful when the first onewill die.
Elok
Comment removed based on user account deletion
Keep smoking that crack.
Starcraft is too much of a toy world to be convincing of AI capabilities.
A grander challenge, but still toy-world, would be some smaller games, like the Space Station 13 series. Unlike Go, Star Craft and other games, SS13 success relies on some degree of co-operation between players.
Teenagers, especially, use natural language and anti-language, with words and expressions that only small groups within the "in-crowd" understand. Once adults (or AIs?) start using the same groovy words, kids will often then start using other words and phrases.
I'd love to see how an AI performs in games where it doesn't even know all the meanings of words, or where the opposite meaning of a word is obvious to a closed group.
USB, USB, USB!
Starcraft confirms it.
This comment is fully compliant with RFC 527.
Comment removed based on user account deletion
Problem with Hacker news is they don't have a funny mod, so everyone is always uptight, because if you don't say something insightful, you'll get voted down.
It's a lot more relaxed at Slashdot.
"First they came for the slanderers and i said nothing."
Comment removed based on user account deletion
I mean seriously. ha ha hahah HAHA HAA H HAH AHAHAHH HAHHAHA ha he he heh heh.
Chess players ,jeopardy players, go players all laugh and weep at the same time in sympathy for you.
She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.
Actually, I bet a sufficiently-well-trained AI could win at CAH (for values of "win" equal to "get the most black cards"; in practice everybody wins in a good game of CAH). Even without a camera watching player expressions and so on, an AI can learn combos that work well (and who they work well for), and see all kinds of relations between cards in terms of how different players react to them. It would take a lot of training - quite possibly an infeasible amount - to be good enough to beat *arbitrary groups*, but a specific group? Yeah, it could do it.
There's no place I could be, since I've found Serenity...
While I don't completely agree with the guy, I don't think he's completely wrong either. I don't think HOTS will be successful
HoTS is over, and it was successful. The most recent release of Starcraft was LoTV, and it was also a huge success. Furthermore, it's brought people back to the game who haven't played in nearly half a decade. Furthermore, LoTV is tons of fun.
"First they came for the slanderers and i said nothing."
As far as I understand micro in games like Starcraft, a computer ai should have a huge advantage. In a battle a human player will usually direct the fire of several units at once. A computer would direct every single unit and point it at the optimal target and also move away units that have taken fire. Starcraft II marines do a stutter type movement, for example, where they use the pause between shots to move, loosing almost no fire time. Thus a computer player would have perfectly balanced health, every time one unit loses health it would be moved to the back. A human player would thus lose units a lot faster. Also a computer player could individually target the best places and units to shoot for every single shot. Since some units have better attack against other units, this would make another difference.
All those problems are solved already. You don't need new ai research, you just need the time and work to put in algorithms that know which unit should prefer to attack which other unit and when it is best to move back a unit a little (when it is under attack itself).
So my question is: Since Starcraft is half about micro and half about macro, doesn't that make a game between a computer player and a human player boring, since the computer has such a huge natural advantage in micro? Or will the computer in Starcraft be forced to mimic the human disadvantage in micro in some way?
Since this is your thesis, what do you think? OTOH you could enable the some of the micro algorithms (that your units automatically attack the weakest enemy, for example) for the humans as well, but that would hurt the human players, since they aren't used to that.
I just don't think Starcraft is a good game to play against a computer in that sense. And I wonder why the computer isn't better at it right now.
"A machine isn't good at lying."
Yeah right. Until some college students find the ultimate "SuperCrushRush(TM) StarCraft Bluff Algorithm" and their box mops the floor with every number of human opponents in a game of SC.
It's a modern microcomputer people! ... This isn't news, this is blatantly obvious.
Take one, give it insane specs, a small army of engineers and a few years time and they will find a way that the machine outperforms every human at a very specific task.
We suffer more in our imagination than in reality. - Seneca
Overmind.
We suffer more in our imagination than in reality. - Seneca
With the influx of liberals and SJWs here the past few years, political stories get a lot of comments and views. I can't blame the new owners for doing something that greatly increases their ad revenues.
Need some ethics with your gaming journalism there do you.
HoTS is over, and it was successful.
was? Is the past tense intentional?
Either way, I guess it depend of your definition of "successful". In my mind, if they don't get even close of DOTA2 and LOL number, it won't be a success and it won't survive for long.
As for Starcraft II, I've read a number of a few million but I wasn't able to find it's total number of sales. But, keeping Diablo III 30 millions in mind, I wouldn't call it a huge success, Diablo III was.
Elok
At least it would be impossible for an AI to beat the best Starcraft player. An AI can't look 5000 moves ahead to pick the best move in Starcraft. It takes a lot of skill to play Starcraft at a high level.
Just impossible to program an AI to deal with all the variables and deceptions. Unless you program it to do random things in the hopes of bring right once in a while.
Space is racist, those dirt republicans just want to leave blacks and minorities behind on earth. It's the new form of white flight.
Space is black motherfucker.
Wanna buy a shirt?
https://www.redbubble.com/people/stealthfinger/shop?asc=u
HOTS was a huge success, because it was almost entirely an accident. Very little actual investment, and a pretty good return.
The spaces AI on many online versions of the game are good at lying. Especially if they say they will cover you when you, their partner, goes null.
I only look human.
My mother is a halfling and my dad is an ogre, so that makes me an Ogreling
When is AI going to tackle the really hard/useful stuff like a robot to talk to telemarketers?
Now there is a gift to mankind worth making.
Yes, 100% truth!
I've been playing a lot of SC2 lately, and it's a really strange thing with that particular game.
I find that it's really enjoyable, trying different variations of strategies and a certain thrill in pulling off such things as heading off a whole army of units as they're on the way to destroy your base, using only a couple of Ghosts with nukes. Or strategically building a bunch of Void Rays in a far corner of the map that the enemy hasn't even explored yet, so you can suddenly bring a fleet of them out unexpectedly to change the course of the battle, just when the enemy thinks he/she is winning.
But at the same time? I just have no interest in ranking up to play at a high level like "Diamond", because all of those "pros" are just playing the game like machines. It's all about how quickly you can click a mouse in exact sequences to maximize build rates of the "optimal" units at so many seconds into the game, etc. etc. While that makes them technically superior players, it also strips all the fun out of it for me. If I get pitted against anyone who plays like that, I find everything I'm constructing gets completely annihilated in a matter of only a few minutes, with hordes of high level units I didn't even think could be built that quickly.... But I have to think people playing the game that way aren't even enjoying it anymore. They're just enjoying going through the motions as precisely as possible to achieve the reward of as high as rank as they can get. It's sort of like the gambler who keeps pulling the lever on that slot machine because the potential payout motivates them to continue, instead of actually finding the slot machine fun to play in and of itself.
How about beating humans at something like Cards Against Humanity? Or Werewolf?
I'm still looking for the computer that can beat me at boxing.
Wanna buy a shirt?
https://www.redbubble.com/people/stealthfinger/shop?asc=u
In my mind, if they don't get even close of DOTA2 and LOL number, it won't be a success and it won't survive for long.
Even starcraft 1 is still surviving and has a scene, heck, even smash brothers has a scene and has survived, so you're not thinking very clearly.
"First they came for the slanderers and i said nothing."
It sounds like some unfortunate acronym confusion here:
Starcraft II: HotS = Heart of the Swarm
HoTS = Heroes of the Storm
The problem is that the build orders and game are designed around results of pro-gamers facing each other which changing strategy over time. If you try to build an AI that beats those players, and I think you can, youve missed the point. If the AI chooses to use its own strategy, it might not be countered by unprepared pro-players. But if the AI was involved in the ladders/tournaments, just as a new player would be, players would learn to beat it. If THEN, it could win more than not, that would be something.
Even starcraft 1 is still surviving and has a scene, heck, even smash brothers has a scene and has survived, so you're not thinking very clearly.
We're talking business here. So unless you show me some source of major income from it, Starcraft 1 popularity today doesn't worth much except for the IP (which show in the sales of Starcraft II).
And I'm not quite sure where you're going with this, all I'm saying is that, in my mind, the last two AAA game of blizzard (Overwatch and HOTS) won't achieve the success of their other IP (Warcraft, Starcraft and Diablo).
Elok
It sounds like some unfortunate acronym confusion here:
Starcraft II: HotS = Heart of the Swarm
HoTS = Heroes of the Storm
Oh my god, I didn't notice!
One of Starcraft II expansion got the same acronym as Heroes of the Storm.
Elok
And I'm not quite sure where you're going with this, all I'm saying is that, in my mind, the last two AAA game of blizzard (Overwatch and HOTS)
Bro, HOTS is not the latest game from Blizzard.
"First they came for the slanderers and i said nothing."
Dying?
http://www.digitaltrends.com/gaming/hearthstone-earning-20-million-monthly/
And I'm not quite sure where you're going with this, all I'm saying is that, in my mind, the last two AAA game of blizzard (Overwatch and HOTS)
Bro, HOTS is not the latest game from Blizzard.
Yeah, another guy just pointed out that HOTS is the acronym for both Heroes of the Storm and StarCraft II: Heart of the Swarm. I'm talking about Heroes of the Storm.
Elok
oh, I see what you're getting at.
"First they came for the slanderers and i said nothing."
Yes Heroes of the Storm is a flop. Question. How many NEW players will start playing Starcraft and WoW? I said Diablo has a following/player base. Overwatch @$60 will flop. In 6 months Overwatch will be just like Battlefront is now. Dead. Blizzard is the Microsoft of games companies. They are now putting out games no one wants to play except the die hard Blizzard fanboys. You know because it's a Blizzard game. Tell me how Overwatch is $60 and TF2 is still free.
Actually I think all of their game have been a financial success compared to what would normally be called a success in the game industry. It is just that WoW was a ridiculous juggernaut cash cow. You simply can't plan for every game to give that kind of financial success and if revenue drops off from WoW, they will have no choice but to downsize into a more normal, if still successful, sized company. If anything it looks like they are managing the transition better that many companies in the past.
Yes Heroes of the Storm is a flop. Question. How many NEW players will start playing Starcraft and WoW? I said Diablo has a following/player base. Overwatch @$60 will flop. In 6 months Overwatch will be just like Battlefront is now. Dead. Blizzard is the Microsoft of games companies. They are now putting out games no one wants to play except the die hard Blizzard fanboys. You know because it's a Blizzard game. Tell me how Overwatch is $60 and TF2 is still free.
You can't blame Blizzard to try creating new IP. And, worst case, they could create a D3 extension and millions will sell. They got many "easy-cash" option.
Elok
hah hahaha. man, you are hysterical.
You know, I hear we'll never need more than 640k on computers either!
She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.
Uh.. Overwatch is only $40, at least on PC. There's a fancy version with extra skins and stuff for other Blizzard games, which is $60. No idea what console versions will cost, though.
And the StarCraft players, too, shall fall from their high thrones as the machine mind pummels them into oblivion...
TF2 cost money when it first came out as well, then it went free... this is the same thing Blizzard is doing.. paid to see if they can recoup devel costs then if it is successful use it to drive their platform (blizzard.net, which they already have a free game that has driven usage in part)