EPTIt definitely seems like it would definitely be possible for an AI to beat the best players most of the time, but Starcraft seems so much more volatile than Go or Chess, especially with the possible denial/hiding of information, or straight up tricking the AI (researching cloak, then cancelling it, cancelling tech after it's scouted, etc). I think that the best/sneakiest/craftiest players would be able to outsmart it once in a while. I doubt that the best players would NEVER win, you know?
BoxeR: "AlphaGo won't beat humans in StarCraft" - Page 8
| Forum Index > SC2 General |
|
Ryncol
United States980 Posts
It definitely seems like it would definitely be possible for an AI to beat the best players most of the time, but Starcraft seems so much more volatile than Go or Chess, especially with the possible denial/hiding of information, or straight up tricking the AI (researching cloak, then cancelling it, cancelling tech after it's scouted, etc). I think that the best/sneakiest/craftiest players would be able to outsmart it once in a while. I doubt that the best players would NEVER win, you know? | ||
|
SK.Testie
Canada11084 Posts
That was pretty serious stuff. | ||
|
BronzeKnee
United States5217 Posts
On March 13 2016 03:19 Brutaxilos wrote: As a programmer, I'm actually quite skeptical that Boxer would be able to beat an intelligent AI given that a team of scientists are given enough time to develop one. It's inevitable. If the constraints are the same, as in the computer must use a keyboard and mouse and can only view one part of the screen at a time I don't think an AI could ever win against a top player, at least in my lifetime. And if we were able to control the game using the human brain, it would be another no contest, we'd have near perfect micro and macro too. I think many of you are underestimating the power of the brain. Only if the computer is unrestrained by a keyboard and mouse and the human is restrained by those factors will the human lose. Humans are far too innovative. If you simply deny scouting and the AI will either guess what you are doing or go for some standard safe play, and either of those things could be exploit. | ||
|
writer22816
United States5775 Posts
On March 13 2016 13:57 BronzeKnee wrote: If the constraints are the same, as in the computer must use a keyboard and mouse and can only view one part of the screen at a time I don't think an AI could ever win against a top player, at least in my lifetime. And if we were able to control the game using the human brain, it would be another no contest, we'd have near perfect micro and macro too. Only if the computer is unrestrained by a keyboard and mouse and the human is restrained by those factors will the human lose. Humans are far too innovative. If you simply deny scouting and the AI will either guess what you are doing or go for some standard safe play, and either of those things could be exploit. What the hell does it even mean for a computer to use a keyboard and a mouse? A computer doesn't have two hands and 10 fingers. | ||
|
BronzeKnee
United States5217 Posts
On March 13 2016 14:03 writer22816 wrote: What the hell does it even mean for a computer to use a keyboard and a mouse? A computer doesn't have two hands and 10 fingers. The computer with the AI should be playing the game using another computer using a keyboard, mouse and monitor because that is how Starcraft is played. I'm didn't create Starcraft so don't blame that that is how the game is played. Having those limitations is what makes Starcraft difficult. If the AI can at once be blink microing Stalkers while warping in units at pylon off screen (off field of vision) then that is cheating. The AI in Chess or Go can do nothing a human cannot, the AI is literally outthinking the players. So even if the APM is limited, the field of vision must be limited also. The AI has no chance given equal constraints. If we are talking about no keyboard, mouse or monitor for the computer, then it should be the same for humans. I can imagine perfect forcefields, if the game responded to my mind, I'd never miss a forcefield. And my macro would be on point too, just subtle sounds would be all I would need to know to send a worker to a mineral line. | ||
|
ZAiNs
United Kingdom6525 Posts
On March 13 2016 14:03 writer22816 wrote: What the hell does it even mean for a computer to use a keyboard and a mouse? A computer doesn't have two hands and 10 fingers. Basically that the game would have to interact with the game in the same way humans do. It would have to give the game inputs via virtual keyboard and mouse (which would have some limitations on speed and accuracy to make this fair) which would also mean it'd have to reason about things like what it uses its hotkeys for. It'd also have to observe the game through the same viewport a player sees and have access to the minimap, as opposed to just having the whole game state available to it constantly as the in-game bots do. | ||
Jibba
United States22883 Posts
On March 13 2016 10:57 Liquid`Nazgul wrote: Totally agree. It doesn't make any sense to beat top pros with limited strategy and perfect execution and then brag about it from an AI perspective. A waste of time in terms of prestige. It would be kind of awesome to see an AI do a probe rush in BW and win with perfect micro. At least it'd be awesome the first time, and it wouldn't be awesome because of the AI but because it's such an infuriating, unsolvable way to lose. | ||
|
todespolka
221 Posts
No human is able to multi select single units for example and if ai has an advantage its not fair. | ||
|
todespolka
221 Posts
On March 13 2016 14:06 BronzeKnee wrote: The computer with the AI should be playing the game using another computer using a keyboard, mouse and monitor because that is how Starcraft is played. I'm didn't create Starcraft so don't blame that that is how the game is played. Having those limitations is what makes Starcraft difficult. If the AI can at once be blink microing Stalkers while warping in units at pylon off screen (off field of vision) then that is cheating. The AI in Chess or Go can do nothing a human cannot, the AI is literally outthinking the players. So even if the APM is limited, the field of vision must be limited also. The AI has no chance given equal constraints. If we are talking about no keyboard, mouse or monitor for the computer, then it should be the same for humans. I can imagine perfect forcefields, if the game responded to my mind, I'd never miss a forcefield. And my macro would be on point too, just subtle sounds would be all I would need to know to send a worker to a mineral line. Exactly! | ||
Liquid`Jinro
Sweden33719 Posts
On March 13 2016 13:33 Ryncol wrote: I'm not very well-educated on the subject of AI or Go, but assuming an APM cap is in place to make the game mechanically possible, it seems like a pretty tall order to make an AI that won't drop a game vs a human player. Like, take Has vs Jaedong with the seven pylon wall off. I'm skeptical of even a God AI would reacting perfectly and wining that, because it's so off the walls fucking insane. It definitely seems like it would definitely be possible for an AI to beat the best players most of the time, but Starcraft seems so much more volatile than Go or Chess, especially with the possible denial/hiding of information, or straight up tricking the AI (researching cloak, then cancelling it, cancelling tech after it's scouted, etc). I think that the best/sneakiest/craftiest players would be able to outsmart it once in a while. I doubt that the best players would NEVER win, you know? I don't think you would necessarily have the same goal posts for a game like Starcraft as you would a game of Go. I.E the best heads-up limit poker bots are (as of a few years) better than the best humans, but you don't expect them to win every hand. Variance is different between games i.e a low level pro will never beat a high ranked pro at Go or chess, seldomly in Starcraft (especially if it's starcraft 2 I think), but quite often in Poker. Doesn't mean they won't make an AI that 5-0s the best human of course, just that it would require a higher skill gap to ensure such an outcome. | ||
|
Yferi
United States90 Posts
On March 13 2016 12:41 EngrishTeacher wrote: Could someone link me to some past AI vs. Pro games? I'd be really interested in seeing how the AIs are currently losing. https://www.youtube.com/playlist?list=PLokMj1YGn8mgnXUgtFk-WmY2dy73Bdhz0 Human >>> AI as of now. There are some other games on YouTube, but those are vs. players who aren't as good. | ||
|
iFU.pauline
France1564 Posts
What's the point having 5000 apm, to tell an unit to attack 5000 times until the next round? At that speed an unit wouldn't even respond anyway coze all you need is one click until the next round. And anyway what an AI could do during that little space of time? It won't produce units faster either... Time resolution is nowhere near that high for a computer to make a significant difference compare to progamer... That doesn't make any sense. Starcraft has its mechanic limits as well and AI has to be bound by it too... Pretty sure we will be sending a human on Mars before an android could beat a human at brood war... | ||
|
Grumbels
Netherlands7031 Posts
On March 13 2016 08:50 Destructicon wrote: I think way too many of you are just going into this thinking of perfect mechanics and mid to late game engagements where the AI just destroys humans. I think that's a pretty narrow view of how things would unfold. The AI would need to learn such intricacies as scouting and interpreting the information it sees, because if the AI gets bunker rushed 3 times in a row all its perfect micro and mechanics will be useless. Yes its true BW and SC2 are very heavily mechanics dependent and a human would probably get destroyed if he'd try to fight an AI toe to toe in any late game situation. However the early to mid game humans can probably juggle a lot of tasks efficiently enough to the point the advantage of an AI would be negligible and then the game sense would kick in. How does the AI learn the subtle differences between a economic 1/1/1 or a offensive one? Or the difference between the different variations of Gateway all-ins (with or without blink). Yeah probably in 5-10 years the programmers will crack it. But I think they'll have one hell of a fight ahead of them when tackling BW and SC2, the information acquisition, interpretation and decisions modules will probably take tons of time to fine tune and refine. It will build an opening database though, if the information it scouts matches with something it has seen before it will know the correct response. And with regards to trickery, I do not think you can easily fool a robust AI with something like that. You would just have to add more openings to the database, like "masked mech into bio allin" or so. I think the opening database will always be small enough to manage, even if you add these variations per opening and even if you allow for masked openings. | ||
|
TheDougler
Canada8304 Posts
On March 13 2016 03:08 Clonester wrote: The same has the complete Go community said about AlphaGo and also Lee Sedol said, he will win so easy against AlphaGo. It became a train wreck... for the Go community. Same will happen with AlphaStarcraft for Boxer, Flash, Bisu and the complete Community. Looks the majority here don't think the Starcraft pros have a chance actually. It's mostly the pros themselves that are confident. The rest of us just want to see the games. | ||
|
Charoisaur
Germany15961 Posts
On March 13 2016 17:37 TheDougler wrote: Looks the majority here don't think the Starcraft pros have a chance actually. It's mostly the pros themselves that are confident. The rest of us just want to see the games. If it would happen in the near future I'm confident the pros would win. However it's inevitable that one day the AI will be able to defeat humans. The question is just how long it'd take | ||
|
rabidch
United States20289 Posts
| ||
hexhaven
Finland931 Posts
http://waitbutwhy.com/2015/01/artificial-intelligence-revolution-1.html | ||
|
Spazzer
Canada139 Posts
On March 13 2016 02:38 Axieoqu wrote: I would assume Starcraft would be even easier for the AI because mechanics are so important. Just consider how well the simple blink/micro bots work. I think this point is the biggest issues. Here is an example of what it could accomplish and some. | ||
|
rednusa
651 Posts
| ||
|
Cuce
Turkey1127 Posts
who are we to question his decree | ||
| ||
