For now. Give them another month. This is like AlphaGo vs Fan Hui all over again -- people knocked that accomplishment at the time because he was just a master, not one of the top players in the world. Well, not much longer, AlphaGo beat Lee Sedol, the best player in the world.
The ceiling here is going to be incredibly high, much higher than the level of play that people are capable of, even when restricted to a single window.
This doesn't nullify the observations that people are making here.
Part of the difficulty here is describing what a 'fair' match might be. Specifically, I think fairness has to do with a goal many people have for AI: to improve human play. The strategies in Chess or Go that were employed could conceivably be used by human players. There aren't any hard restrictions preventing humans from learning from that play, even if the AI is entirely superior.
It would follow that a 'fair' SCII match would employ strategies that humans could implement. Making extra workers, for instance, might be a real lesson from AlphaStar play. The insane stalker micro, however, could never be done by a human.
From this perspective, I think the important takeaways were:
* The AI leaned heavily on super-human stalker micro.
* The AI had some strategic blind-spots, namely the immortal harass.
* The APM comparison isn't terribly meaningful; a lot of human APM is spammy/twitchy button presses that doesn't do all that much, whereas the AI can presumably make each action count. There were also AlphaStar APM spikes that likely go along with the stalker-micro issue.
None of this really matters though. The AI is improving every day through training. Give it another few months of development and it'll be able to trounce humans under any "fair" set of handicaps you can think of, like limiting average and max APM throughout the game. We saw the same pattern with AlphaGo. There's no reason whatsoever to suppose that humans are fundamentally better at this game than an AI can be.
When AlphaGo first one, people said it wasn't fair because it was running on a whole cluster of computers. Well, within not much time at all, it was good enough to run on a single computer and still beat top humans. We are dealing with exponential progress here. The writing is on the wall.
It's tempting to assume the AI will just keep getting better and better, but that's not guaranteed, and I was happy to see that the Deepmind folks in the video clearly acknowledged this. In the game that MaNa won, it's possible that he did so by finding a strategy the AI agent had never encountered before, causing it to respond with nonsense (e.g. not building a Phoenix and pulling its entire stalker army back to deal with warp prism harrassment). In a game with a strategy space as large as SC2, it's possible that an AI will never be able to saturate the space of viable strategies, and it will always be possible to find edge cases that the AI has no idea how to handle.
The point isn't that the AI won't improve or win with those conditions; I agree it likely will, and soon. The point is that the conditions of the match matter and that this one missed the mark.
It absolutely does matter whether the AI can use obviously super-human techniques, because then it's not nearly as interesting for human observers. I'd much rather watch an AI that was a strategic genius that won despite being hamstrung in terms of micro/techniques.
> There's no reason whatsoever to suppose that humans are fundamentally better at this game than an AI can be.
Lee Sedol was not the best player anymore at that time (not saying it wasn't an impressive/important achievement, but overstating it doesn't help either - the "beat best human players part" came later in 2017).
Lee Sedol was still top 5, certainly no worse than top 10 at the time. By all mean he wasn't the best and most dominant, but the difference with the top was tiny.
I don't understand who's downvoting you, this is accurate. While AlphaGo/Zero improved quickly to superhuman play, we are just in this thread comparing timelines, so that is relevant.
What kind of evidence is going into this analogical reasoning? Do we also extrapolate similarly for other things? We went to the Moon in 1960s. Was Mars a month, or a year, or a decade away? Then we sent robots to Mars. Did we yet send any robots to Alpha Centauri?
Different problems have different difficulties. Solving simple problems quickly doesn't mean we'd also be able to just as easily solve the hard problems. Often the comparably simpler problems have the best reward/effort ratio and thus make quick progress, which doesn't need to be the case for hard problems.
Going to the Moon is a completely different endeavor than making an AI better at a game that it's already quite good at. This is a red herring.
If you had bet against AIs reaching parity with top human players in any previous game, whether it be Checkers, Chess, Go, etc., you'd have lost. I see no reason why StarCraft II should be any different.
We can reconvene in the comments here a year from now and see where AlphaStar is then.
It doesn't seem like hype to me -- it seems like a genuine, significant accomplishment. Sure, they might not be able to beat the best pro players consistently right now, but I suspect that is right around the corner. Would you rather they stay completely mum until they've reached that goal too? And why? I'd rather know now, and then be able to follow along as it gets better and beats higher and higher-ranked players.
This was my read as well. It seems that Mana simply found a strategy that the AI had not found. Due to not having trained against it, the AI produced nonsense results. The commentators noted that the obvious response was to build a Pheonix and just completely shut down the harassment. The situation is similar to Alpha Go vs Lee Sedol match 4.
One of the hardest parts about these kinds of human vs ai expositions is making sure the AI has explored the full possibility space, so that can handle all situations. The techniques at play lack the ability to perceive a completely new situation and formulate a good response. (Though anyone who's lost to cheese in games they later learned easy counters for know that humans, while better than state of the art AI, aren't perfect here either.)
Mana got himself in the same situation where he was surrounded by stalkers on multiple sides, but this time the micro wasn’t so crazy that he couldn’t manage it, and he was able to take on one group at a time.
The immortal drop, while unanswered, was not really that effectual.
But it was answered: AlphaStar pulled a huge stalker army that was about to hit MaNa's base all the way back home to (attempt to) answer the drop, repeatedly. If you have more complexity to your army but fewer army units, as MaNa did, a delay like that is how you win the game.
That's what I said on Lobsters. They were always good at builds, micro's, etc. The one thing they couldn't do was judge human intent, esp if they were being mislead (esp time wasting). I was waiting for one of the players to try to screw with its head to see what it did. Mana showed two gaps: the back and forth thing; that it ignored the observers giving up constant strategy information. Then, he got the first win.
Now, the questions are how many more such glitches will show up and can they eliminate them with better algorithms?
And against human players up to Masters 3 or so :) When you're still using the all-army hotkey, defending with a small and precise group isn't happening.
That time, the ai didn’t really even try to engage. In fact, the ending of the match was marked by the entirely absent group of stalkers as the natural was engaged.
It’s likely safer to say the AI was confused in general at that point, possibly related to the camera change, but we didn’t really get to see the quality of stalker micro that game
"possibly related to the camera change, but we didn’t really get to see the quality of stalker micro that game "
In software, changes in assumptions can break what depended on them. There could be many assumptions in its neural net centered on full visibility. They should probably retrain all or just some from scratch with the camera change in from the beginning to see what happens. Then, it will be firmly encoded into the strategies over time.
The immortal drop let him keep AlphaStar occupied while he built up a critical mass of immortals (it becomes harder and harder to effectively micro stalkers against immortals as numbers go up, probably even for an AI), then let him put AlphaStar in an awkward position when it was camping near where the warp prism was hiding.
The ceiling here is going to be incredibly high, much higher than the level of play that people are capable of, even when restricted to a single window.