5 ms·
I'm not proficient in Starcraft, but if they have to make Pros play against 5 different agents for a Bo5 is it because the same agent would merely repeat the sa
by devilmoon 8y ago
I'm not proficient in Starcraft, but if they have to make Pros play against 5 different agents for a Bo5 is it because the same agent would merely repeat the same game overall and the human player would be able to see through its strategy?
AFAIK at least for Go/Chess DeepMind wasn't handpicking agents to send against human opponents, but it was simply a trained agent who would try its own strategy and respond to the opponent, in this case isn't it like the single agent only ever plays a single type of meta strat?
If so, I think this is a bit less impressive than what I had predicted.
- aurelwu 8y agoDifferent agents for now, and 1 of the 6 possible matchups and just on one map (starcraft usually is played on multiple different maps/boards with different layouts over the course of a series). Given that this was just a demonstration match it's reasonable restrictions and they surely will lift them one by one (just like they removed the zoomed out vision in the final, live-played match). I'm really impressed with their performance, and even if my brother was just playing offrace I didn't expect him to lose.
- codeflo 8y agoThere's a rock-paper-scissors like aspect to StarCraft openings where you essentially have to randomize your strategy so that the other player can't blindly counter what you're doing. I assume each of their agents learns one particular strategy; in that case, they could simply create a "meta-agent" that acts like one of their trained agents at random (probably using some weighted distribution).
- devilmoon 8y agoWell, why wasn't this implemented then? For an hyped and live streamed event I'm expecting to see something that blows me away, not hand picked agents still in embryonic stage