2 ms·
Surely all this RL process did was speed up what the computer would have learned, by a large stretch? The "cost" factors they chose would have hit them ultimat
by Twirrim 9y ago
Surely all this RL process did was speed up what the computer would have learned, by a large stretch? The "cost" factors they chose would have hit them ultimately, regardless of if the bot stood still in the base, or wandered off elsewhere.
- aoeuasdf1 9y agoNot necessarily - maybe a mixed strategy of (1) Not creepblocking at all, and letting your opponent have your creeps under their tower (2) Modest creepblocking to punish (1) (3) Severe creepblocking to punish (2) Would be better than some hand-trained RL creepblocking which is divorced from game outcomes.
- mannykannot 9y agoMy guess is that in general, many complex strategies are effectively unreachable without something like analysis, on account of intermediate states being disfavored, leading to algorithms being trapped by local minima in the cost function (I don't know whether that would be an issue for this game, specifically.)