Our hypothesis, which our experiments suggest, is that learning to play this game will cause useful representations of multi-agent dynamics emerge. We train the ...
確定! 回上一頁