Re:Reinforcement learning’s foundational flaw
by dgant on Monday July 09, 2018 @05:51PM
(#17493741)
Enjoyed the article. But I'm a bit confused by the Venn diagram; neither StarCraft nor DOTA are deterministic or fully observable. And they are discrete only at such extreme resolutions that they may as well be continuous.
[ Parent
| on HN ]
Re:Reinforcement learning’s foundational flaw
by batmansmk on Monday July 09, 2018 @06:30PM
(#17493931)
Sc2 is deterministic. A replay file is simply a record of all actions performed. There is no randomness in the action resolutions, which allows actions to be the only thing needed to be streamed over to dync game states.
Re:Reinforcement learning’s foundational flaw
by alcinos on Tuesday July 10, 2018 @06:06AM
(#17496746)
Well it's deterministic once you know the random seed, which is stored in the replay file. An agent doesn't know the seed, hence cannot predict the exact outcome of its actions, only a probability over the outcomes. So, from the agent's perspective, the game is indeed random.
Re:Reinforcement learning’s foundational flaw
by andreyk on Monday July 09, 2018 @07:52PM
(#17494368)
True, not all properties are there to the same extent as in eg Go, but they are there to some extent and more broadly Starcraft and Dota still 'game-like' ie not as complex as driving a self driving car in the real world. The Venn diagram is tricky to get right...