News for hackers. Stuff that matters.
Re:Deep Mind Playing Montezuma's Revenge with Intrinsic Moti... by schoen on Wednesday June 08, 2016 @01:34PM (#11863966)
I'm confused by the subtitle of the linked article, which says "The AI system was able to solve the complex game in just four tries". But the video shows the AI dying many more times than that, and not ultimately winning the game, just learning to explore a portion of it successfully.

What did the Wired editor mean by "solve" and "four tries"? (Or, for that matter, "complete"?)

Re:Deep Mind Playing Montezuma's Revenge with Intrinsic Moti... by wrsh07 on Wednesday June 08, 2016 @03:11PM (#11864769)
Oh, I think I get it now. I think they mean that in one play [ie 4 lives], it solves significantly more than any previous network.
Re:Deep Mind Playing Montezuma's Revenge with Intrinsic Moti... by dang on Wednesday June 08, 2016 @01:53PM (#11864120)
Given that this is what the paper (linked by jerf) says:

within a fraction of the training time, our agent explores a significant portion of the first level and obtains significantly higher scores than previously published agents

... it's probably just the usual hype.