I *think* @halhod was doing some A+ nudge-wink trolling of @GaryMarcus with the CNN comment.
But yes, NNs are not at all suited to many-step-until-reward puzzles like this without very strong hints (like carefully engineered reward functions or watching lots of successful runs).
-
did you read the last paragraph? “In summary, combining MCTS with neural networks is a powerful technique that helps to bridge the gap between symbolic AI and connectionism.“ it’s (also) a hybrid model.
you are getting warmer, yes. it was certainly a seriously misleading piece for DRL, which is what they mostly work on. since they have become a private company, they stand to make money from that misrepresentation. isn’t that what hype is about?
-
-
I mean, hal, robots can totally solve rubiks cubes. This is not a big open problem Here's a robot built with legos! Legos, hal. And a cell phone! And it's infinitely more capable than the openai "demo."https://www.youtube.com/watch?v=X0pFZG7j5cE …
- 18 more replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.