Very nice paper on Cartesian Genetic Programming applied to Atari games. https://arxiv.org/abs/1806.05695 -- State-of-the-art on a number of games, and very competitive on many more.
Show this thread
This is remarkable because getting strong results on ALE using deep RL took many refinement iterations, whereas this is a "first try" of sorts. Usually, if you can get competitive results on a historically hard problem with relatively little effort, you are onto something.
-
-
Yes, like its very easy to game the objective function. "These simple strategies create local optima which can deceive evolution."
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Although I understand the basic principle of evolving a program that takes an image as input, I have to say I have a hard time understanding the specific detailed in the paper :(pic.twitter.com/XwbhOZoM9D
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
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.