Deep reinforcement learning overfits. Often, a trained network can only play the particular level(s) you trained it on! In our new paper, we show how to train more general networks with procedural level generation, generating progressively harder levels. https://arxiv.org/abs/1806.10729 pic.twitter.com/JDNYPZuAHV
-
Show this thread
-
-
Replying to @hardmaru
Bonjour you can read it here: Thread by
@togelius: "Deep reinforcement learning overfits. Often, a trained network can only play the particular level(s) you trained it on! […]" https://threadreaderapp.com/thread/1012726654261702658.html … Have a good day.
1 reply 4 retweets 7 likes
But can it play untrusted? Generate new interesting levels for it? https://alexnisnevich.github.io/untrusted/
10:50 AM - 30 Jun 2018
0 replies
0 retweets
1 like
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.
Read my blog!