I just started working on the openai retro contest, and was wondering who is out there currently competing?
here is the site: https://contest.openai.com/details
I have been trying simple agents that just keep going right, and JERK, and was about to utilize Deep learning. If anyone also is thinking of utilizing some of what we learnt in class, please let me know. Would be an awesome thought exercise!
I just started did in terms of downloading code, hasn’t gotten very far, but would love to exchange ideas!
awesome , lets do it 
i submitted a couple of submissions
Sounds interesting, thank for sharing this info.
I wonder what knowledge from class we can utilize here? I’d be happy to work on it.
There are few things that are applicable:
a) the code provided is using tensorflow and the competition is open to using pytorch
b) the implementations are based on published papers, and i learnt a lot from how jeremy dissected papers during the lectures that it might be awesome to try and do the same
c) There are two deep learning baselines implementations offered one us rainbow DQN it might be interesting to explore and make better