4
$\begingroup$

I'm looking for academic papers or other credible sources focusing on the topic of parralelized reinforcement learning, specifically Q-learning. I'm mostly interested in methods of sharing Q-table between processes (or joining/syncing them together if each process have it's own). I'd also appreciate a brief description of method used in linked/mentioned sources.

I should mention that I use neural network (PyBrain) as approximation.

|improve this question
$\endgroup$
  • $\begingroup$ I couldn't add pybrain as tag (too little rep.) $\endgroup$ – Luke Jan 15 '16 at 7:44
  • 1
    $\begingroup$ Done it for you :) $\endgroup$ – Dawny33♦ Jan 15 '16 at 10:47
2
$\begingroup$

I think you will like the following two papers:

Available from: http://arxiv.org/abs/1507.04296

Nair A, Srinivasan P, Blackwell S, Alcicek C, Fearon R, De Maria A, et al. Massively Parallel Methods for Deep Reinforcement Learning. arXiv preprint arXiv:150704296

Available from: http://arxiv.org/abs/1602.01783

Mnih V, Badia AP, Mirza M, Graves A, Lillicrap TP, Harley T, et al. Asynchronous Methods for Deep Reinforcement Learning. arXiv:160201783

|improve this answer
$\endgroup$
  • $\begingroup$ I use simple Q-learning but I guess articles about deep Q-learning would also be useful, not for my current work however. $\endgroup$ – Luke Feb 23 '16 at 17:42
  • $\begingroup$ In my opinion, although the papers are about deep q-learning the concepts are applicable to what you are doing: different agents collecting trajectories in parallel and sharing and exchanging the network parameters. Definitely can serve as inspiration. $\endgroup$ – Juan Leni Feb 23 '16 at 18:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.