Transcription of Asynchronous Methods for Deep Reinforcement Learning
{{id}} {{{paragraph}}}
Example: confidence
Asynchronous Methods for Deep Reinforcement Learning
Asynchronous Methods for Deep Reinforcement LearningVolodymyr Puigdom nech P. DeepMind2 Montreal Institute for Learning Algorithms (MILA), University of MontrealAbstractWeproposeaconceptuallysi mpleandlightweight framework for deep reinforce-ment Learning that uses Asynchronous gradientdescent for optimization of deep neural networkcontrollers. We present Asynchronous variants offour standard Reinforcement Learning algorithmsand show that parallel actor-learners have astabilizing effect on training allowing all fourmethods to successfully train neural best performing method, anasynchronous variant of actor-critic, surpassesthe current state-of-the-art on the Atari domainwhile training for half the time on a singlemulti-core CPU instead of a GPU.
In earlier work, (Li & Schuurmans,2011) applied the Map Reduce framework to parallelizing batch reinforce-ment learning methods with linear function approximation. Parallelism was used to speed up large matrix operations but not to parallelize the collection of experience or sta-bilize learning. (Grounds & Kudenko,2008) proposed a
Loading..
Tags:
Information
Domain:
Source:
Link to this page:
Please notify us if you found a problem with this document: