1. [Publications](/publications)
2. Reinforcement Learning through Asynchronous Advantage Actor-Critic on a GPU
 
 # Reinforcement Learning through Asynchronous Advantage Actor-Critic on a GPU

  ![](/sites/default/files/styles/wide/public/pubs/2017-04_Reinforcement-Learning-through/GA3C_v1.png?itok=QnweVWgW)

 We introduce a hybrid CPU/GPU version of the Asynchronous Advantage ActorCritic
(A3C) algorithm, currently the state-of-the-art method in reinforcement
learning for various gaming tasks. We analyze its computational traits and concentrate
on aspects critical to leveraging the GPU’s computational power. We
introduce a system of queues and a dynamic scheduling strategy, potentially
helpful for other asynchronous algorithms as well. Our hybrid CPU/GPU version
of A3C, based on TensorFlow, achieves a significant speed up compared
to a CPU implementation; we make it publicly available to other researchers at
<https://github.com/NVlabs/GA3C>.



 ## Authors



[Iuri Frosio](/person/iuri-frosio)

[Stephen Tyree](/person/stephen-tyree)

[Jason Clemons](/person/jason-clemons)

[Jan Kautz](/person/jan-kautz)

Mohammad Babaeizadeh (University of Illinois at Urbana-Champaign)

 

 

 ## Publication Date



Saturday, April 1, 2017

 

 ## Published in



[Proceeding of ICLR 2017](https://openreview.net/pdf?id=r1VGvBcxl)

 

 ## Research Area



[Artificial Intelligence and Machine Learning ](/research-area/machine-learning-artificial-intelligence)

 

 

 ## External Links



[Reinforcement Learning through Asynchronous Advantage Actor-Critic on a GPU](https://openreview.net/pdf?id=r1VGvBcxl)