From 1e6e9640f0b7d30902bc12479e58faaa51b965ff Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Taneli=20Mielik=C3=A4inen?= Date: Mon, 6 Jan 2020 15:20:40 -0800 Subject: [PATCH] Nerual -> Neural --- pkg/suggestion/v1alpha3/NAS_Reinforcement_Learning/README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/pkg/suggestion/v1alpha3/NAS_Reinforcement_Learning/README.md b/pkg/suggestion/v1alpha3/NAS_Reinforcement_Learning/README.md index c06a4ff592d..e2df242d335 100644 --- a/pkg/suggestion/v1alpha3/NAS_Reinforcement_Learning/README.md +++ b/pkg/suggestion/v1alpha3/NAS_Reinforcement_Learning/README.md @@ -1,4 +1,4 @@ -# About the Nerual Architecture Search with Reinforcement Learning Suggestion +# About the Neural Architecture Search with Reinforcement Learning Suggestion The algorithm follows the idea proposed in *Neural Architecture Search with Reinforcement Learning* by Zoph & Le (https://arxiv.org/abs/1611.01578), and the implementation is based on the github of *Efficient Neural Architecture Search via Parameter Sharing* (https://github.com/melodyguan/enas). It uses a recurrent neural network with LSTM cells as controller to generate neural architecture candidates. And this controller network is updated by policy gradients. However, it currently does not support parameter sharing.