Login / Signup
A novel multi-step Q-learning method to improve data efficiency for deep reinforcement learning.
Yinlong Yuan
Zhu Liang Yu
Zhenghui Gu
Yao Yeboah
Wei Wu
Xiaoyan Deng
Jingcong Li
Yuanqing Li
Published in:
Knowl. Based Syst. (2019)
Keyphrases
</>
multi step
reinforcement learning
input data
data sets
missing data
dynamic programming
model free
multi agent
support vector machine svm
single step
function approximation
covariance matrix
optimal policy
training samples
state space
data points
pairwise
objective function
similarity measure
learning algorithm