As an important method to solve sequential decision-making problems, reinforcement learning learns the policy of tasks through the interaction with environment. But it has difficulties scaling to large-scale problems. One of the reasons is the exploration and exploitation dilemma which may lead to inefficient learning. We present an approach that addresses this shortcoming by introducing qualitative knowledge into reinforcement learning using cloud control systems to represent ‘if-then’ rules. We use it as the heuristics exploration strategy to guide the action selection in deep reinforcement learning. Empirical evaluation results show that our approach can make significant improvement in the learning process.
Chenxi LI
PLA University of Science and Technology
Lei CAO
PLA University of Science and Technology
Xiaoming LIU
PLA University of Science and Technology
Xiliang CHEN
PLA University of Science and Technology
Zhixiong XU
PLA University of Science and Technology
Yongliang ZHANG
PLA University of Science and Technology
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Chenxi LI, Lei CAO, Xiaoming LIU, Xiliang CHEN, Zhixiong XU, Yongliang ZHANG, "A Study of Qualitative Knowledge-Based Exploration for Continuous Deep Reinforcement Learning" in IEICE TRANSACTIONS on Information,
vol. E100-D, no. 11, pp. 2721-2724, November 2017, doi: 10.1587/transinf.2017EDL8112.
Abstract: As an important method to solve sequential decision-making problems, reinforcement learning learns the policy of tasks through the interaction with environment. But it has difficulties scaling to large-scale problems. One of the reasons is the exploration and exploitation dilemma which may lead to inefficient learning. We present an approach that addresses this shortcoming by introducing qualitative knowledge into reinforcement learning using cloud control systems to represent ‘if-then’ rules. We use it as the heuristics exploration strategy to guide the action selection in deep reinforcement learning. Empirical evaluation results show that our approach can make significant improvement in the learning process.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2017EDL8112/_p
Copy
@ARTICLE{e100-d_11_2721,
author={Chenxi LI, Lei CAO, Xiaoming LIU, Xiliang CHEN, Zhixiong XU, Yongliang ZHANG, },
journal={IEICE TRANSACTIONS on Information},
title={A Study of Qualitative Knowledge-Based Exploration for Continuous Deep Reinforcement Learning},
year={2017},
volume={E100-D},
number={11},
pages={2721-2724},
abstract={As an important method to solve sequential decision-making problems, reinforcement learning learns the policy of tasks through the interaction with environment. But it has difficulties scaling to large-scale problems. One of the reasons is the exploration and exploitation dilemma which may lead to inefficient learning. We present an approach that addresses this shortcoming by introducing qualitative knowledge into reinforcement learning using cloud control systems to represent ‘if-then’ rules. We use it as the heuristics exploration strategy to guide the action selection in deep reinforcement learning. Empirical evaluation results show that our approach can make significant improvement in the learning process.},
keywords={},
doi={10.1587/transinf.2017EDL8112},
ISSN={1745-1361},
month={November},}
Copy
TY - JOUR
TI - A Study of Qualitative Knowledge-Based Exploration for Continuous Deep Reinforcement Learning
T2 - IEICE TRANSACTIONS on Information
SP - 2721
EP - 2724
AU - Chenxi LI
AU - Lei CAO
AU - Xiaoming LIU
AU - Xiliang CHEN
AU - Zhixiong XU
AU - Yongliang ZHANG
PY - 2017
DO - 10.1587/transinf.2017EDL8112
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E100-D
IS - 11
JA - IEICE TRANSACTIONS on Information
Y1 - November 2017
AB - As an important method to solve sequential decision-making problems, reinforcement learning learns the policy of tasks through the interaction with environment. But it has difficulties scaling to large-scale problems. One of the reasons is the exploration and exploitation dilemma which may lead to inefficient learning. We present an approach that addresses this shortcoming by introducing qualitative knowledge into reinforcement learning using cloud control systems to represent ‘if-then’ rules. We use it as the heuristics exploration strategy to guide the action selection in deep reinforcement learning. Empirical evaluation results show that our approach can make significant improvement in the learning process.
ER -