Deep Reinforcement Learning A brief survey

被引:2735
作者
Arulkumaran, Kai [1 ]
Deisenroth, Marc Peter [2 ,3 ]
Brundage, Miles [4 ,5 ]
Bharath, Anil Anthony [1 ]
机构
[1] Imperial Coll London, Dept Bioengn, London, England
[2] Imperial Coll London, Dept Comp, London, England
[3] PROWLER Io, Cambridge, England
[4] Arizona State Univ, Sci & Technol Dept, Human & Social Dimens, Tempe, AZ 85287 USA
[5] Univ Oxford, Future Humanity Inst, Oxford, England
关键词
NETWORKS;
D O I
10.1109/MSP.2017.2743240
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
Deep reinforcement learning (DRL) is poised to revolutionize the field of artificial intelligence (AI) and represents a step toward building autonomous systems with a higher-level understanding of the visual world. Currently, deep learning is enabling reinforcement learning (RL) to scale to problems that were previously intractable, such as learning to play video games directly from pixels. DRL algorithms are also applied to robotics, allowing control policies for robots to be learned directly from camera inputs in the real world. In this survey, we begin with an introduction to the general field of RL, then progress to the main streams of value-based and policy-based methods. Our survey will cover central algorithms in deep RL, including the deep Q-network (DQN), trust region policy optimization (TRPO), and asynchronous advantage actor critic. In parallel, we highlight the unique advantages of deep neural networks, focusing on visual understanding via RL. To conclude, we describe several current areas of research within the field.
引用
收藏
页码:26 / 38
页数:13
相关论文
共 100 条
[1]  
[Anonymous], P INT C LEARN REPR
[2]  
[Anonymous], 2017, P INT C LEARN REPR
[3]  
[Anonymous], NIPS WORKSH DEEP REI
[4]  
[Anonymous], 2016, P INT C LEARN REPR
[5]  
[Anonymous], BRIDGING GAP VALUE P
[6]  
[Anonymous], P INT C LEARN REPR
[7]  
[Anonymous], P INT C LEARN REPR
[8]  
[Anonymous], ICML WORKSH DEEP LEA
[9]  
[Anonymous], NIPS WORKSH DEEP REI
[10]  
[Anonymous], FROM ANIMALS TO ANIMATS