Deep Reinforcement Learning: A Brief Survey
Open accessThe DOI matches its Crossref record; 177 of the 182 references it cites were found in a registry or an index.
Abstract
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.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Foundational surveys on RL [7,72] provide broad overviews of classical and deep RL, respectively, but do not focus on human feedback or LLMs.”
From RLHF Deciphered: A Critical Analysis of Reinforcement Learning from Human Feedback for LLMs (2026) · cites this paper as [7]
In the citing paper
“Classical approaches employ reinforcement learning (RL) [6, 7] and imitation learning [8-10] that operate on primitive actions, which could be challenging for systematic exploration [11-15], interpretability [16-18], and generalization [19-21].”
From Voyager: An Open-Ended Embodied Agent with Large Language Models (2023) · cites this paper as [7]
In the citing paper
“This can be frequently modeled as decision making and learning in a Markov decision process (MDP) [160].”
From Edge artificial intelligence for 6g: vision, enabling technologies, and applications (2022) · cites this paper as [160]
Citations
Cites
LeCun, Y., Bengio, Y., & Hinton, G. (2015) · Nature · doi:10.1038/nature14539 · Cited by 84,126
Learning representations by back-propagating errors
Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1986) · Nature · doi:10.1038/323533a0 · Cited by 31,892 · cited 2×
“The workhorse of DRL, however, remains backpropagation [162, 111].”
Human-level control through deep reinforcement learning
Mnih, V., Kavukcuoglu, K., Silver, D., Rusu, A. A., Veness, J., Bellemare, M. G., Graves, A., Riedmiller, M., Fidjeland, A. K., Ostrovski, G., et al. (2015) · Nature · doi:10.1038/nature14236 · Cited by 30,357 · cited 10×
“[34] train locally linear models for use with the NAF algorithm— the continuous equivalent of the DQN [71]—to improve the algorithm’s sample complexity in the robotic domain where”
InfluentialReinforcement Learning: An Introduction
Sutton, R. S., & Barto, A. G. (1998) · IEEE Transactions on Neural Networks · doi:10.1109/tnn.1998.712192 · Cited by 27,225 · cited 7×
“(temporal) credit assignment problem [115].”
InfluentialReinforcement Learning: An Introduction
(2005) · IEEE Transactions on Neural Networks · doi:10.1109/tnn.2004.842673 · Cited by 24,496
Auto-Encoding Variational Bayes
Kingma, D. P., & Welling, M. (2013) · arXiv · doi:10.48550/arxiv.1312.6114 · Cited by 20,067 · cited 1×
“[22] devised a method for calculating gradients to optimize stochastic policies by “reparameterizing” [30], [60] the stochasticity away from the network, thereby allowing standard gradients to be used (instead of the high-variance REINFORCE estimator [97]).”
Labaca-Castro, R. (2023) · Machine Learning under Malware Attack · doi:10.1007/978-3-658-40442-0_9 · Cited by 19,272
GAN(Generative Adversarial Nets)
(2017) · Journal of Japan Society for Fuzzy Theory and Intelligent Informatics · doi:10.3156/jsoft.29.5_177_2 · Cited by 19,270
Mastering the game of Go with deep neural networks and tree search
Silver, D., Huang, A., Maddison, C. J., Guez, A., Sifre, L., van den Driessche, G., Schrittwieser, J., Antonoglou, I., Panneershelvam, V., Lanctot, M., et al. (2016) · Nature · doi:10.1038/nature16961 · Cited by 15,612 · cited 7×
“dynamic programming over all possible actions (Figure 3, top left), sample trajectories for heuristic search (as was done by AlphaGo [108]), or even perform an exhaustive search (Figure 3, top right).”
InfluentialDistilling the Knowledge in a Neural Network
Hinton, G., Vinyals, O., & Dean, J. (2015) · arXiv · doi:10.48550/arxiv.1503.02531 · Cited by 14,204 · cited 1×
“technique known as distillation [40] has been utilised to both transfer policies learned by large DQNs to smaller DQNs, and transfer policies learned by several DQNs trained on separate games to one single DQN [85, 95].”
Cited by
Items 501–600 of 1,021
Privacy Leakage on DNNs: A Survey of Model Inversion Attacks and Defenses
Fang, H., Qiu, Y., Yu, H., Yu, W., Kong, J., Chong, B., Chen, B., Wang, X., Xia, S., & Xu, K. (2026) · International Journal of Computer Vision · doi:10.1007/s11263-026-03049-x · Cited by 77
Guo, Y., Yu, F. R., An, J., Yang, K., Yu, C., & Leung, V. C. M. (2020) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2020.2968498 · Cited by 76
“…each time stage t , the agent OE observes the current network state s ( t ) ( s ( t ) ∈ S ) and feeds the network state s ( t ) to the deep neural network for computing an action a ( t ) ( a ( t ) ∈ A ), i.e., the video transcoding and video quality adaptation decision for the next time stage [28].”
BackgroundNetwork Intrusion Detection Based on Extended RBF Neural Network With Offline Reinforcement Learning
López‐Martín, M., Sánchez-Esguevillas, A., Arribas, J. I., & Carro, B. (2021) · IEEE Access · doi:10.1109/access.2021.3127689 · Cited by 79
Deep reinforcement learning guided graph neural networks for brain network analysis
Zhao, X., Wu, J. X., Peng, H., Beheshti, A., Monaghan, J. J. M., McAlpine, D., Hernández-Pérez, H., Dras, M., Dai, Q., Li, Y., et al. (2022) · Neural Networks · doi:10.1016/j.neunet.2022.06.035 · Cited by 81
Exploring the multifunctional roles of quantum dots for unlocking the future of biology and medicine
Ali, M. K., Javaid, S., Afzal, H., Zafar, I., Fayyaz, K., Ain, Q. U., Rather, M. A., Hossain, M. J., Rashid, S., Khan, K. A., et al. (2023) · Environmental Research · doi:10.1016/j.envres.2023.116290 · Cited by 78
Li, H., Collins, C. R., Ribelli, T. G., Matyjaszewski, K., Gordon, G. J., Kowalewski, T., & Yaron, D. J. (2018) · Molecular Systems Design & Engineering · doi:10.1039/c7me00131b · Cited by 68
Deep Reinforcement Learning for User Association and Resource Allocation in Heterogeneous Networks
Zhao, N., Liang, Y., Niyato, D. T., Pei, Y., & Jiang, Y. (2018) · Globecom · doi:10.1109/glocom.2018.8647611 · Cited by 74
A multifaceted vision of the human-AI collaboration: a comprehensive review
Puerta-Beldarrain, M., Gómez-Carmona, O., Sánchez-Corcuera, R., Casado-Mansilla, D., López-de-Ipiña, D., & Chen, L. (2025) · IEEE Access · doi:10.1109/access.2025.3536095 · Cited by 73
“A rapidly growing sub-field within reinforcement learning is Deep Reinforcement Learning, which leverages deep neural networks to learn the policy for RL problems [134], [135], [136], [137].”
Hierarchical Tracking by Reinforcement Learning-Based Searching and Coarse-to-Fine Verifying
Zhong, B., Bai, B., Li, J., Zhang, Y., & Fu, Y. (2018) · IEEE Transactions on Image Processing · doi:10.1109/tip.2018.2885238 · Cited by 76
“(DRL) [47] has gained considerable attention and been suc-”
BackgroundTransfer learning as an enabler of the intelligent digital twin
Maschler, B., Braun, D., Jazdi, N., & Weyrich, M. (2021) · Procedia CIRP · doi:10.1016/j.procir.2021.05.020 · Cited by 73
“Even on simple tasks this can take thousands of trials, costing considerable amounts of time [27].”
BackgroundStrategies for Controlling Microgrid Networks with Energy Storage Systems: A Review
Al-Saadi, M., Al‐Greer, M., & Short, M. (2021) · Energies · doi:10.3390/en14217234 · Cited by 78
DQRA: Deep Quantum Routing Agent for Entanglement Routing in Quantum Networks
Le, L., & Nguyen, T. N. (2022) · IEEE Transactions on Quantum Engineering · doi:10.1109/tqe.2022.3148667 · Cited by 87
Li, S., Wei, C. T., & Wang, Y. (2022) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2022.3148085 · Cited by 84
Energy-Efficient Design for a NOMA Assisted STAR-RIS Network With Deep Reinforcement Learning
Guo, Y., Fang, F., Cai, D., & Ding, Z. (2022) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2022.3224926 · Cited by 77
“DRL is one of artificial intelligence (AI) technology, which can train fully autonomous agents via interacting with the environment and applying specific optimal strategies, improving over time through trial and error [10].”
BackgroundElbaz, K., Zhou, A., & Shen, S. (2023) · Tunnelling and Underground Space Technology · doi:10.1016/j.tust.2023.105104 · Cited by 73
Wu, G., Chen, X., Shen, Y., Xu, Z., Zhang, H., Shen, S., & Yu, S. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3357110 · Cited by 75
“Additionally, it enhances stability and facilitates quicker learning of a more effective strategy [50], [51].”
BackgroundZhong, S., Wang, X., Zhao, J., Li, W., Li, H., Wang, Y., Deng, S., & Zhu, J. (2021) · Applied Energy · doi:10.1016/j.apenergy.2021.116623 · Cited by 72
The Impact of Artificial Intelligence on Sustainable Development in Electronic Markets
Thamik, H., & Wu, J. (2022) · Sustainability · doi:10.3390/su14063568 · Cited by 79
Toward Data-Driven Optimal Control: A Systematic Review of the Landscape
Prag, K., Woolway, M., & Çelik, T. (2022) · IEEE Access · doi:10.1109/access.2022.3160709 · Cited by 84
Deep Learning for Optical Sensor Applications: A Review
Al‐Ashwal, N. H., Soufy, K. A. M. A., Hamza, M. E., & Swillam, M. A. (2023) · Sensors · doi:10.3390/s23146486 · Cited by 78
“Examples of this type include generative adversarial networks (GAN) [79] and deep reinforcement learning (DRL) [80].”
Hadi, M. U., tashi, Q. A., Qureshi, R., Shah, A., muneer, A., Irfan, M., Zafar, A., Shaikh, M. B., Akhtar, N., Wu, J., et al. (2023) · doi:10.36227/techrxiv.23589741.v3 · Cited by 53
Blumensaat, F., Leitão, J. P., Ort, C., Rieckermann, J., Scheidegger, A., Vanrolleghem, P. A., & Villez, K. (2019) · Environmental Science & Technology · doi:10.1021/acs.est.8b06481 · Cited by 79
A deep reinforcement learning for user association and power control in heterogeneous networks
Ding, H., Zhao, F., Tian, J., Li, D., & Zhang, H. (2020) · Ad Hoc Networks · doi:10.1016/j.adhoc.2019.102069 · Cited by 75
Skordilis, E., & Moghaddass, R. (2020) · Computers & Industrial Engineering · doi:10.1016/j.cie.2020.106600 · Cited by 71
Huang, L., Zhu, Z., & Zou, Z. (2022) · Automation in Construction · doi:10.1016/j.autcon.2022.104691 · Cited by 78
Data-driven planning via imitation learning
Choudhury, S., Bhardwaj, M., Arora, S., Kapoor, A., Ranade, G. V., Scherer, S., & Dey, D. (2018) · The International Journal of Robotics Research · doi:10.1177/0278364918781001 · Cited by 105
“We refer the reader to [62] for a concise survey on RL and to [6] for a survey on deep RL. Training such policies can be classified into two approaches - either value function-based approach , where a value function for an action is learnt, or policy search , where a policy is directly learnt.”
BackgroundThe Case for Automatic Database Administration using Deep Reinforcement LearningPreprint
Sharma, A., Schuhknecht, F., & Dittrich, J. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1801.05643 · Cited by 87
Exploring optimal control of epidemic spread using reinforcement learning
Ohi, A. Q., Mridha, M. F., Monowar, M. M., & Hamid, M. A. (2020) · Scientific Reports · doi:10.1038/s41598-020-79147-8 · Cited by 89
Applications of Artificial Intelligence in Distribution Power System Operation
Stock, S., Babazadeh, D., & Becker, C. U. (2021) · IEEE Access · doi:10.1109/access.2021.3125102 · Cited by 75
Multi-Echelon Inventory Optimization Using Deep Reinforcement Learning
Geevers, K., van Hezewijk, L., & Mes, M. (2022) · Central European Journal of Operations Research · doi:10.1007/s10100-023-00872-2 · Cited by 88
“Deep learning was combined with RL, because they have a common goal of creating general-purpose artificial intelligence systems that can interact with and learn from their environment (Arulkumaran et al. 2017).”
Ensemble reinforcement learning: A survey
Song, Y., Suganthan, P. N., Pedrycz, W., Ou, J., He, Y., Chen, Y., & Wu, Y. (2023) · Applied Soft Computing · doi:10.1016/j.asoc.2023.110975 · Cited by 88
Transformer-Based Reinforcement Learning for Scalable Multi-UAV Area Coverage
Chen, D., Qi, Q., Fu, Q., Wang, J., Liao, J., & Han, Z. (2024) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2024.3358010 · Cited by 77
“Recently, some studies [12] have discovered the potential of multi-agent deep reinforcement learning (MA-DRL) [13], [14], [15], [16], [17] in the field of long-term decision-making multi-player games, such as AlphaGo Zero and StarCraft II.”
BackgroundAdversarial Machine Learning Attacks and Defences in Multi-Agent Reinforcement Learning
Standen, M., Kim, J., & Szabo, C. (2025) · ACM Computing Surveys · doi:10.1145/3708320 · Cited by 89
A Comprehensive Survey of Agents for Computer Use: Foundations, Challenges, and Future Directions
Sager, P., Meyer, B., Yan, P., von Wartburg-Kottler, R., Etaiwi, L., Enayati, A., Nobel, G., Abdulkadir, A., Grewe, B. F., & Stadelmann, T. (2026) · Journal of Artificial Intelligence Research · doi:10.1613/jair.1.19490 · Cited by 80
Ramola, A., Shakya, A. K., & Bergman, A. (2025) · Photonics · doi:10.3390/photonics12060565 · Cited by 75
Deep Reinforcement Learning Based QoS-Aware Routing in Knowledge-Defined Networking
Quang, P. T. A., Hadjadj‐Aoul, Y., & Outtagarts, A. (2019) · Lecture notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering · doi:10.1007/978-3-030-14413-5_2 · Cited by 70
“Recent breakthroughs in deep neural networks [8,9] has improved the performance of DRL; thus paving a way for adopting DRL in the context of networking.”
MethodA Survey on Slice Admission Control Strategies and Optimization Schemes in 5G Network
Ojijo, M. O., & Falowo, O. E. (2020) · IEEE Access · doi:10.1109/access.2020.2967626 · Cited by 84
“Indeed, if the state becomes too large the action space my grow exponentially [50] and the computational cost involved in solving the MDP becomes too intensive.”
BackgroundKhishe, M., Caraffini, F., & Kühn, S. (2021) · Mathematics · doi:10.3390/math9091002 · Cited by 71
Miglani, A., & Kumar, N. (2021) · Computer Communications · doi:10.1016/j.comcom.2021.07.009 · Cited by 81
Deep Reinforcement Learning for Decentralized Multi-Robot Exploration With Macro Actions
Tan, A. H., Bejarano, F. P., Zhu, Y., Ren, R., & Nejat, G. (2022) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2022.3224667 · Cited by 86
Tang, Y., Hu, W., Zhang, B., Cao, D., Hou, N., Li, Y., Chen, Z., & Blaabjerg, F. (2021) · IEEE Transactions on Energy Conversion · doi:10.1109/tec.2021.3126754 · Cited by 70
A survey of deep learning applications in cryptocurrency
Zhang, J., Cai, K., & Wen, J. (2023) · iScience · doi:10.1016/j.isci.2023.108509 · Cited by 81
Tall Buildings with Dynamic Facade Under Winds
Ding, F., & Kareem, A. (2020) · Engineering · doi:10.1016/j.eng.2020.07.020 · Cited by 69
Reinforcement Learning: An OverviewPreprint
Murphy, K. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2412.05265 · Cited by 92
Jameel, F., Javaid, U., Khan, W. U., Aman, M. N., Pervaiz, H., & Jäntti, R. (2020) · Sustainability · doi:10.3390/su12125161 · Cited by 79
Blockchain-Based Secure Computation Offloading in Vehicular Networks
Zheng, X., Li, M., Chen, Y., Guo, J., Alam, M., & Hu, W. (2020) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2020.3014229 · Cited by 75
Alhazmi, K., Albalawi, F., & Sarathy, S. M. (2021) · Chemical Engineering Journal · doi:10.1016/j.cej.2021.130993 · Cited by 68
“An optimal policy can be learned in several ways [36].”
Fully Adaptive Framework: Neural Computerized Adaptive Testing for Online Education
Zhuang, Y., Liu, Q., Huang, Z., Li, Z., Shen, S., & Ma, H. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i4.20399 · Cited by 68
“Deep reinforcement learning, as one of state-of-the-art techniques (Arulkumaran et al. 2017), has shown superior abilities in many fields, such as games (Hessel et al. 2018; Rajeswaran, Mordatch, and Kumar 2020), robotics (Hu et al. 2020; Haarnoja et al. 2018) and recommender systems (Chen et al.…”
BackgroundSaleem, M. A., Khadim, A., Fatima, M., Khan, M. A., Nair, H. K. G., & Asif, M. K. (2022) · International Conference on Cyber Resilience (ICCR) · doi:10.1109/iccr56254.2022.9995824 · Cited by 72
Kaewdornhan, N., Srithapon, C., Liemthong, R., & Chatthaworn, R. (2023) · Energies · doi:10.3390/en16052357 · Cited by 75
Proximal Policy Optimization for RIS-Assisted Full Duplex 6G-V2X Communications
Saikia, P., Pala, S., Singh, K. K., Singh, S. K., & Huang, W.-J. (2023) · IEEE Transactions on Intelligent Vehicles · doi:10.1109/tiv.2023.3275632 · Cited by 74
Zhang, Y., Zhang, D., & Jiang, H. (2023) · Journal of Marine Science and Engineering · doi:10.3390/jmse11071440 · Cited by 81
The State of AI-Empowered Backscatter Communications: A Comprehensive Survey
Xu, F., Hussain, T., Ahmed, M., Ali, K., Mirza, M. A., Khan, W. U., Ihsan, A., & Han, Z. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3299210 · Cited by 76
“In order to obtain reward feedback in a loop, the DRL method takes inspiration from neural networks, precisely function approximators [84].”
MethodService Function Chain Dynamic Scheduling in Space-Air-Ground Integrated Networks
Jia, Z., Cao, Y., He, L., Wu, Q., Zhu, Q., Niyato, D. T., & Han, Z. (2025) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2025.3543259 · Cited by 76
Deep learning in metasurfaces: from automated design to adaptive metadevices
Saifullah, Y., Wu, N., Wang, H., Zheng, B., Qian, C., & Chen, H. (2025) · Advanced Photonics · doi:10.1117/1.ap.7.3.034005 · Cited by 75
Lindemann, L., Zhao, Y., Yu, X., Pappas, G. J., & Deshmukh, J. V. (2025) · IEEE Control Systems · doi:10.1109/mcs.2025.3611545 · Cited by 100
Yao, H., Zhang, B., Zhang, P., Wu, S., Jiang, C., & Guo, S. (2018) · IEEE Transactions on Emerging Topics in Computing · doi:10.1109/tetc.2018.2871549 · Cited by 72
Nguyen, K. K., Duong, T. Q., Vien, N. A., Le‐Khac, N., & Nguyen, L. D. (2019) · IEEE Access · doi:10.1109/access.2019.2952411 · Cited by 72
“There are value function-based methods, policy search-based methods, and an actor-critic approach that employs both value functions and policy search [30].”
MethodRobot control parameters auto-tuning in trajectory tracking applications
Roveda, L., Forgione, M., & Piga, D. (2020) · Control Engineering Practice · doi:10.1016/j.conengprac.2020.104488 · Cited by 74
Service Function Chain Embedding for NFV-Enabled IoT Based on Deep Reinforcement Learning
Fu, X., Yu, F. R., Wang, J., Qi, Q., & Liao, J. (2019) · IEEE Communications Magazine · doi:10.1109/mcom.001.1900097 · Cited by 73
“In this article, with current attention to deep reinforcement learning (DRL) [7], we present a DRL-based SFC embedding scheme in NFV-enabled IoT.”
MethodLearning to Solve 3-D Bin Packing Problem via Deep Reinforcement Learning and Constraint Programming
Jiang, Y., Cao, Z., & Zhang, J. (2021) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2021.3121542 · Cited by 81
“…have been emerging vigorously, mainly due to the advances of: 1) reinforcement learning frameworks, such as DQN, actor–critic, PPO, and DDPG [23] and 2) deep architectures (as the policy network), such as the pointer network (PtrNet) [24] and its variants, which are designated to handle the…”
BackgroundParameter estimation in quantum sensing based on deep reinforcement learning
Xiao, T., Fan, J., & Zeng, G. (2022) · npj Quantum Information · doi:10.1038/s41534-021-00513-z · Cited by 82
“In RL, the problem to resolve is described as an MDP. Theoretical results in RL rely on the MDP description being a correct match to the problem 38,39 .”
BackgroundDeep reinforcement learning challenges and opportunities for urban water systems
Negm, A. S., Ma, X., & Aggidis, G. (2024) · Water Research · doi:10.1016/j.watres.2024.121145 · Cited by 70
“…policy of the current action-value pair of the current policy (e.g. SARSA) whilst off-policy methods explore the best value policy without necessarily following the current behavioural policy; they are also called optimal methods (e.g. Q-217 learning) (Arulkumaran et al. , 2017; Li, 2017).”
BackgroundMethodInfluentialMakris, I., Karampasi, A. S., Grammatikis, P. I. R., Episkopos, N., Iturbe, E., Rios, E., Piperigkos, N., Lalos, A. S., Xenakis, C., Λάγκας, Θ., et al. (2024) · Computer Science Review · doi:10.1016/j.cosrev.2024.100717 · Cited by 76
Towards a Hands-Free Query Optimizer through Deep LearningPreprint
Marcus, R., & Papaemmanouil, O. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1809.10212 · Cited by 73
Conservative Q-Improvement: Reinforcement Learning for an Interpretable Decision-Tree PolicyPreprint
Roth, A. M., Topin, N., Jamshidi, P., & Veloso, M. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1907.01180 · Cited by 73
“…specifically at the field of Reinforcement Learning (RL), which in recent years has demonstrated impressive results on tasks such as the games of Go and Dota2, walking as a simulated human-like agent, and accomplishing a variety of real-world robotic tasks [Gitau, 2019; Arulkumaran et al. , 2017].”
BackgroundWhen 5G Meets Deep Learning: A Systematic Review
Santos, G. L., Endo, P. T., Sadok, D. F. H., & Kelner, J. (2020) · Algorithms · doi:10.3390/a13090208 · Cited by 76
Agrawal, S. K., Samant, A., & Yadav, S. K. (2022) · Physical Communication · doi:10.1016/j.phycom.2022.101673 · Cited by 66
Lei, X., Xia, Y., Deng, L., & Sun, L. (2022) · Structural and Multidisciplinary Optimization · doi:10.1007/s00158-022-03210-3 · Cited by 66
“Combining the Q-learning algorithm, DQN could be established to iteratively update the neural network hyper-parameters in the action-value function Q(s; a; ) toward the final target (Arulkumaran et al. 2017).”
MethodVesNet-RL: Simulation-Based Reinforcement Learning for Real-World US Probe Navigation
Bi, Y., Jiang, Z., Gao, Y., Wendler, T., Karlas, A., & Navab, N. (2022) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2022.3176112 · Cited by 78
Edge Computing Task Offloading for Environmental Perception of Autonomous Vehicles in 6G Networks
Lv, P., Xu, W., Nie, J., Yuan, Y., Cai, C., Chen, Z., & Xu, J. W. (2022) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2022.3211193 · Cited by 72
“Considering the complexity and real-time requirements of task offload-ing decisions in the edge environment, an algorithm based on deep reinforcement learning (DRL) [28] is adopted to obtain task offloading decisions.”
MethodA review of deep learning segmentation methods for carotid artery ultrasound images
Huang, Q., Tian, H., Jia, L., Li, Z., & Zhou, Z. (2023) · Neurocomputing · doi:10.1016/j.neucom.2023.126298 · Cited by 68
Neurosymbolic Reinforcement Learning and Planning: A Survey
Acharya, K., Raza, W., Dourado, C. M. J. D. M., Velasquez, A., & Song, H. (2023) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2023.3311428 · Cited by 92
“A large number of surveys are available on RL on various aspects, given as follows: 1) RL in general [1], DRL [8], [9], and causal RL [10]; 2) safety and security in RL [11], [12]; 3) environment [13]; 4) agent [14], [15], [16]; 5) application, such as natural language processing [17], [18],…”
BackgroundMethodDong, Y., Zhang, H., Wang, C., & Zhou, X. (2023) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2023.107230 · Cited by 69
Nearest neighbours reveal fast and slow components of motor learning
Kollmorgen, S., Hahnloser, R. H. R., & Mante, V. (2020) · Nature · doi:10.1038/s41586-019-1892-x · Cited by 66
An Empirical Analysis of Deep Learning for Cardinality EstimationPreprint
Ortíz, J., Bałazińska, M., Gehrke, J. E., & Keerthi, S. S. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1905.06425 · Cited by 80
Kim, H., Lim, D.-E., & Lee, S. (2020) · IEEE Transactions on Semiconductor Manufacturing · doi:10.1109/tsm.2020.2965293 · Cited by 69
“The biggest advantage of DNN is its ability to extract low-dimensional representation of high-dimensional data [41].”
BackgroundSelective Transfer With Reinforced Transfer Network for Partial Domain Adaptation
Chen, Z., Chen, C., Cheng, Z., Jiang, B., Fang, K., & Jin, X. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.01272 · Cited by 74
A Control Strategy Based on Deep Reinforcement Learning Under the Combined Wind-Solar Storage System
Huang, S., Li, P., Yang, M. Y., Gao, Y., Yun, J., & Zhang, C. (2021) · IEEE Transactions on Industry Applications · doi:10.1109/tia.2021.3105497 · Cited by 65
“RL is used to make decisions and optimize targets while DL is good at perceiving features and data mining [38].”
MethodLi, G., Wu, J., Deng, C., Xu, X., & Shao, X. (2021) · IEEE/ASME Transactions on Mechatronics · doi:10.1109/tmech.2021.3124415 · Cited by 70
“The deep learning provides the ability for agent to perceive dynamic environment, while the reinforcement learning enables the agnet to learn the optimal strategy [30].”
BackgroundAvoiding Catastrophe: Active Dendrites Enable Multi-Task Learning in Dynamic Environments
Iyer, A., Grewal, K., Velu, A., Souza, L. O., Forest, J., & Ahmad, S. (2022) · Frontiers in Neurorobotics · doi:10.3389/fnbot.2022.846219 · Cited by 87
“Deep RL uses deep learning networks to represent the policy function (see Arulkumaran et al. [2017] for a review).”
BackgroundHUNTER: An Online Cloud Database Hybrid Tuning System for Personalized Requirements
Cai, B., Liu, Y., Zhang, C., Zhang, G.-Y., Zhou, K., Liu, L., Li, C., Cheng, B., Yang, J., & Xing, J. (2022) · ACM SIGMOD International Conference on Management of Data (SIGMOD) · doi:10.1145/3514221.3517882 · Cited by 82
“In fact, although the superior samples can help DDPG acquire better performance, it is hard to reduce the recommendation time without modification, because DDPG, as a DRL algorithm, is difficult to converge quickly [2, 16].”
BackgroundRongcai, Z., Hongwei, X., & Yuan, K. (2023) · Ocean Engineering · doi:10.1016/j.oceaneng.2023.113779 · Cited by 65
Jiang, C., Li, X., Lin, J., Liu, M. Z., & Ma, Z. (2023) · Automation in Construction · doi:10.1016/j.autcon.2023.104817 · Cited by 65
“The control problems must be formulated as a MDP to adopt DRL algorithms [16], e.”
BackgroundMethodInfluentialRen, K., Zeng, Y., Zhong, Y., Sheng, B., & Zhang, Y. (2023) · Journal Of Big Data · doi:10.1186/s40537-023-00814-4 · Cited by 82
“Anomaly-based intrusion detection systems have gained dominance in IDS by incorporating machine learning techniques such as decision tree (DT), deep learning, and reinforcement learning [7–12].”
BackgroundMethodZhang, W., Xiao, G., Gen, M., Geng, H., Wang, X., Deng, M., & Zhang, G. (2024) · Frontiers in Industrial Engineering · doi:10.3389/fieng.2024.1337174 · Cited by 75
Reinforcement Learning Solutions for Microgrid Control and Management: A Survey
Barbalho, P. I. N., Moraes, A. L., Lacerda, V. A., Barra, P. H. A., Fernandes, R. A. S., & Coury, D. V. (2025) · IEEE Access · doi:10.1109/access.2025.3546578 · Cited by 75
“In the MFs, the training is based only on trial-and-error actions, and the agent learns an optimal policy without predicting the state transitions of the environment and the reward [19].”
Optimising a Microgrid System by Deep Reinforcement Learning Techniques
Domínguez-Barbero, C., García‐González, J., Sanz-Bobi, M. A., & Sánchez-Úbeda, E. F. (2020) · Energies · doi:10.3390/en13112830 · Cited by 69
Spatio-temporal feature fusion for dynamic taxi route recommendation via deep reinforcement learning
Ji, S., Wang, Z., Li, T., & Zheng, Y. (2020) · Knowledge-Based Systems · doi:10.1016/j.knosys.2020.106302 · Cited by 72
Hirchoua, B., Ouhbi, B., & Frikh, B. (2021) · Expert Systems with Applications · doi:10.1016/j.eswa.2020.114553 · Cited by 78
Lasers that learn: The interface of laser machining and machine learning
Mills, B., & Grant‐Jacob, J. A. (2021) · IET Optoelectronics · doi:10.1049/ote2.12039 · Cited by 71
A review on the recent applications of deep learning in predictive drug toxicological studies
Sinha, K., Ghosh, N., & Sil, P. C. (2023) · Chemical Research in Toxicology · doi:10.1021/acs.chemrestox.2c00375 · Cited by 64
Achieving better connections between deposited lines in additive manufacturing via machine learning
Jiang, J., Yu, C., Xu, X. W., Ma, Y., & Liu, J. (2020) · Mathematical Biosciences & Engineering · doi:10.3934/mbe.2020191 · Cited by 67
“The most commonly used machine learning techniques include three types: Supervised learning [36], unsupervised learning [37] and reinforcement learning [38].”
MethodA deep reinforcement learning approach for the meal delivery problem
Jahanshahi, H., Bozanta, A., Çevik, M., Kavuk, E. M., Tosun, A., Sonuç, S. B., Kosucu, B., & Bener, A. (2022) · Knowledge-Based Systems · doi:10.1016/j.knosys.2022.108489 · Cited by 71
“These improvements remove the correlation between consecutive samples and also allow the neural network updates to use each step multiple times; hence, it addresses the fundamental instability problem in RL [2].”
BackgroundCloud–Edge Collaborative SFC Mapping for Industrial IoT Using Deep Reinforcement Learning
Xu, S., Li, Y., Guo, S., Lei, C., Liu, D., & Qiu, X. (2021) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2021.3113875 · Cited by 64
Artificial Intelligence and the Common Sense of Animals
Shanahan, M., Crosby, M., Beyret, B., & Cheke, L. (2020) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2020.09.002 · Cited by 58
“This is especially true in a reinforcement learning (RL) context, where, thanks to progress in deep learning [22], it is now possible to bring the methods of comparative cognition directly to bear [23,24].”
Deep Reinforcement Learning for Active High Frequency TradingPreprint
Briola, A., Turiel, J., Marcaccioli, R., Cauderan, A., & Aste, T. (2021) · RePEc: Research Papers in Economics · doi:10.48550/arxiv.2101.07107 · Cited by 76
“In their basic configuration, RL algorithms have proved to be not efficient in dealing with the curse of dimensionality [1].”
BackgroundDeep Learning for Camera Autofocus
Wang, C., Huang, Q., Cheng, M., Ma, Z., & Brady, D. J. (2021) · IEEE Transactions on Computational Imaging · doi:10.1109/tci.2021.3059497 · Cited by 64
“Deep reinforcement learning is an active area with extensive applications [52], and here we briefly review the vision-related tasks that learn a policy to explore uncertainty.”
BackgroundLearning to Drive Like Human Beings: A Method Based on Deep Reinforcement Learning
Tian, Y., Cao, X., Huang, K., Fei, C., Zheng, Z., & Ji, X. (2022) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2021.3055899 · Cited by 67
Remote Sensing Object Tracking With Deep Reinforcement Learning Under Occlusion
Cui, Y., Hou, B., Wu, Q., Ren, B., Wang, S., & Jiao, L. (2021) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2021.3096809 · Cited by 71
Metadata
- Authors
- Arulkumaran, K., Deisenroth, M. P., Brundage, M., & Bharath, A. A.
- Affiliations
- University of Oxford (GB)
- Institution of Engineering and Technology (GB)
- Venue
- IEEE Signal Processing Magazine
- Published
- 2017-11
- Volume
- 34
- Issue
- 6
- Pages
- 26-38
- ISSN
- 1053-5888 (ISSN-L)
- OpenAlex
- W3100789280
- Cited by
- 4,733
- References
- 182
- Citations by year
- 2026: 444 · 2025: 855 · 2024: 806 · 2023: 749 · 2022: 590 · 2021: 511 · 2020: 388 · 2019: 260 · 2018: 122 · 2017: 6
- FWCI
- 126.58
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE Signal Processing Magazine (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- UCL Discovery (University College London) (submittedVersion)
- Spiral (Imperial College London) (submittedVersion)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Reinforcement Learning in Robotics
- Physical Sciences › Computer Science › Artificial Intelligence › Artificial Intelligence in Games
- Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
- Keywords
- deep reinforcement learning, value-based methods, deep Q-network, trust region policy optimization, asynchronous advantage actor-critic, visual understanding, autonomous systems
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.