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 401–500 of 1,021
Gautam, M. (2023) · Electricity · doi:10.3390/electricity4040020 · Cited by 114
“Contrary to conventional rule-based chess systems, AlphaGo utilized the strength of neural networks trained using both supervised learning and RL, along with a standard heuristic search approach [38].”
BackgroundMethodApplying Neural Network to Health Estimation and Lifetime Prediction of Lithium-Ion Batteries
Li, P., Wu, X., Grosu, R., Hou, J., Илолов, М., & Xiang, S. (2024) · IEEE Transactions on Transportation Electrification · doi:10.1109/tte.2024.3457621 · Cited by 107
A Survey of Deep Learning Applications to Autonomous Vehicle Control
Kuutti, S., Bowden, R., Jin, Y., Barber, P., & Fallah, S. (2021) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2019.2962338 · Cited by 96
“Therefore, intrinsic motivation is implemented in the system, encouraging the agent to learn about its environment, whilst exploitation can be taken advantage of in states which have already been explored adequately [51], [117]–[119].”
BackgroundHu, J., Zhang, H., Song, L., Han, Z., & Poor, H. V. (2020) · IEEE Wireless Communications · doi:10.1109/mwc.001.1900262 · Cited by 96
“In deep reinforcement learning, deep neural networks are utilized to handle high-dimensional state spaces [11].”
MethodWon, D.-O., Müller, K., & Lee, S. (2020) · Science Robotics · doi:10.1126/scirobotics.abb9764 · Cited by 106
A survey of modeling for prognosis and health management of industrial equipment
Yucesan, Y., Dourado, A. D. P., & Viana, F. A. C. (2021) · Advanced Engineering Informatics · doi:10.1016/j.aei.2021.101404 · Cited by 107
A Survey of Industrial AIoT: Opportunities, Challenges, and Directions
Awaisi, K. S., Ye, Q., & Sampalli, S. (2024) · IEEE Access · doi:10.1109/access.2024.3426279 · Cited by 109
A Survey on Interactive Reinforcement Learning: Design Principles and Open Challenges
Cruz, C. A., & Igarashi, T. (2020) · Designing Interactive Systems Conference · doi:10.1145/3357236.3395525 · Cited by 131
Review and Evaluation of Reinforcement Learning Frameworks on Smart Grid Applications
Vamvakas, D. G., Michailidis, P., Korkas, C. D., & Kosmatopoulos, E. B. (2023) · Energies · doi:10.3390/en16145326 · Cited by 104
Advances and Challenges of Single‐Pixel Imaging Based on Deep Learning
Song, K., Bian, Y., Wang, D., Li, R., Wu, K., Liu, H., Qin, C., Hu, J., & Xiao, L. (2024) · Laser & Photonics Review · doi:10.1002/lpor.202401397 · Cited by 106
Market sentiment-aware deep reinforcement learning approach for stock portfolio allocation
Koratamaddi, P., Wadhwani, K., Gupta, M., & Sanjeevi, S. G. (2021) · Engineering Science and Technology an International Journal · doi:10.1016/j.jestch.2021.01.007 · Cited by 121
On AI Approaches for Promoting Maternal and Neonatal Health in Low Resource Settings: A Review
Khan, M., Khurshid, M., Vatsa, M., Singh, R., Duggal, M., & Singh, K. (2022) · Frontiers in Public Health · doi:10.3389/fpubh.2022.880034 · Cited by 106
“ML includes various methods including supervised (6, 7), unsupervised (8, 9), semi-supervised (10, 11), and reinforcement learning (12, 13).”
MethodComputation Offloading and Resource Allocation in NOMA–MEC: A Deep Reinforcement Learning Approach
Shang, C., Sun, Y., Luo, H., & Guizani, M. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3264206 · Cited by 99
“neural networks (DNNs) to approximate value functions and policies and achieve the end-to-end optimization [15].”
BackgroundWang, T., Zhu, Z., Zhang, J., Tian, J., & Zhang, W. (2024) · Transportation Research Part C Emerging Technologies · doi:10.1016/j.trc.2024.104582 · Cited by 101
Towards the application of machine learning in digital twin technology: a multi-scale review
Nele, L., Mattera, G., Yap, E. W., Vozza, M., & Vespoli, S. (2024) · Discover Applied Sciences · doi:10.1007/s42452-024-06206-4 · Cited by 103
“RL has emerged as a powerful approach for solving complex decision-making problems in various domains, including robotics, game play, and autonomous systems [94, 95].”
RLHF Deciphered: A Critical Analysis of Reinforcement Learning from Human Feedback for LLMs
Chaudhari, S., Aggarwal, P., Murahari, V., Rajpurohit, T., Kalyan, A., Narasimhan, K., Deshpande, A., & Castro da Silva, B. (2026) · ACM Computing Surveys · doi:10.1145/3743127 · Cited by 92
“We next discuss a few commonly used RL algorithms for the process.”
A deep reinforcement learning based offloading scheme in ad-hoc mobile clouds
Le, D. V., & Tham, C. (2018) · IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS) · doi:10.1109/infcomw.2018.8406881 · Cited by 102
“[8] Deep RL has emerged as a class of RL schemes which uses the deep learning [9] to enable reinforcement learning to scale to decision-making problems with high-dimensional state and action spaces [10].”
Wu, F., Zhang, H., Wu, J., Han, Z., Poor, H. V., & Song, L. (2021) · IEEE Transactions on Communications · doi:10.1109/tcomm.2021.3065135 · Cited by 99
Hybrid Reinforcement Learning for STAR-RISs: A Coupled Phase-Shift Model Based Beamformer
Zhong, R., Liu, Y., Mu, X., Chen, Y., Wang, X., & Hanzo, L. (2022) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2022.3192053 · Cited by 102
“The DDPG and DQN algorithms are capable of optimizing independent continuous and discrete problems [38].”
When large language model meets optimization
Huang, S., Yang, K., Qi, S., & Wang, R. (2024) · Swarm and Evolutionary Computation · doi:10.1016/j.swevo.2024.101663 · Cited by 116
Explainable reinforcement learning for broad-XAI: a conceptual framework and survey
Dazeley, R., Vamplew, P. W., & Cruz, F. (2023) · Neural Computing and Applications · doi:10.1007/s00521-023-08423-1 · Cited by 107
Park, M.-J., Lee, J., Cho, K., Park, J., Moon, J., Lee, S.-H., Kim, T., Oh, S., Choi, S.-W., Choi, Y., et al. (2022) · IEEE Journal of Solid-State Circuits · doi:10.1109/jssc.2022.3193354 · Cited by 91
A value-based deep reinforcement learning model with human expertise in optimal treatment of sepsis
Wu, X., Li, R., He, Z., Yu, T., & Cheng, C. (2023) · npj Digital Medicine · doi:10.1038/s41746-023-00755-5 · Cited by 116
“The blind trust of AI algorithms in decision making for healthcare management without clinician supervision has led to increased medical risks and safety issues 24,25 .”
BackgroundZhao, Z., Lee, C. K. M., & Ren, J. (2023) · Applied Energy · doi:10.1016/j.apenergy.2023.122278 · Cited by 98
Pinthurat, W., Surinkaew, T., & Hredzak, B. (2024) · Renewable and Sustainable Energy Reviews · doi:10.1016/j.rser.2024.114648 · Cited by 103
Zhang, H., Zhao, H., Liu, R., Kaushik, A., Gao, X., & Xu, S. (2024) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2024.3405642 · Cited by 96
Robot Navigation of Environments with Unknown Rough Terrain Using deep Reinforcement Learning
Zhang, K., Niroui, F., Ficocelli, M., & Nejat, G. (2018) · 2018 IEEE International Symposium on Safety, Security, and Rescue Robotics (SSRR) · doi:10.1109/ssrr.2018.8468643 · Cited by 86
“DRL can directly use raw sensory data to determine robot navigation actions without the need of pre-labeled data [24].”
Heo, S., Nam, K., Loy-Benitez, J., Li, Q., Lee, S.-C., & Yoo, C. (2019) · Energy and Buildings · doi:10.1016/j.enbuild.2019.109440 · Cited by 90
Intelligent Network Slicing for V2X Services Toward 5G
Mei, J., Wang, X., & Zheng, K. (2019) · IEEE Network · doi:10.1109/mnet.001.1800528 · Cited by 90
“Compared to value function methods, policy search methods do not need to maintain a value function model, but directly search for an optimal policy at the cost of needing more training data samples and a slower rate of convergence [15].”
MethodXue, Z., Zhou, P., Xu, Z., Wang, X., Xie, Y., Ding, X., & Wen, S. (2021) · IEEE Internet of Things Journal · doi:10.1109/jiot.2021.3057653 · Cited by 84
“Then, to make full use of deep learning, the deep RL approach [28], [29], by effectively combining deep learning and RL, has emerged as the continuous decision-making method for clinician domain [30].”
MethodLang, S., Behrendt, F., Lanzerath, N., Reggelin, T., & Müller, M. (2020) · Winter Simulation Conference · doi:10.1109/wsc48552.2020.9383997 · Cited by 90
“However, the main difference between deep reinforcement learning and traditional methods is that the agent is always represented by one or more deep neural networks (Arulkumaran et al. 2017; François-Lavet et al. 2018; Sutton and Barto 2018).”
BackgroundA survey of machine learning techniques for improving Global Navigation Satellite Systems
Mohanty, A., & Gao, G. (2024) · EURASIP Journal on Advances in Signal Processing · doi:10.1186/s13634-024-01167-7 · Cited by 98
“Notable RL algorithms include Q-learning, deep Q-net-works (DQN), proximal policy optimization (PPO), and advantage actor-critic (A2C) [49, 50].”
BackgroundPolicy search in continuous action domains: An overview
Sigaud, O., & Stulp, F. (2019) · Neural Networks · doi:10.1016/j.neunet.2019.01.011 · Cited by 105
Analogues of mental simulation and imagination in deep learning
Hamrick, J. B. (2019) · Current Opinion in Behavioral Sciences · doi:10.1016/j.cobeha.2018.12.011 · Cited by 83
“Much of the research in deep RL is model-free in that it aims to learn policies without knowing anything about the transition, observation, recognition, or reward functions (see [23] for a review).”
Generative Adversarial Networks: A Literature Review
Cheng, J., Yang, Y., Tang, X., Xiong, N. N., Zhang, Y., & Lei, F. (2020) · KSII Transactions on Internet and Information Systems · doi:10.3837/tiis.2020.12.001 · Cited by 126
“Reinforcement learning is used to describe and solve the problem that agents use learning strategies to maximize returns or achieve specific goals in their interaction with the environment [12, 13].”
BackgroundEdge Intelligence—Research Opportunities for Distributed Computing Continuum Systems
Pujol, V. C., Donta, P. K., Morichetta, A., Murturi, I., & Dustdar, S. (2023) · IEEE Internet Computing · doi:10.1109/mic.2023.3284693 · Cited by 111
“through experiences by interacting with the environment.(73) DRL is used in various fields, including natural language processing, vehicular communication, robotics, health care, computer vision, etc.”
MethodDeep Learning Challenges and Prospects in Wireless Sensor Network Deployment
Qiu, Y., Ma, L., & Priyadarshi, R. (2024) · Archives of Computational Methods in Engineering · doi:10.1007/s11831-024-10079-6 · Cited by 97
Challenges and Prospects of Digital Core-Reconstruction Research
Zhu, L., Zhang, C., Zhang, C., Zhou, X., Zhang, Z., Nie, X., Liu, W., & Zhu, B. (2019) · Geofluids · doi:10.1155/2019/7814180 · Cited by 88
Survey on Machine Learning for Traffic-Driven Service Provisioning in Optical Networks
Panayiotou, T., Michalopoulou, M., & Ellinas, G. (2023) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2023.3247842 · Cited by 94
“For more information on value-based, policy-based, and actor-critic methods, RL advances, as well as multi-agent systems, the reader is referred to [177], [178].”
Deep Reinforcement Learning for Multimedia Traffic Control in Software Defined Networking
Huang, X., Yuan, T., Qiao, G., & Ren, Y. (2018) · IEEE Network · doi:10.1109/mnet.2018.1800097 · Cited by 87
“According to [8], Deep Reinforcement Learning ( DRL) combines deep learning and reinforcement learning experience interacting with the environment to achieve successful strategies that lead to the greatest long-term rewards.”
BackgroundNavigating through the Maze of Homogeneous Catalyst Design with Machine Learning
Gomes, G., Pollice, R., & Aspuru‐Guzik, A. (2021) · Trends in Chemistry · doi:10.1016/j.trechm.2020.12.006 · Cited by 94
DRLinFluids: An open-source Python platform of coupling deep reinforcement learning and OpenFOAM
Wang, Q., Yan, L., Hu, G., Li, C., Xiao, Y., Xiong, H., Rabault, J., & Noack, B. R. (2022) · Physics of Fluids · doi:10.1063/5.0103113 · Cited by 99
“As the foundations and main aspects of common DRL algorithms have been discussed extensively in both the DRL-centric literature [Arulkumaran et al., 2017, Schulman et al., 2017, Lillicrap et al., 2019, Haarnoja et al., 2018], and the fluid mechanics literature [Garnier et al., 2021, Rabault et al.,…”
Reinforcement Learning for Disassembly System Optimization Problems: A Survey
Guo, X., Bi, Z., Wang, J., Qin, S., Liu, S., & Qi, L. (2023) · International Journal of Network Dynamics and Intelligence · doi:10.53941/ijndi0201001 · Cited by 103
Price-Matching-Based Regional Energy Market With Hierarchical Reinforcement Learning Algorithm
Zhang, N., Yan, J., Hu, C., Sun, Q., Yang, L., Gao, W., Guerrero, J. M., & Li, Y. (2024) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2024.3390595 · Cited by 87
Demand Response Management for Industrial Facilities: A Deep Reinforcement Learning Approach
Huang, X., Hong, S. H., Yu, M., Ding, Y., & Jiang, J. (2019) · IEEE Access · doi:10.1109/access.2019.2924030 · Cited by 86
“The words actor and critic are synonyms of the policy and value function, respectively [10].”
Path planning for asteroid hopping rovers with pre-trained deep reinforcement learning architectures
Jiang, J., Zeng, X., Guzzetti, D., & You, Y. (2020) · Acta Astronautica · doi:10.1016/j.actaastro.2020.03.007 · Cited by 89
Big data analytics deep learning techniques and applications: A survey
Selmy, H. A., Mohamed, H. K., & Medhat, W. (2023) · Information Systems · doi:10.1016/j.is.2023.102318 · Cited by 99
Yang, N., Chen, S., Zhang, H., & Berry, R. (2024) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2024.3405075 · Cited by 114
“The agents can learn by making predictions about the consequences of their actions [27].”
BackgroundMobility-Aware Seamless Service Migration and Resource Allocation in Multi-Edge IoV Systems
Chen, Z., Huang, S., Min, G., Ning, Z., Li, J., & Zhang, Y. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3540407 · Cited by 86
Iriondo, A., Lazkano, E., Susperregi, L., Urain, J., Fernández, A., & Molina, J. (2019) · Applied Sciences · doi:10.3390/app9020348 · Cited by 95
Machine Learning and Deep Learning: Open Issues and Future Research Directions for the Next 10 Years
Pramod, A., Naicker, H. S., & Tyagi, A. K. (2021) · Computational Analysis and Deep Learning for Medical Care · doi:10.1002/9781119785750.ch18 · Cited by 85
Attacking Black-box Recommendations via Copying Cross-domain User Profiles
Fan, W., Derr, T., Zhao, X., Ma, Y., Liu, H., Wang, J., Tang, J., & Li, Q. (2021) · 2021 IEEE 37th International Conference on Data Engineering (ICDE) · doi:10.1109/icde51399.2021.00140 · Cited by 77
“Most existing RL techniques cannot handle such a large discrete action space problem as the time complexity of making a decision is linear to the size of action space [30], [31], [33], [35], [36].”
BackgroundReinforcement Learning in Game Industry—Review, Prospects and Challenges
Souchleris, K., Sidiropoulos, G. K., & Papakostas, G. A. (2023) · Applied Sciences · doi:10.3390/app13042443 · Cited by 98
Talaat, F. M. (2022) · Multimedia Tools and Applications · doi:10.1007/s11042-022-13000-0 · Cited by 100
“In Q-learning, there is no direct assignment of Q values such as the table-based view, instead, an error function is presented to calculate the difference between the existing Q value and the new value.”
MethodInfluentialSun, D., Jamshidnejad, A., & Schutter, B. D. (2024) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2023.3342651 · Cited by 95
“These two control measures can be either used independently or coordinated together within a control method, such as model predictive control (MPC) [5] or deep reinforcement learning (DRL) [6].”
MethodKhadivi, M., Charter, T., Yaghoubi, M., Jalayer, M., Ahang, M., Shojaeinasab, A., & Najjaran, H. (2025) · Computers & Industrial Engineering · doi:10.1016/j.cie.2025.110856 · Cited by 105
“This work addressed instability issues of function approximation techniques and excited the community with the ability to learn from high dimensional observations, kickstarting a revolution of advancements in the field of DRL (Arulkumaran et al., 2017).”
BackgroundDeep reinforcement learning for multi-objective placement of virtual machines in cloud datacenters
Caviglione, L., Gaggero, M., Paolucci, M., & Ronco, R. (2020) · Soft Computing · doi:10.1007/s00500-020-05462-x · Cited by 89
“Toward this end, in this paperwe propose amechanism forVMplacement based on deep reinforcement learning (DRL) (Arulkumaran et al. 2017).”
Differential evolution with mixed mutation strategy based on deep reinforcement learning
Tan, Z., & Li, K. (2021) · Applied Soft Computing · doi:10.1016/j.asoc.2021.107678 · Cited by 103
Chien, C., & Lan, Y.-B. (2021) · Computers & Industrial Engineering · doi:10.1016/j.cie.2021.107782 · Cited by 83
Zhang, J., Cheng, L., Liu, C., Zhao, Z., & Mao, Y. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.120972 · Cited by 83
Improving Short Text Classification With Augmented Data Using GPT-3
Balkus, S. V., & Yan, D. (2022) · Natural Language Engineering · doi:10.1017/s1351324923000438 · Cited by 68
“5 Comparison to other data augmentation techniques Finally, even though example optimization did not produce consistent results on the test set, future work may consider alternative optimization techniques for selecting the best example-label pairs to provide to GPT-3, such as reinforcement learning (Arulkumaran et al…”
BackgroundRETRACTED ARTICLE: A review on quantum computing and deep learning algorithms and their applications
Valdez, F., & Melín, P. (2022) · Soft Computing · doi:10.1007/s00500-022-07037-4 · Cited by 92
RetractedAn obstacle avoidance approach for UAV path planning
Bashir, N., Boudjit, S., Dauphin, G., & Zeadally, S. (2023) · Simulation Modelling Practice and Theory · doi:10.1016/j.simpat.2023.102815 · Cited by 87
A comprehensive survey on client selection strategies in federated learning
Li, J., Chen, T., & Teng, S. (2024) · Computer Networks · doi:10.1016/j.comnet.2024.110663 · Cited by 103
From Chess and Atari to StarCraft and Beyond: How Game AI is Driving the World of AI
Risi, S., & Preuss, M. (2020) · Künstliche Intell. · doi:10.1007/s13218-020-00647-w · Cited by 94
“[3] provides a good overview over deep RL.”
BackgroundStrong Mixed-Integer Programming Formulations for Trained Neural Networks
Anderson, R., Huchette, J., Tjandraatmadja, C., & Vielma, J. P. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-17953-3_3 · Cited by 99
Fragkos, G., Johnson, J. D., & Tsiropoulou, E. E. (2022) · IEEE Transactions on Human-Machine Systems · doi:10.1109/thms.2022.3163185 · Cited by 91
Ahmed, N., Ngadi, M. A., Sharif, J. M., Hussain, S., Uddin, M., Rathore, M. S., Iqbal, J., Abdelhaq, M., Alsaqour, R. A., Ullah, S. S., et al. (2022) · Sensors · doi:10.3390/s22207896 · Cited by 97
“Group of players, utility functions, and techniques are involved in a game [61].”
BackgroundMulti-UAV Path Learning for Age and Power Optimization in IoT With UAV Battery Recharge
Eldeeb, E., Sant’Ana, J. M. D. S., Pérez, D. E., Shehab, M. A., Mahmood, N. H., & Alves, H. (2022) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2022.3222092 · Cited by 81
“The state-action value function (Q-function) Qπ(s, a) describes how good an action a is at state s while following the policy π [14].”
BackgroundIntelligent Robots and Human-Robot Collaboration in the Construction Industry: A Review
Wei, H.-H., Zhang, Y., Sun, X., Chen, J., & Li, S. (2023) · Journal of Intelligent Construction · doi:10.26599/jic.2023.9180002 · Cited by 95
Kang, D., Kang, D., Hwangbo, S., Niaz, H., Lee, W. B., Liu, J. J., & Na, J. (2023) · Energy · doi:10.1016/j.energy.2023.128623 · Cited by 87
Deep reinforcement learning for inverse inorganic materials design
Karpovich, C. J., Pan, E., & Olivetti, E. (2024) · npj Computational Materials · doi:10.1038/s41524-024-01474-5 · Cited by 94
“Deep RL is particularly suited to guided materials generation tasks as it is adept at learning from high-dimensional data that is com-monlyencounteredinthematerialsspace 45 .”
Du, Y., Mukherjee, T., Li, R., Hou, Z., Dutta, S., Arnold, C. B., Elwany, A. H., Kung, S., Tang, J., & DebRoy, T. (2025) · Progress in Materials Science · doi:10.1016/j.pmatsci.2025.101587 · Cited by 100
Machine learning based video coding optimizations: A survey
Zhang, Y., Kwong, S., & Wang, S. (2019) · Information Sciences · doi:10.1016/j.ins.2019.07.096 · Cited by 88
Recommending Cryptocurrency Trading Points with Deep Reinforcement Learning Approach
Sattarov, O., Muminov, A., Lee, C. W., Kang, H. K., Oh, R.-D., Ahn, J., Oh, H. J., & Jeon, H. S. (2020) · Applied Sciences · doi:10.3390/app10041506 · Cited by 90
Deep reinforcement learning for robust emotional classification in facial expression recognition
Li, H., & Xu, H. (2020) · Knowledge-Based Systems · doi:10.1016/j.knosys.2020.106172 · Cited by 88
Qin, Z., Liu, Z., Han, G., Lin, C., Guo, L., & Xie, L. (2021) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2021.3117792 · Cited by 85
Li, R., Zhang, C., Xie, W., Gong, Y., Ding, F., Dai, H., Chen, Z., Yin, F., & Zhang, Z. (2023) · Nanophotonics · doi:10.1515/nanoph-2022-0692 · Cited by 93
Mourtzis, D., Angelopoulos, J., & Panopoulos, N. (2023) · International Journal of Production Research · doi:10.1080/00207543.2023.2232470 · Cited by 90
Kim, S., Choi, Y., Park, J., Adams, D., Heo, S., & Lee, J. H. (2023) · Renewable and Sustainable Energy Reviews · doi:10.1016/j.rser.2023.114049 · Cited by 86
On Improving Cross-dataset Generalization of Deepfake Detectors
Nadimpalli, A. V., & Rattani, A. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw56347.2022.00019 · Cited by 79
“Reinforcement learning (RL) [4, 32] is a branch of machine learning that studies how an intelligent agent should operate in a given environment to maximize the concept of cumulative reward.”
BackgroundAdaptive Modulation and Coding Based on Reinforcement Learning for 5G Networks
Mota, M. P., Araújo, D. C., Neto, F. H. C., de Almeida, A. L. F., & Cavalcanti, F. R. P. (2019) · IEEE Globecom Workshops (GC Wkshps) · doi:10.1109/gcwkshps45667.2019.9024384 · Cited by 79
“Moreover, a comparison with other RLbased algorithms such as multi-armed bandits (MABs) [26] or deep RL solutions [27] is envisioned.”
BackgroundLuis, S. Y., Reina, D. G., & Toral, S. L. (2021) · IEEE Access · doi:10.1109/access.2021.3053348 · Cited by 90
Qian, G., & Liu, J. (2022) · Progress in Nuclear Energy · doi:10.1016/j.pnucene.2022.104401 · Cited by 79
Mei, P., Karimi, H. R., Xie, H., Chen, F., Huang, C., & Yang, S. (2023) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2023.106239 · Cited by 79
Target Search Control of AUV in Underwater Environment With Deep Reinforcement Learning
Cao, X. S., Sun, C., & Yan, M. (2019) · IEEE Access · doi:10.1109/access.2019.2929120 · Cited by 79
“DRL learns the low-dimensional state features of the high-dimensional state from the sensory data and solves the limitation of traditional RL technology due to the large state and action space [18], [19].”
BackgroundLearning to select cuts for efficient mixed-integer programming
Huang, Z., Wang, K., Liu, F., Zhen, H., Zhang, W., Yuan, M., Hao, J., Yu, Y. H., & Wang, Y.-X. (2021) · Pattern Recognition · doi:10.1016/j.patcog.2021.108353 · Cited by 117
Multi-Agent Deep Reinforcement Learning for Multi-Echelon Inventory Management
Liu, X., Hu, M., Peng, Y., & Yang, Y. (2022) · Production and Operations Management · doi:10.1177/10591478241305863 · Cited by 110
From screens to scenes: A survey of embodied AI in healthcare
Liu, Y., Cao, X., Chen, T., Jiang, Y., You, J., Wu, M., Wang, X., Feng, M., Jin, Y., & Chen, J. (2025) · Information Fusion · doi:10.1016/j.inffus.2025.103033 · Cited by 91
A Review of Artificial Intelligence Algorithms Used for Smart Machine Tools
Chang, C., Lee, H.-W., & Liu, C.-H. (2018) · Inventions · doi:10.3390/inventions3030041 · Cited by 89
“[66] highlighted the unique advantages of deep neural networks, focusing on visual understanding via reinforcement learning.”
RL-Cache: Learning-Based Cache Admission for Content Delivery
Kirilin, V., Sundarrajan, A., Gorinsky, S., & Sitaraman, R. K. (2020) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2020.3000415 · Cited by 85
“One can roughly classify model-free RL algorithms into TD (Temporal Difference), MC (Monte Carlo), and DPS (Direct Policy Search) types [26].”
MethodWu, F., Zhang, H., Wu, J., & Song, L. (2020) · IEEE Transactions on Communications · doi:10.1109/tcomm.2020.2986289 · Cited by 83
“We investigate on the UAV trajectory design problem using multi-agent DRL, and propose a DQN-based multi-UAV trajectory design algorithm in Section V.”
BackgroundMethodInfluentialFusion of Machine Learning and MPC under Uncertainty: What Advances Are on the Horizon?
Mesbah, A., Wabersich, K. P., Schoellig, A. P., Zeilinger, M. N., Lucia, S., Badgwell, T. A., & Paulson, J. A. (2022) · Proceedings of the ... American Control Conference/Proceedings of the American Control Conference · doi:10.23919/acc53348.2022.9867643 · Cited by 101
“In essence, DRL uses deep neural networks to parameterize the policy and value functions and, thus, is particularly suitable for dealing with continuous state and action spaces.”
BackgroundA Review of Deep Reinforcement Learning Algorithms for Mobile Robot Path Planning
Singh, R., Ren, J., & Lin, X. (2023) · Vehicles · doi:10.3390/vehicles5040078 · Cited by 92
Reinforcement learning in reliability and maintenance optimization: A tutorial
Zhang, Q., Liu, Y., Xiang, Y., & Xiahou, T. (2024) · Reliability Engineering & System Safety · doi:10.1016/j.ress.2024.110401 · Cited by 82
A Comprehensive Survey on Evidential Deep Learning and its Applications
Gao, J., Chen, M., Xiang, L., & Xu, C. (2025) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2025.3625258 · Cited by 115
“The goal of the agent is to learn the optimal policy that maximizes the expected cumulative reward [141].”
Multi-Agent DRL for Resource Allocation and Cache Design in Terrestrial-Satellite Networks
Li, X., Zhang, H., Zhou, H., Wang, N., Long, K., Al–Rubaye, S., & Karagiannidis, G. K. (2022) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2022.3231379 · Cited by 80
“selecting actions through gradient descent in formula (26) [40].”
BackgroundAbdellatif, A. A., Mhaisen, N., Mohamed, A. M. S., Erbad, A., & Guizani, M. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3288050 · Cited by 91
“For instance, [34] and [35] review the usage of DRL algorithms for computer vision and natural language processing.”
BackgroundHoxha, J., Çodur, M. Y., Mustafaraj, E., Kanj, H., & Masri, A. E. (2023) · Applied Energy · doi:10.1016/j.apenergy.2023.121765 · Cited by 80
Shalumov, V. (2020) · Aerospace Science and Technology · doi:10.1016/j.ast.2020.105996 · Cited by 82
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.