Human-level control through deep reinforcement learning
The DOI matches its Crossref record; 38 of the 38 references it cites were found in a registry or an index.
Abstract
The theory of reinforcement learning provides a normative account, deeply rooted in psychological and neuroscientific perspectives on animal behaviour, of how agents may optimize their control of an environment. To use reinforcement learning successfully in situations approaching real-world complexity, however, agents are confronted with a difficult task: they must derive efficient representations of the environment from high-dimensional sensory inputs, and use these to generalize past experience to new situations. Remarkably, humans and other animals seem to solve this problem through a harmonious combination of reinforcement learning and hierarchical sensory processing systems, the former evidenced by a wealth of neural data revealing notable parallels between the phasic signals emitted by dopaminergic neurons and temporal difference reinforcement learning algorithms. While reinforcement learning agents have achieved some successes in a variety of domains, their applicability has previously been limited to domains in which useful features can be handcrafted, or to domains with fully observed, low-dimensional state spaces. Here we use recent advances in training deep neural networks to develop a novel artificial agent, termed a deep Q-network, that can learn successful policies directly from high-dimensional sensory inputs using end-to-end reinforcement learning. We tested this agent on the challenging domain of classic Atari 2600 games. We demonstrate that the deep Q-networ…
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
“In previous studies, the agents are assumed to act based on simple and heuristic policy functions, and learned in isolated and restricted environments [1][2][3][4][5][6].”
From A survey on large language model based autonomous agents (2024) · cites this paper as [1]
In the citing paper
“This technique was developed in 2013 with Google Deep Mind [55].”
From Review of deep learning: concepts, CNN architectures, challenges, applications, future directions (2021) · cites this paper as [55]
In the citing paper
“Whereas previous reinforcement-learning algorithms often rely only on task-specific learning signals9,10, Dreamer rests predominantly on the unsupervised objective of its world model.”
From Mastering diverse control tasks through world models (2025) · cites this paper as [9]
In the citing paper
“Traditional RL such as AlphaGo Zero and the DQN series (Silver et al., 2017; Mnih et al., 2015; Yue et al., 2023) can continuously improve the performance of a policy in environments like Go and Atari games without an explicit upper bound .”
From Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model? (2025) · cites this paper as Mnih et al., 2015
In the citing paper
“Unlike previous reinforcement learning algorithms that often rely only on task-specific learning signals 7,8 , Dreamer rests predominantly on the unsupervised objective of its world model.”
From Mastering Diverse Domains through World Models (2023) · cites this paper as [7]
24 more claims about this paper
Create a free account to see every claim other papers make about this one.
Citations
Cites
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155
Gradient-based learning applied to document recognition
Lecun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998) · Proceedings of the IEEE · doi:10.1109/5.726791 · Cited by 58,805 · cited 1×
“We use one particularly successful architecture, the deep convolutional network 17 , which uses hierarchical layers of tiled convolutional filters tomimictheeffectsofreceptivefields—inspiredbyHubelandWiesel’s seminalworkonfeedforwardprocessinginearlyvisualcortex 18 —thereby…”
Method(2008) · Journal of Machine Learning Research · Cited by 33,051
Reinforcement 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
Reinforcement Learning: An Introduction
(2005) · IEEE Transactions on Neural Networks · doi:10.1109/tnn.2004.842673 · Cited by 24,496
Reducing the dimensionality of data with neural networks
Hinton, G. E., & Salakhutdinov, R. R. (2006) · Science · doi:10.1126/science.1127647 · Cited by 20,928
(2010) · International Conference on Machine Learning · Cited by 13,995
A neural substrate of prediction and reward
Schultz, W., Dayan, P., & Montague, P. R. (1997) · Science · doi:10.1126/science.275.5306.1593 · Cited by 9,388 · cited 1×
“…4,5 , the former evidenced by a wealth of neural data revealingnotableparallelsbetweenthephasicsignalsemittedbydopa-minergicneuronsandtemporaldifferencereinforcementlearning algorithms 3 .”
BackgroundLearning Deep Architectures for AI
Bengio, Y. (2009) · Foundations and Trends® in Machine Learning · doi:10.1561/2200000006 · Cited by 9,239 · cited 2×
“Notably, recent advances in deep neural networks 9–11 , in which several layers of nodes areused to buildup progressivelymore abstractrepresentations of the data, have made it possible for artificial neural networks to learn concepts such as object categories directly from raw sensory data.”
BackgroundWatkins, C. J. C. H., & Dayan, P. (1992) · Machine Learning · doi:10.1007/bf00992698 · Cited by 9,083
Cited by
Items 1,301–1,400 of 1,418
Sensing flow gradients is necessary for learning autonomous underwater navigation
Jiao, Y., Hang, H., Merel, J., & Kanso, E. (2025) · Nature Communications · doi:10.1038/s41467-025-58125-6 · Cited by 23
Cyber-physical defense in the quantum Era
Barbeau, M., & Garcia-Alfaro, J. (2022) · Scientific Reports · doi:10.1038/s41598-022-05690-1 · Cited by 22
SAGCI-System: Towards Sample-Efficient, Generalizable, Compositional, and Incremental Robot Learning
Lv, J., Yu, Q., Shao, L., Liu, W., Xu, W., & Lu, C. (2022) · 2022 International Conference on Robotics and Automation (ICRA) · doi:10.1109/icra46639.2022.9811859 · Cited by 22
Learning adversarial attack policies through multi-objective reinforcement learning
García, J., Majadas, R., & Fernández, F. (2020) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2020.104021 · Cited by 22
Machine Learning, Functions and Goals
Butlin, P. (2022) · Croatian journal of philosophy · doi:10.52685/cjp.22.66.5 · Cited by 22
Replay-Guided Adversarial Environment Design
Jiang, M., Dennis, M., Parker-Holder, J., Foerster, J., Grefenstette, E., & Rocktäschel, T. (2021) · arXiv · doi:10.48550/arxiv.2110.02439 · Cited by 21
“While deep reinforcement learning (RL) approaches have led to many successful applications in challenging domains like Atari [21], Go [35], Chess [36], Dota [4], and StarCraft [40] in recent years, deep RL agents still prove to be brittle, often failing to transfer to environments only slightly different from those en…”
BackgroundADAPT: Zero-Shot Adaptive Policy Transfer for Stochastic Dynamical Systems
Harrison, J., Garg, A., Ivanovic, B., Zhu, Y., Savarese, S., Fei-Fei, L., & Pavone, M. (2017) · arXiv · doi:10.48550/arxiv.1707.04674 · Cited by 21
(2025) · doi:10.1038/s41467-025-58439-5 · Cited by 21
“The DQN algorithm [42], which is widely used in reinforcement learning methods, is employed to train the policy of RL-GenRisk.”
Learning to navigate cloth using haptics
Clegg, A., Yu, W., Erickson, Z., Tan, J., Liu, C. K., & Turk, G. (2017) · 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2017.8206110 · Cited by 20
“Recent advances in Reinforcement Learning have enabled the training of complex robotic motor skills with highdimensional continuous state and action spaces [19]–[23] and human-expert-level agents in playing Atari games [24] and Go [25].”
BackgroundTrajectory-wise Multiple Choice Learning for Dynamics Generalization in Reinforcement Learning
Seo, Y., Lee, K., Clavera, I., Kurutach, T., Shin, J., & Abbeel, P. (2020) · arXiv · doi:10.48550/arxiv.2010.13303 · Cited by 20
“Early successful deep RL approaches had been mostly model-free, which do not require an explicit model of the environment, but instead directly learn a policy [25, 28, 38].”
BackgroundLiveBench: A Challenging, Contamination-Limited LLM Benchmark
White, C., Dooley, S., Roberts, M., Pal, A., Feuer, B., Jain, S., Shwartz-Ziv, R., Jain, N., Saifullah, K., Dey, S., et al. (2024) · arXiv · doi:10.48550/arxiv.2406.19314 · Cited by 20
Count-based exploration with the successor representation
Machado, M. C., Bellemare, M. G., & Bowling, M. (2020) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v34i04.5955 · Cited by 19
Self-citationKadi, H. A., & Terzić, K. (2023) · Sensors · doi:10.3390/s23052389 · Cited by 27
“Deep Q-learning (DQN) [126] extends Q-learning to solve high-dimensional/continuous state problems in practice, where actions are still discrete.”
BackgroundMethodMultimodal representation learning for place recognition using deep Hebbian predictive coding
Pearson, M. J., Dora, S., Struckmeier, O., Knowles, T. C., Mitchinson, B., Tiwari, K., Kyrki, V., Bohte, S., & Pennartz, C. M. A. (2021) · Frontiers in Robotics and AI · doi:10.3389/frobt.2021.732023 · Cited by 19
Few-Shot Class-Incremental Learning with Meta-Learned Class Structures
Zheng, G., & Zhang, A. (2021) · 2021 International Conference on Data Mining Workshops (ICDMW) · doi:10.1109/icdmw53433.2021.00058 · Cited by 19
Wait, But Why?: Assessing Behavior Explanation Strategies for Real-Time Strategy Games
Robertson, J., Kokkinakis, A. V., Hook, J., Kirman, B., Block, F., Ursu, M. F., Patra, S., Demediuk, S., Drachen, A., & Olarewaju, O. (2021) · 26th International Conference on Intelligent User Interfaces · doi:10.1145/3397481.3450699 · Cited by 18
“Deep neural models have learned to play at human or better levels in classic Atari games [26] without hard-coded domain knowledge or supervised training.”
BackgroundNew Automation for Social Bots: From Trivial Behavior to AI-Powered Communication
Grimme, C., Pohl, J., Cresci, S., Lüling, R., & Preuss, M. (2022) · Lecture Notes in Computer Science · doi:10.1007/978-3-031-18253-2_6 · Cited by 19
Zheng, J., Cao, H., Chen, D., Ansari, R., Chu, K.-C., & Huang, M.-C. (2020) · IEEE Sensors Journal · doi:10.1109/jsen.2020.2986768 · Cited by 17
“1) Deep Deterministic Policy Gradient (DDPG): The deep deterministic policy gradient (DDPG) [22] algorithm combines the advances in Deep Q Network (DQN) algorithm [21] and deterministic policy gradient (DPG) algorithm [28].”
BackgroundMethodInfluentialExplainable AI-Aided Feature Selection and Model Reduction for DRL-Based V2X Resource Allocation
Khan, N., Abdallah, A., Celik, A., Eltawil, A. M., & Coleri, S. (2025) · IEEE Transactions on Communications · doi:10.1109/tcomm.2025.3554655 · Cited by 20
Vehicular Edge Computing via Deep Reinforcement Learning
Qi, Q., & Ma, Z. (2019) · arXiv · doi:10.48550/arxiv.1901.04290 · Cited by 17
“The DQN (Deep Q-Network) model firstly proposed by DeepMind combines CNN and Q-learning, trains the CNN by the rewards obtained by the Q-learning in each epoch and realizes the directly control from input states to the output policy [18].”
MethodApplications of Deep Reinforcement Learning in Communications and Networking: A Survey
Luong, N. C., Hoang, D. T., Gong, S., Niyato, D., Wang, P., Liang, Y.-C., & Kim, D. I. (2018) · arXiv · doi:10.48550/arxiv.1810.07862 · Cited by 17
Recurrent Off-policy Baselines for Memory-based Continuous Control
Yang, Z., & Nguyen, H. (2021) · arXiv · doi:10.48550/arxiv.2110.12628 · Cited by 17
Benchmarking the Spectrum of Agent Capabilities
Hafner, D. (2021) · arXiv · doi:10.48550/arxiv.2109.06780 · Cited by 16
“It comprises around 54 individual games, depending on the evaluation protocol (Mnih et al., 2015; Schulman et al., 2017; Badia et al., 2020; Hafner et al., 2020).”
BackgroundWei, J., Qi, S., Wang, W., Jiang, L., Gao, H., Zhao, F., Al-Bukhaiti, K., & Wan, A. (2025) · Contemporary Mathematics · doi:10.37256/cm.6220256459 · Cited by 15
“…learning (RL) networks, which conglomerate neural networks with RL algorithms, enable systems to learn over environmental interactions, making them particularly effective for decision-making chores involving tardy recompenses, such as those in gaming, robotics, and autonomous systems [61].”
The Difficulty of Passive Learning in Deep Reinforcement Learning
Ostrovski, G., Castro, P. S., & Dabney, W. (2021) · arXiv · doi:10.48550/arxiv.2110.14020 · Cited by 15
“QR-DQN indeed appears to be a nontrivially different passive learning algorithm, significantly better in some games, while curiously worse in others (Fig.”
MethodInfluentialSelf-citation(2020) · doi:10.1145/3383668.3419905 · Cited by 15
“This is constantly demonstrated by successful integration examples, where – in the reverse direction – scientific research on AI for games frequently builds on industrial games, as in game playing [1, 17, 35], automated testing [22, 26, 34] or balancing [9, 20], world-building [27, 36], dynamic difficulty adjustment (…”
BackgroundBarth, A., & Ma, O. (2024) · Acta Astronautica · doi:10.1016/j.actaastro.2023.11.014 · Cited by 15
An Instruction Set Architecture for Machine Learning
Chen, Y., Lan, H., Du, Z., Liu, S., Tao, J., Han, D., Luo, T., Guo, Q., Li, L., Xie, Y., et al. (2018) · ACM Transactions on Computer Systems · doi:10.1145/3331469 · Cited by 14
“…by deep learning [30, 34, 36, 37] have become the state-of-the-art across a broad range of applications (such as pattern recognition [12] and web search [26]), some have even achieved human-level performance on specific tasks such as ImageNet recognition [23] and Atari 2600 video games [45].”
BackgroundDeep reinforcement learning for task-based feature learning in prosthetic vision
White, J., Kameneva, T., & McCarthy, C. (2019) · 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc.2019.8856541 · Cited by 14
In-context Reinforcement Learning with Algorithm Distillation
Laskin, M., Wang, L., Oh, J., Parisotto, E., Spencer, S., Steigerwald, R., Strouse, D., Hansen, S., Filos, A., Brooks, E., et al. (2022) · arXiv · doi:10.48550/arxiv.2210.14215 · Cited by 14
“To date, deep RL research has mostly focused on powerful single-task agents Mnih et al. (2015); Schrittwieser et al. (2019); Hessel et al. (2021).”
BackgroundSelf-citationEvolution of Brains and Computers: The Roads Not Taken
Solé, R., & Seoane, L. F. (2022) · Entropy · doi:10.3390/e24050665 · Cited by 14
A New Anti-Jamming Strategy Based on Deep Reinforcement Learning for MANET
Xu, Y., Lei, M., Li, M., Zhao, M., & Hu, B. (2019) · 2019 IEEE 89th Vehicular Technology Conference (VTC2019-Spring) · doi:10.1109/vtcspring.2019.8746494 · Cited by 14
“2) Deep Q-Network: Deep Q-network, developed by DeepMind [11], is a recent breakthrough for model-free control which enables computers to enjoy wonderful performance on many Atari video games as human level [12].”
BackgroundSemi-automated level design via auto-playtesting for handheld casual game creation
Powley, E. J., Colton, S., Gaudl, S., Saunders, R., & Nelson, M. J. (2016) · 2016 IEEE Conference on Computational Intelligence and Games (CIG) · doi:10.1109/cig.2016.7860438 · Cited by 14
“Reinforcement learning methods such as deep Q learning have also been demonstrated to work well for general videogame playing [16].”
MethodGoing Beyond Linear Transformers with Recurrent Fast Weight Programmers
Irie, K., Schlag, I., Csordás, R., & Schmidhuber, J. (2021) · arXiv · doi:10.48550/arxiv.2106.06295 · Cited by 14
“[12]’s work, which presents multi-task Atari as a side experiment, we study the Transformer family of models on the standard Atari 2600 setting [31, 63, 64] by training game-specific agents.”
MethodSim-to-Real in Reinforcement Learning for Everyone
Vacaro, J., Marques, G., Oliveira, B., Paz, G., Paula, T., Staehler, W., & Murphy, D. (2019) · 2019 Latin American Robotics Symposium (LARS), 2019 Brazilian Symposium on Robotics (SBR) and 2019 Workshop on Robotics in Education (WRE) · doi:10.1109/lars-sbr-wre48964.2019.00060 · Cited by 13
“Deep reinforcement learning (DRL) achieved remarkable results in recent years, such as learning to play Atari games directly from raw pixels [2].”
BackgroundPAC Reinforcement Learning with Rich Observations
Krishnamurthy, A., Agarwal, A., & Langford, J. (2016) · arXiv · doi:10.48550/arxiv.1602.02722 · Cited by 13
“The Atari Reinforcement Learning research program (Mnih et al., 2015) has highlighted a critical deficiency of reinforcement learning algorithms: they cannot effectively solve problems that require systematic exploration.”
(2023) · doi:10.3390/s23218924 · Cited by 13
Koster, R., Pîslar, M., Tacchetti, A., Balaguer, J., Liu, L., Elie, R., Hauser, O. P., Tuyls, K., Botvinick, M., & Summerfield, C. (2025) · Nature Communications · doi:10.1038/s41467-025-58043-7 · Cited by 12
Zhou, Q., Li, H., & Wang, J. (2020) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v34i04.6177 · Cited by 12
Chen, Q., & Heydari, B. (2022) · Journal of Mechanical Design · doi:10.1115/1.4055057 · Cited by 12
Learning Interactive Real-World Simulators
Yang, S., Du, Y., Ghasemipour, K., Tompson, J., Kaelbling, L., Schuurmans, D., & Abbeel, P. (2023) · arXiv · doi:10.48550/arxiv.2310.06114 · Cited by 346
“Reinforcement learning (RL) has achieved superhuman performance on difficult tasks such as playing Go and Atari games [35, 36], but has limited real world applications due, among other reasons, to the lack of a realistic environment simulator [37].”
BackgroundNguyen, C. H., Shihua, J. M., & Liem, R. P. (2025) · Communications in Transportation Research · doi:10.1016/j.commtr.2025.100165 · Cited by 12
Synthetic Returns for Long-Term Credit Assignment
Raposo, D., Ritter, S., Santoro, A., Wayne, G., Weber, T., Botvinick, M., van Hasselt, H., & Song, F. (2021) · arXiv · doi:10.48550/arxiv.2102.12425 · Cited by 12
“…required to obtain that performance, measured by the total number of environment steps, for each agent’s best seed reported in the literature (Mnih et al., 2015; Hessel et al., 2018; Horgan et al., 2018; Espeholt et al., 2018; Kapturowski et al., 2018; Puigdom`enech Badia et al., 2020a;b).…”
BackgroundMethodBYOL-Explore: Exploration by Bootstrapped Prediction
Guo, Z. D., Thakoor, S., Pîslar, M., Pires, B. A., Altché, F., Tallec, C., Saade, A., Calandriello, D., Grill, J.-B., Tang, Y., et al. (2022) · arXiv · doi:10.48550/arxiv.2206.08332 · Cited by 12
“We follow the classical 30 random no-ops evaluation regime [46, 73], and average performance over 10 episodes and over 3 seeds.”
MethodAgent based dynamic thresholding for adaptive anomaly detection using reinforcement learning
Yang, X., Howley, E., & Schukat, M. (2025) · Neural Computing and Applications · doi:10.1007/s00521-024-10536-0 · Cited by 11
“Specifically, we apply the popular DRL algorithm DQN [25] to solve the MDP given its efficiency in the anomaly detection task.”
Combining MCTS and A3C for Prediction of Spatially Spreading Processes in Forest Wildfire Settings
Ganapathi Subramanian, S., & Crowley, M. (2018) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-89656-4_28 · Cited by 11
Robust Domain Randomised Reinforcement Learning through Peer-to-Peer Distillation
Zhao, C., & Hospedales, T. (2020) · arXiv · doi:10.48550/arxiv.2012.04839 · Cited by 11
“Deep reinforcement learning (RL) has been successfully applied to various tasks, including Go [22], Atari [13] and robot control tasks [21], etc.”
Rudroff, T., Rainio, O., & Klén, R. (2024) · Brain Sciences · doi:10.3390/brainsci14111111 · Cited by 11
Data Assimilation in Chaotic Systems Using Deep Reinforcement Learning
Hammoud, M. A. E. R., Raboudi, N., Titi, E. S., Knio, O., & Hoteit, I. (2024) · Journal of Advances in Modeling Earth Systems · doi:10.1029/2023ms004178 · Cited by 11
“For instance, convolutional, recurrent and graph neural networks would allow for significant dimensionality reduction, and are commonly adopted for RL tasks (e.g., Jumper et al., 2021; Mnih et al., 2015).”
Avoiding Side Effects in Complex Environments
Turner, A. M., Ratzlaff, N., & Tadepalli, P. (2020) · arXiv · doi:10.48550/arxiv.2006.06547 · Cited by 11
“[13]’s DQN, trained on the primary SafeLife reward function R without a side effect penalty.”
MethodLearning of Motor Control from Motor Babbling**This research is supported by CREST, JST.
Aoki, T., Nakamura, T., & Nagai, T. (2016) · IFAC-PapersOnLine · doi:10.1016/j.ifacol.2016.10.478 · Cited by 10
Efficient collective swimming by harnessing vortices through deep reinforcement learning
Verma, S., Novati, G., & Koumoutsakos, P. (2018) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1800923115 · Cited by 10
“Here, we use deep reinforcement learning [deep RL (23)] to discover such strategies for two autonomous and self-propelled swimmers and elucidate the physical mechanisms that enable efficient and sustained coordinated swimming.”
MethodMeasuring Visual Generalization in Continuous Control from Pixels
Grigsby, J., & Qi, Y. (2020) · arXiv · doi:10.48550/arxiv.2010.06740 · Cited by 10
“…that are trained separately, and helps to reduce overestimation bias (Fujimoto et al., 2018). φ (cid:48) refers to target networks, which are moving averages of the critic networks’ weights and help to stabilize learning - a common trick in Deep RL (Lillicrap et al., 2015) (Mnih et al., 2015).”
BackgroundMacro, Finance, and Macro Finance: Solving Nonlinear Models in Continuous Time with Machine Learning
Duarte, V. (2017) · SSRN Electronic Journal · doi:10.2139/ssrn.3012602 · Cited by 9
Scaling Instructable Agents Across Many Simulated Worlds
SIMA Team, Raad, M. A., Ahuja, A., Barros, C., Besse, F., Bolt, A., Bolton, A., Brownfield, B., Buttimore, G., Cant, M., et al. (2024) · arXiv · doi:10.48550/arxiv.2404.10179 · Cited by 9
“There is a long history of research in creating agents that can interact with video games or simulated 3D environments (e.g., Mnih et al., 2015; Berner et al., 2019; Vinyals et al., 2019; Baker et al., 2022) and even follow language instructions in a limited range of environments (e.g., Abramson et…”
Self-citationA multi-agent deep reinforcement learning based energy management for behind-the-meter resources
Wilk, P., Wang, N., & Li, J. (2022) · The Electricity Journal · doi:10.1016/j.tej.2022.107129 · Cited by 9
“The recent advancement of deep learning has enabled deep reinforcement learning (DRL) (Bengio et al., 2013; Arulkumaran et al., 2017) to drive optimal policies for sophisticated and capable agents, which can outperform conventional rule-based operation policies in applications such as games (e.g., Go, WarCraft) (Mnih…”
Deep reinforcement learning for multi-agent coordination
Aina, K. O., & Ha, S. (2026) · Artificial Life and Robotics · doi:10.1007/s10015-025-01089-z · Cited by 8
“The recent advancement of deep reinforcement learning (DRL) in tackling real-world tasks has established it as a powerful tool for solving complex problems characterized by high-dimensional information representation through trial-and-error methods [11].”
Phenomenal interface theory: a model for basal consciousness
Klein, C., & Barron, A. B. (2025) · Philosophical Transactions of the Royal Society B: Biological Sciences · doi:10.1098/rstb.2024.0301 · Cited by 8
“On the one hand, there is a close (and non-accidental) relationship between what a PI does and various forms of so-called deep reinforcement learning [87,88].”
A knowledge-driven framework for Robotic Odor Source Localization using large language models
Mahmud, K. R., Wang, L., Hassan, S., & Zhang, Z. (2025) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2025.104915 · Cited by 10
Podracer architectures for scalable Reinforcement Learning
Hessel, M., Kroiss, M., Clark, A., Kemaev, I., Quan, J., Keck, T., Viola, F., & van Hasselt, H. (2021) · arXiv · doi:10.48550/arxiv.2104.06272 · Cited by 8
“Indeed a steady increase can be observed in the workloads associated to popular RL agents (Hessel et al., 2018; Mnih et al., 2015, 2016).”
BackgroundPlaytesting: What is Beyond Personas
Ariyurek, S., Surer, E., & Betin-Can, A. (2023) · IEEE Transactions on Games · doi:10.1109/tg.2022.3165882 · Cited by 7
“Researchers measured the uncertainty using bootstrapped DQN [24], statespace features [14], and error of a neural function [25].”
MethodLoss of Plasticity in Continual Deep Reinforcement Learning
Abbas, Z., Zhao, R., Modayil, J., White, A., & Machado, M. C. (2023) · arXiv · doi:10.48550/arxiv.2303.07507 · Cited by 10
Stratospheric Aerosol Injection as a Deep Reinforcement Learning Problem
de Witt, C. S., & Hornigold, T. (2019) · arXiv · doi:10.48550/arxiv.1905.07366 · Cited by 7
“To overcome the inherent sample inefficiency of DRL, we propose to emulate a Global Circulation Model (GCM) using deep learning techniques.”
BackgroundMethodInfluentialUnderstanding Learned Reward Functions
Michaud, E. J., Gleave, A., & Russell, S. (2020) · arXiv · doi:10.48550/arxiv.2012.05862 · Cited by 7
Interactive Perception for Deformable Object Manipulation
Weng, Z., Zhou, P., Yin, H., Kravberg, A., Varava, A., Navarro-Alarcon, D., & Kragic, D. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3431943 · Cited by 7
Elements of episodic memory: insights from artificial agents
Boyle, A., & Blomkvist, A. (2024) · Philosophical Transactions of the Royal Society B: Biological Sciences · doi:10.1098/rstb.2023.0416 · Cited by 6
Colas, J. T., O’Doherty, J. P., & Grafton, S. T. (2024) · PLOS Computational Biology · doi:10.1371/journal.pcbi.1011950 · Cited by 6
Measuring Progress in Deep Reinforcement Learning Sample Efficiency
Dorner, F. E. (2021) · arXiv · doi:10.48550/arxiv.2102.04881 · Cited by 6
“While the effect is of similar magnitude for different algorithms it is second strongest for DQN, the earliest algorithm we considered, which might inflate perceived progress by making the baseline result worse.”
BackgroundMethodInfluentialShigeno, S. (2017) · Diversity and Commonality in Animals · doi:10.1007/978-4-431-56469-0_19 · Cited by 6
Lahariya, M., Sadeghianpourhamami, N., & Develder, C. (2022) · SSRN Electronic Journal · doi:10.2139/ssrn.4090508 · Cited by 6
Using Unity to Help Solve Intelligence
Ward, T., Bolt, A., Hemmings, N., Carter, S., Sanchez, M., Barreira, R., Noury, S., Anderson, K., Lemmon, J., Coe, J., et al. (2020) · arXiv · doi:10.48550/arxiv.2011.09294 · Cited by 6
“In the past decade, reinforcement learning has made significant advances in the pursuit of achieving general purpose intelligence (e.g. Espeholt et al., 2018; Mnih et al., 2015; Silver et al., 2016).”
BackgroundWang, B., Sun, T., & Zheng, X. S. (2019) · Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment · doi:10.1609/aiide.v15i1.5244 · Cited by 6
The Sos Conductor: Orchestrating Resources With Iterative Agent-Based Reinforcement Learning
Chen, Q., & Heydari, B. (2024) · Systems Engineering · doi:10.1002/sys.21747 · Cited by 5
“Further details can be found in ref.”
MethodInfluentialEmergence of Collaborative Hunting via Multi-Agent Deep Reinforcement Learning
Tsutsui, K., Takeda, K., & Fujii, K. (2023) · Lecture Notes in Computer Science · doi:10.1007/978-3-031-37660-3_15 · Cited by 5
Artificial Intelligence for Dynamic Spectrum Management
Liang, Y.-C. (2020) · Signals and Communication Technology · doi:10.1007/978-981-15-0776-2_6 · Cited by 5
“Deep Q-network : Different from the tabular method in traditional RL, a neural network called deep Q-network (DQN) is adopted to approximate the relationship between state space and action space [12].”
MethodRen, T., & Zeng, X.-J. (2024) · Proceedings of the Thirty-ThirdInternational Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2024/22 · Cited by 8
“In this study, we construct a computational model using deep Q-learning (DQN) [Mnih et al. , 2015] to explore how agents can simultaneously learn both interaction and dilemma strategies from their long-term experiences.”
MethodDeep reinforcement learning enabling a BCFbot to learn various undulatory patterns
Hameed, I., Chao, X., Navarro-Alarcon, D., & Jing, X. (2025) · Ocean Engineering · doi:10.1016/j.oceaneng.2025.120322 · Cited by 5
Tian, Z., He, Y., Sun, L., Li, Y., & Du, S. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.128184 · Cited by 5
Thompson sampling via local uncertainty
Wang, Z., & Zhou, M. (2019) · arXiv · doi:10.48550/arxiv.1910.13673 · Cited by 5
“…recent interest in employing deep neural networks to better solve sequential decision-making problems, such as these in reinforcement learning (Mnih et al., 2013, 2015, 2016; Arulkumaran et al., 2017; Fran¸cois-Lavet et al., 2018) and contextual bandits (Riquelme et al., 2018; Russo et al.,…”
BackgroundOn Value Functions and the Agent-Environment Boundary
Jiang, N. (2019) · arXiv · doi:10.48550/arxiv.1905.13341 · Cited by 5
“Since the original game screen has a high resolution, it is common in practice to downsample the screen as a pre-processing step [Mnih et al., 2015].”
BackgroundMethodGenerative Network Design for Multi-Agent Coordination: A VAE-Assisted Deep RL Approach
Chen, Q., & Heydari, B. (2025) · Journal of Mechanical Design · doi:10.1115/1.4068796 · Cited by 4
Berrevoets, J., Kacprzyk, K., Qian, Z., & van der Schaar, M. (2023) · arXiv · doi:10.48550/arxiv.2303.02186 · Cited by 8
“While research on bandits seem to provide algorithms which are agnostic to estimation, RL seems to stem more from a deep learning community; hence most papers in RL rely on neural networks to approximate the reward (based on, for example, Mnih et al. [173]).”
BackgroundValue-Consistent Representation Learning for Data-Efficient Reinforcement Learning
Yue, Y., Kang, B., Xu, Z., Huang, G., & Yan, S. (2022) · arXiv · doi:10.48550/arxiv.2206.12542 · Cited by 4
“Later, Data-Efficient Rainbow (van Hasselt, Hessel, and Aslanides 2019) and OTRainbow (Kielak 2020) show that Rainbow DQN with hyperparameter tuning can be a strong baseline for the low-data regime by simply increasing the number of steps in multi-step return and allowing more frequent parameter update.”
BackgroundInfluentialEco-evolutionary Dynamics of Non-episodic Neuroevolution in Large Multi-agent Environments
Hamon, G., Nisioti, E., & Moulin-Frier, C. (2023) · Proceedings of the Companion Conference on Genetic and Evolutionary Computation · doi:10.1145/3583133.3590703 · Cited by 4
“representation learning and reinforcement learning in high-dimensional spaces [20].”
BackgroundDoes Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?
Yue, Y., Chen, Z., Lu, R., Zhao, A., Wang, Z., Yue, Y., Song, S., & Huang, G. (2025) · arXiv · doi:10.48550/arxiv.2504.13837 · Cited by 804
Towards Practical Credit Assignment for Deep Reinforcement Learning
Alipov, V., Simmons-Edler, R., Putintsev, N., Kalinin, P., & Vetrov, D. (2021) · arXiv · doi:10.48550/arxiv.2106.04499 · Cited by 3
“We use the AtariCNN architecture first proposed by DQN [10] for the agent’s policy πθ(a|s), with an additional value head for V πθ .”
MethodDeep Learning for Solving Loading, Packing, Routing, and Scheduling Problems
Elkefi, S., & Layeb, S. B. (2023) · Handbook of Formal Optimization · doi:10.1007/978-981-19-8851-6_16-1 · Cited by 3
Reinforcement Learning with Trajectory Feedback
Efroni, Y., Merlis, N., & Mannor, S. (2020) · arXiv · doi:10.48550/arxiv.2008.06036 · Cited by 3
A General Framework on Enhancing Portfolio Management with Reinforcement Learning
Li, Y., Wang, J., & Cao, Y. (2019) · arXiv · doi:10.48550/arxiv.1911.11880 · Cited by 3
Untying cable by combining 3D deep neural network with deep reinforcement learning
Fan, Z., Shao, W., Hayashi, T., & Ohashi, T. (2023) · Advanced Robotics · doi:10.1080/01691864.2022.2126729 · Cited by 2
Generalization through Diversity: Improving Unsupervised Environment Design
Li, W., Varakantham, P., & Li, D. (2023) · arXiv · Cited by 3
Li, G., Zhang, G., & Qin, C. (2021) · Sensors · doi:10.3390/s21186100 · Cited by 2
“Techniques such as Double DQN (DDQN) [19] and Dueling DQN [20] have been developed to improve the performance of the DQN algorithm.”
MethodInfluentialAlphaMut: A Deep Reinforcement Learning Model to Suggest Helix-Disrupting Mutations
Bhargav, P., & Mukherjee, A. (2025) · Journal of Chemical Theory and Computation · doi:10.1021/acs.jctc.4c01387 · Cited by 2
The Principle of Unchanged Optimality in Reinforcement Learning Generalization
Irpan, A., & Song, X. (2019) · arXiv · doi:10.48550/arxiv.1906.00336 · Cited by 2
A Unified Theory of Dual-Process Control
Moskovitz, T., Miller, K., Sahani, M., & Botvinick, M. M. (2022) · arXiv · doi:10.48550/arxiv.2211.07036 · Cited by 2
DeKraker, J., Eichert, N., Hong, S.-J., Evans, A. C., & Bernhardt, B. C. (2026) · Neuroscience & Biobehavioral Reviews · doi:10.1016/j.neubiorev.2026.106890 · Cited by 1
Generalization through Diversity: Improving Unsupervised Environment Design
Li, W., Varakantham, P., & Li, D. (2023) · Proceedings of the Thirty-Second International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2023/601 · Cited by 3
Player-Guided AI outperforms standard AI in Sequence Alignment Puzzles
Mutalova, R., Sarrazin-Gendron, R., Ghasemloo Gheidari, P., Cai, E., Richard, G., Caisse, S., Knight, R., Blanchette, M., Szantner, A., & Waldispühl, J. (2023) · Proceedings of The ACM Collective Intelligence Conference · doi:10.1145/3582269.3615597 · Cited by 1
Kinetix: Investigating the Training of General Agents through Open-Ended Physics-Based Control Tasks
Matthews, M., Beukman, M., Lu, C., & Foerster, J. (2024) · arXiv · doi:10.48550/arxiv.2410.23208 · Cited by 1
“Through massive parallelisation and elimination of CPU-GPU transfer, this gives tremendous speed benefits (Lu et al., 2022).”
InfluentialComparing Task Simplifications to Learn Closed-Loop Object Picking Using Deep Reinforcement Learning
Breyer, M., Furrer, F., Novkovic, T., Siegwart, R., & Nieto, J. (2018) · arXiv · doi:10.48550/arxiv.1803.04996 · Cited by 1
Metadata
- Authors
- Mnih, V., Kavukcuoglu, K., Silver, D., Rusu, A. A., Veness, J., Bellemare, M. G., Graves, A., Riedmiller, M., Fidjeland, A. K., Ostrovski, G., Petersen, S., Beattie, C., Sadik, A., Antonoglou, I., King, H., Kumaran, D., Wierstra, D., Legg, S., & Hassabis, D.
- Affiliations
- Google DeepMind (United Kingdom) (GB)
- Corresponding author
- Kavukcuoglu, Koray
- Hassabis, Demis
- ORCID
- Venue
- Nature
- Published
- 2015-02-25
- Volume
- 518
- Issue
- 7540
- Pages
- 529-533
- ISSN
- 0028-0836 (ISSN-L)
- 1476-4687
- OpenAlex
- W2145339207
- Semantic Scholar
- 340f48901f72278f6bf78a04ee5b01df208cc508
- Cited by
- 30,357
- References
- 37
- Citations by year
- 2026: 2851 · 2025: 3619 · 2024: 3335 · 2023: 3592 · 2022: 3367 · 2021: 3971 · 2020: 3466 · 2019: 2907 · 2018: 1776 · 2017: 912 · 2016: 416 · 2015: 132
- FWCI
- 1057.57
- Type
- journal article
- Language
- en
- Open access
- closed
- Licence
- www.springer.com/tdm (tdm)
- www.springer.com/tdm (vor)
- Locations
- Nature (publishedVersion)
- PubMed (publishedVersion)
- RePEc: Research Papers in Economics (submittedVersion)
- UCL Discovery (University College London) (submittedVersion)
- Funders
- DeepMind
- 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 Q-network, deep reinforcement learning, game-playing agents, end-to-end learning, deep neural network, temporal difference learning, dopaminergic neurons, human-level performance, DeepMind, AlphaGo, AlphaZero
- MeSH terms
- Algorithms, Artificial Intelligence, Humans, Models, Psychological, Reinforcement, Psychology, Reward, Neural Networks, Computer, Video Games
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.