A general reinforcement learning algorithm that masters chess, shogi, and Go through self-play
Open accessThe DOI matches its Crossref record; 30 of the 30 references it cites were found in a registry or an index; 5 of the 16 sources its claims cite were checked for support.
Abstract
One program to rule them all: Computers can beat humans at increasingly complex games, including chess and Go. However, these programs are typically constructed for a particular game, exploiting its properties, such as the symmetries of the board on which it is played. Silver et al. developed a program called AlphaZero, which taught itself to play Go, chess, and shogi (a Japanese version of chess) (see the Editorial, and the Perspective by Campbell). AlphaZero managed to beat state-of-the-art programs specializing in these three games. The ability of AlphaZero to adapt to various game rules is a notable step toward achieving a general game-playing system. Science , this issue p. 1140 ; see also pp. 1087 and 1118
Overview
CoChat Summary
TL;DR
AlphaZero is a single reinforcement-learning and MCTS algorithm that, given only game rules, learned superhuman play in chess, shogi, and Go through self-play. It defeated world-champion programs Stockfish, Elmo, and AlphaGo Zero using the same algorithm and network architecture across all three games.
Key findings
- In chess, AlphaZero defeated Stockfish 8, winning 155 games and losing 6 out of 1,000; it also won matches from common human openings and 2016 TCEC opening positions.
- In shogi, AlphaZero defeated Elmo, winning 98.2% of games as black and 91.2% overall; in Go, it defeated AlphaGo Zero, winning 61% of games.
- AlphaZero first outperformed Stockfish after 4 hours (300,000 steps), Elmo after 2 hours (110,000 steps), and AlphaGo Lee after 30 hours (74,000 steps); training ran for 700,000 steps, taking about 9 hours in chess, 12 hours in shogi, and 13 days in Go.
- AlphaZero searched only about 60,000 positions per second in chess and shogi, versus 60 million for Stockfish and 25 million for Elmo, yet still defeated Stockfish with 1/10 thinking time and won 46% against Elmo with 1/100 thinking time.
Methods
- AlphaZero uses a deep neural network \(f_\theta(s)\) that outputs move probabilities \(p\) and a scalar value \(v\), trained by reinforcement learning from self-play starting from random parameters; MCTS uses these outputs to guide search, with 800 simulations per move during training.
- The same algorithm, network architecture, and hyperparameters were used for chess, shogi, and Go, with no domain knowledge except game rules; no opening book, endgame tablebases, or domain-specific heuristics were used.
- Training used 5,000 first-generation TPUs to generate self-play games and 16 second-generation TPUs to train networks; evaluation matches used 3 hours per game plus 15 seconds per move, with opponents on 44 CPU cores and AlphaZero on four first-generation TPUs and 44 CPU cores.
Limitations
- Stated: AlphaZero did not use opening books, endgame tablebases, or domain-specific heuristics; the authors note that some such techniques could further improve performance but were left for future research.
- Stated: hyperparameters were reused across games without game-specific tuning, except for exploration noise and the learning rate schedule; matches against an opening book were low in diversity, and forced randomization increased losses.
- Evident from the text: results cover only three games with separate trained instances, no cross-game transfer is demonstrated, and evaluation compares AlphaZero on TPU hardware against CPU-based opponents, so hardware and time-control differences qualify direct strength comparisons.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
3 Valid · 2 Not valid · 8 Not checked · 43 Own results and interpretations
Every source the paper cites for this claim was checked and supports it.
In the paper
“Apart from the representation of positions and actions described above, AlphaZero uses the same network architecture as AlphaGo Zero ( 9 ), briefly recapitulated here.”
What the sources say
- Supportedcites [9] source
The Neural Network Architecture Comparison (p. 29) identifies AlphaGo Zero’s architecture as a residual tower with policy and value heads. The citation supplies the AlphaGo Zero architecture that the citing authors say AlphaZero uses; the comparison to AlphaZero is their own statement. (LLM, conf=0.88)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“and in Go, AlphaZero first outperformed AlphaGo Lee ( 9 ) after 30 hours (74,000 steps).”
What the sources say
- Partly supportedcites [9] sourceExperimental — accuracy not yet measured
Section 2, p. 7 reports that AlphaGo Zero outperformed AlphaGo Lee, but says this happened after 36 hours, not 30. The passage does not give 74,000 steps, and it names the system AlphaGo Zero rather than AlphaZero. (LLM, conf=0.90)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The training algorithm achieved similar performance in all independent runs (see fig. S3), suggesting that the high performance of AlphaZero's training algorithm is repeatable.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“AlphaZero also defeated Stockfish when given 1 / 10 as much thinking time as its opponent (i.e. searching ~ 1 / 10 , 000 as many positions),”
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“and won 46% of games against Elmo when given 1 / 100 as much time (i.e. searching ~ 1 / 40 , 000 as many positions), see Fig. 2.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
23 more claims
Create a free account to see all 56 claims and to verify the 3 references still open.
- Probably not a substantial claim. 10
- It may state two separate claims as one. 23
- The quote only partly states this claim. 1
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?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Games have historically served as highly effective metrics for evaluating progress in deep reinforcement learning research”
From BALROG: Benchmarking Agentic LLM and VLM Reasoning On Games (2024) · cites this paper as Silver et al., 2018
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“In 2017, AlphaGo, using the same learning algorithms in TD-Gammon and a billion times more computing power, soundly beat Ke Jie, the world Go champion, in a game that is to chess what chess is to backgammon in complexity (Silver et al., 2018).”
From Large Language Models and the Reverse Turing Test (2023) · cites this paper as Silver et al., 2018
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Most of the attention was on who won, but more important was how it won: AlphaGo was creative and had discovered new moves and board positions that were outside human experience.”
From Large Language Models and the Reverse Turing Test (2023) · cites this paper as Silver et al., 2018
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“The primary method for applying more computation at inference, common in the AlphaZero family of agents2,72, is scaling the tree search and running multiple independent search attempts.”
From Olympiad-level formal mathematical reasoning with reinforcement learning (2026) · cites this paper as [2]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“Modern AI can write computer programs (Li et al., 2022) or play chess at super-human level (Silver et al., 2018), but the ability of AI to perceive and act in the world remains far below human level.”
From Scaling Instructable Agents Across Many Simulated Worlds (2024) · cites this paper as Silver et al., 2018
1 more claim about this paper
Create a free account to see every claim other papers make about this one.
Citations
Cites
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
Self-citationAdvances in neural information processing systems 7
(1996) · Computers & Mathematics with Applications · doi:10.1016/s0898-1221(96)90463-0 · Cited by 14,281
Identity mappings in deep residual networks
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-46493-0_38 · Cited by 10,741
Identity Mappings in Deep Residual Networks
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · arXiv · doi:10.48550/arxiv.1603.05027 · Cited by 10,665
Mastering the game of Go without human knowledge
Silver, D., Schrittwieser, J., Simonyan, K., Antonoglou, I., Huang, A., Guez, A., Hubert, T., Baker, L., Lai, M., Bolton, A., et al. (2017) · Nature · doi:10.1038/nature24270 · Cited by 9,084
Self-citationIn-Datacenter Performance Analysis of a Tensor Processing Unit
Jouppi, N. P., Young, C., Patil, N., Patterson, D., Agrawal, G., Bajwa, R., Bates, S., Bhatia, S., Boden, N., Borchers, A., et al. (2017) · Proceedings of the 44th Annual International Symposium on Computer Architecture · doi:10.1145/3079856.3080246 · Cited by 4,537
Encyclopedia of Artificial Intelligence
(2009) · IGI Global eBooks · doi:10.4018/978-1-59904-849-9 · Cited by 1,396
Campbell, M., Hoane, A. J., & Hsu, F.-H. (2002) · Artificial Intelligence · doi:10.1016/s0004-3702(01)00129-1 · Cited by 1,226
Lim, T.-S., Loh, W.-Y., & Shih, Y.-S. (2000) · Machine Learning · doi:10.1023/a:1007608224229 · Cited by 1,133
An analysis of alpha-beta pruning
Knuth, D. E., & Moore, R. W. (1975) · Artificial Intelligence · doi:10.1016/0004-3702(75)90019-3 · Cited by 1,004
Research works printed but not linked: 7
- C. J. Maddison A. Huang I. Sutskever D. Silver paper presented at the International Conference on Learning Representations 2015 San Diego CA 7 to 9 May 2015.
- D. N. L. Levy M. Newborn How Computers Play Chess (Ishi Press 2009).
- W. Steinitz The Modern Chess Instructor (Edition Olms 1990).
- O. Arenz “Monte Carlo chess ” master’s thesis Technische Universität Darmstadt (2012).
- 10.1023/a:1007634325138 · DOI
- T. Kaneko K. Hoki in Advances in Computer Games – 13th International Conference ACG 2011 Revised Selected Papers Tilburg Netherlands 20 to 22 November 2011 (Springer 2012) pp. 158–169.
- R. Ramanujan A. Sabharwal B. Selman in Proceedings of the 26th Conference on Uncertainty in Artificial Intelligence (UAI 2010) Catalina Island CA 8 to 11 July (AUAI Press 2010).
Cited by
Items 301–400 of 3,632
Human-Level Control Through Directly Trained Deep Spiking Q -Networks
Liu, G., Deng, W., Xie, X., Huang, L., & Tang, H. (2022) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2022.3198259 · Cited by 101
“The development of DSRL lags behind DRL, while DRL has made tremendous successes, achieved and even surpassed human-level performance in many reinforcement learning (RL) tasks [16], [17], [18], [19], [20], [21].”
BackgroundOn the necessity of abstraction
Konidaris, G. (2018) · Current Opinion in Behavioral Sciences · doi:10.1016/j.cobeha.2018.11.005 · Cited by 105
“AlphaZero essentially solves chess [2].”
ChainerRL: A Deep Reinforcement Learning Library
Fujita, Y., Nagarajan, P., Kataoka, T., & Ishikawa, T. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1912.03905 · Cited by 68
“In recent years, significant strides have been made in numerous complex sequential decision-making problems including game-playing [1, 2] and robotic manipulation [3, 4].”
BackgroundNeural forecasting: Introduction and literature overview.Preprint
Benidis, K., Rangapuram, S. S., Flunkert, V., Wang, B., Maddix, D. C., Türkmen, A. C., Gasthaus, J. A., Bohlke‐Schneider, M., Salinas, D., Stella, L., et al. (2020) · arXiv (Cornell University) · Cited by 95
Xu, S., Wang, J., Wang, R., Chen, J., & Zou, W. (2019) · Optics Express · doi:10.1364/oe.27.019778 · Cited by 72
Tripathi, N., Goshisht, M. K., Sahu, S. K., & Arora, C. (2021) · Molecular Diversity · doi:10.1007/s11030-021-10237-z · Cited by 80
Self-play reinforcement learning guides protein engineering
Wang, Y., Tang, H., Huang, L., Pan, L., Yang, L., Yang, H., Mu, F., & Yang, M. (2023) · Nature Machine Intelligence · doi:10.1038/s42256-023-00691-9 · Cited by 104
“To address this, we propose EvoPlay, a self-play reinforcement learning framework based on the single-player version of AlphaZero.”
MethodDeep Reinforcement Learning: A Chronological Overview and Methods
Terven, J. R. (2025) · AI · doi:10.3390/ai6030046 · Cited by 122
“• AlphaZero [107]: Generalized the AlphaGo approach to chess and shogi, emphasizing how powerful planning with deep policy and value networks can produce superhuman play from scratch.”
Edge Intelligence—On the Challenging Road to a Trillion Smart Connected IoT Devices
Keshavarzi, A., & van den Hoek, W. (2019) · IEEE Design and Test · doi:10.1109/mdat.2019.2899075 · Cited by 73
“For example, systems used by Google DeepMind are based on this concept [39].”
MethodProbing Slow Earthquakes With Deep Learning
Rouet‐Leduc, B., Hulbert, C., McBrearty, I. W., & Johnson, P. A. (2020) · Geophysical Research Letters · doi:10.1029/2019gl085870 · Cited by 70
“CNNs are at the core of recent dramatic advances in computer vision, natural language processing, and recommender systems (Balestriero & Baraniuk, 2018; Ioffe & Szegedy, 2015; Kingma & Ba, 2014; Krizhevsky et al., 2012; LeCun et al., 2015; Nair & Hinton, 2010; Srivastava et al., 2014; Silver et al., 2016, 2018).”
BackgroundQuantum enhancements for deep reinforcement learning in large spaces
Jerbi, S., Trenkwalder, L. M., Nautrup, H. P., Briegel, H. J., & Dunjko, V. (2019) · arXiv · doi:10.48550/arxiv.1910.12760 · Cited by 69
Dynamic Origin–Destination Matrix Prediction with Line Graph Neural Networks and Kalman Filter
Xiong, X., Özbay, K. M. A., Jin, L., & Feng, C. (2020) · Transportation Research Record Journal of the Transportation Research Board · doi:10.1177/0361198120919399 · Cited by 76
“In recent years, deep neural networks have shown to be effective in approximating nonlinear features in classification, regression and control problems [18, 19].”
Modelling Stock Markets by Multi-agent Reinforcement Learning
Lussange, J., Lazarevich, I. A., Bourgeois‐Gironde, S., Palminteri, S., & Gutkin, B. S. (2020) · Computational Economics · doi:10.1007/s10614-020-10038-w · Cited by 90
“…breakthroughs: the steady advances of cognitive neuroscience and neuroeconomics (Eickhoff et al. 2018; Konovalov and Krajbich 2016), and the progress in reinforcement learning theory due to a resurgence of machine learning methods and especially multi-agent learning (Silver et al. 2018a, b).”
BackgroundCroll, H. C., Ikuma, K., Ong, S. K., & Sarkar, S. (2023) · Critical Reviews in Environmental Science and Technology · doi:10.1080/10643389.2023.2183699 · Cited by 79
“…study in the game-play arena showed that pre-training using expert knowledge can hinder the performance of an RL agent by teaching it bad habits (Silver et al., 2018), another study in the wastewater industry showed that pretraining improves the RL agent’s learning speed and final performance,…”
BackgroundQuantum circuit optimization with deep reinforcement learningPreprint
Fösel, T., Niu, M. Y., Marquardt, F., & Li, L. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2103.07585 · Cited by 248
“Due to the generality of this approach, RL is being employed in a broad and growing spectrum of applications, such as robotics [18], video games [19, 20], board games like chess and Go [21, 22], chemistry [23– 25], neural architecture search [26], and chip placement [27].”
MethodReinforcement 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
A survey of approaches for implementing optical neural networks
Xu, R., Lv, P., Xu, F., & Shi, Y. (2020) · Optics & Laser Technology · doi:10.1016/j.optlastec.2020.106787 · Cited by 78
Saliba, S., Bowes, B. D., Adams, S., et al. (2020) · Water · doi:10.3390/w12113222 · Cited by 65
“The relationship of the Actor and the Critic dictates how the overall agent learns based on a reward function that reinforces the learning [20].”
BackgroundExpertise increases planning depth in human gameplay
van Opheusden, B., Kuperwajs, I., Galbiati, G., Bnaya, Z., Li, Y., & Ma, W. J. (2023) · Nature · doi:10.1038/s41586-023-06124-2 · Cited by 74
Atomically Precise Manufacturing of Silicon Electronics
Pitters, J., Croshaw, J., Achal, R., Livadaru, L., Ng, S. H., Lupoiu, R. C., Chutora, T., Huff, T. R., Walus, K., & Wolkow, R. A. (2024) · ACS Nano · doi:10.1021/acsnano.3c10412 · Cited by 72
Recent Advances in Generative Adversarial Networks for Gene Expression Data: A Comprehensive Review
Lee, M. (2023) · Mathematics · doi:10.3390/math11143055 · Cited by 83
Learning in continuous action space for developing high dimensional potential energy models
Manna, S., Loeffler, T. D., Batra, R., Banik, S., Chan, H., Varughese, B., Sasikumar, K., Sternberg, M., Peterka, T., Cherukara, M. J., et al. (2022) · Nature Communications · doi:10.1038/s41467-021-27849-6 · Cited by 74
“In a significant departure from traditional discrete MCTS14 and recent continuous action space MCTS approaches11,13, we introduce three concepts to tackle continuous action space problems: (1) a uniqueness function to avoid degeneracy, (2) correlating treedepth to action space, and (3) implementing an adaptive samplin…”
MethodEmbedded 3D Printing of Multimaterial Polymer Lattices via Graph‐Based Print Path Planning
Weeks, R. D., Truby, R. L., Uzel, S. G. M., & Lewis, J. A. (2022) · Advanced Materials · doi:10.1002/adma.202206958 · Cited by 70
Bioinspired nanofluidic iontronics for brain-like computing
Yu, L., Li, X., Luo, C., Lei, Z., Wang, Y., Hou, Y., Wang, M., & Hou, X. (2023) · Nano Research · doi:10.1007/s12274-023-5900-y · Cited by 72
The Potential of Neural Network Potentials
Duignan, T. T. (2024) · ACS Physical Chemistry Au · doi:10.1021/acsphyschemau.4c00004 · Cited by 82
EdgeSlice: Slicing Wireless Edge Computing Network with Decentralized Deep Reinforcement Learning
Liu, Q., Han, T., & Moges, E. (2020) · IEEE International Conference on Distributed Computing Systems (ICDCS) · doi:10.1109/icdcs47774.2020.00028 · Cited by 86
“Although DRL techniques have been extensively studied in many areas such as robotic control [16], traffic control [17], and chess games [18], the existing DRL models are not appropriate to solve problem P3 for two reasons.”
BackgroundComparative analysis of machine learning methods for active flow control
Pino, F., Schena, L., Rabault, J., & Mendez, M. A. (2023) · Journal of Fluid Mechanics · doi:10.1017/jfm.2023.76 · Cited by 76
Wang, W., Zhang, Y., Gao, J., Jiang, Y., Yang, Y., Zheng, Z., Zou, W., Li, J., Zhang, C., Cao, W., et al. (2023) · Communications in Transportation Research · doi:10.1016/j.commtr.2023.100096 · Cited by 74
Graph neural networks-based scheduler for production planning problems using reinforcement learning
Hameed, M. S. A., & Schwung, A. (2023) · Journal of Manufacturing Systems · doi:10.1016/j.jmsy.2023.06.005 · Cited by 102
Recent advances in deep learning for retrosynthesis
Zhong, Z., Song, J., Feng, Z., Liu, T., Jia, L., Yao, S., Hou, T., & Song, M. (2023) · Wiley Interdisciplinary Reviews Computational Molecular Science · doi:10.1002/wcms.1694 · Cited by 107
Dynamic Scene Path Planning of UAVs Based on Deep Reinforcement Learning
Tang, J., Liang, Y., & Li, K. (2024) · Drones · doi:10.3390/drones8020060 · Cited by 79
“The fundamental unit of a neural network is the neuron, and the prevalent “M-P neuron model” [24] is currently adopted.”
Deep reinforcement learning for process design: Review and perspective
Gao, Q., & Schweidtmann, A. M. (2024) · Current Opinion in Chemical Engineering · doi:10.1016/j.coche.2024.101012 · Cited by 100
Emergent World Representations: Exploring a Sequence Model Trained on a Synthetic Task
Li, K., Hopkins, A. K., Bau, D., Viégas, F., Pfister, H., & Wattenberg, M. (2022) · arXiv · doi:10.48550/arxiv.2210.13382 · Cited by 61
AIR5: Five Pillars of Artificial Intelligence Research
Ong, Y.-S., & Gupta, A. (2019) · IEEE Transactions on Emerging Topics in Computational Intelligence · doi:10.1109/tetci.2019.2928344 · Cited by 76
“2928344 of chess [6], and Carnegie Mellon University’s AI defeating four of the”
BackgroundTaxonomy of Reinforcement Learning Algorithms
Zhang, H., & Yu, T. (2020) · Springer eBooks · doi:10.1007/978-981-15-4095-0_3 · Cited by 114
Quantum Algorithms Applied to Satellite Mission Planning for Earth Observation
Rainjonneau, S., Tokarev, I., Iudin, S., Rayaprolu, S., Pinto, K., Lemtiuzhnikova, D., Koblan, M., Barashov, E., Kordzanganeh, M., Pflitsch, M., et al. (2023) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2023.3287154 · Cited by 83
“AlphaZero uses a computational tree of environment states whose values and probabilities are determined from 10 the outputs of value and policy networks.”
BackgroundMethodInfluentialFoundation models and intelligent decision-making: Progress, challenges, and perspectives
Huang, J., Xu, Y., Wang, Q., Wang, Q., Liang, X., Wang, F., Zhang, Z., Wei, W., Zhang, B., Huang, L., et al. (2025) · The Innovation · doi:10.1016/j.xinn.2025.100948 · Cited by 186
Leibo, J. Z., Hughes, E., Lanctot, M., & Graepel, T. (2019) · arXiv · doi:10.48550/arxiv.1903.00742 · Cited by 60
“A similar phenomenon was observed in AlphaZero: its preferences towards certain Chess openings fluctuated in time; the most frequently played openings at certain intermediate points in training were no longer seen in its later stages (Silver et al., 2018).”
BackgroundResultSelf-citationDeep-Neural-Network Discrimination of Multiplexed Superconducting-Qubit States
Lienhard, B., Vepsäläinen, A. P., Govia, L. C. G., Hoffer, C. R., Qiu, J. Y., Ristè, D., Ware, M., Kim, D. K., Winik, R., Melville, A. J., et al. (2022) · Physical Review Applied · doi:10.1103/physrevapplied.17.014024 · Cited by 84
Adaptive Supply Chain: Demand–Supply Synchronization Using Deep Reinforcement Learning
Kegenbekov, Z., & Jackson, I. (2021) · Algorithms · doi:10.3390/a14080240 · Cited by 78
“Many skeptics can undervalue the astonishing success of AlphaZero in mastering such classic board games as chess, shogi, and go through self-play [7] by pointing out the relative simplicity of rules and determinism of games.”
BackgroundExplainability in Deep Reinforcement Learning: A Review into Current Methods and Applications
Hickling, T., Zenati, A., Aouf, N., & Spencer, P. (2023) · ACM Computing Surveys · doi:10.1145/3623377 · Cited by 89
“(2018)[50]) to be able to strategise many moves in advance.”
BackgroundEvans, R., Hernández-Orallo, J., Welbl, J., Kohli, P., & Sergot, M. (2021) · Artificial Intelligence · doi:10.1016/j.artint.2020.103438 · Cited by 58
A Review of Uncertainty for Deep Reinforcement Learning
Lockwood, O., & Si, M. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment · doi:10.1609/aiide.v18i1.21959 · Cited by 128
Game of Drones: Intelligent Online Decision Making of Multi-UAV Confrontation
Liu, D., Zong, Q., Zhang, X., Zhang, R., Dou, L., & Tian, B. (2024) · IEEE Transactions on Emerging Topics in Computational Intelligence · doi:10.1109/tetci.2024.3360282 · Cited by 65
“…abilities in high-dimensional data compared with traditional optimization methods, allows the automatic discovery of efficient confrontation strategies, for the decision-makingproblemssuchasAlphaGo[3],[4]andAlphaStar[5],[6] series,theself-taughtagentscouldbeatupthetopprofessional humanplayers.”
Algorithm-Aided Prediction of Patient Preferences — An Ethics Sneak Peek
Biller-Andorno, N., & Biller, A. (2019) · New England Journal of Medicine · doi:10.1056/nejmms1904869 · Cited by 57
Overview of Machine Learning Part 1
Maleki, F., Ovens, K. L., Najafian, K., Forghani, B., Reinhold, C., & Forghani, R. (2020) · Neuroimaging Clinics of North America · doi:10.1016/j.nic.2020.08.007 · Cited by 92
Distribution System Resilience Under Asynchronous Information Using Deep Reinforcement Learning
Bedoya, J. C., Wang, Y., & Liu, C. (2021) · IEEE Transactions on Power Systems · doi:10.1109/tpwrs.2021.3056543 · Cited by 61
A Novel Hierarchical Soft Actor-Critic Algorithm for Multi-Logistics Robots Task Allocation
Tang, H., Wang, A., Xue, F., Yang, J., & Cao, Y. (2021) · IEEE Access · doi:10.1109/access.2021.3062457 · Cited by 71
“In recent years, with the great success ofAlphaGo [11]–[13] in the field of go, deep reinforcement learning algorithms have been greatly applied in various fields of study.”
BackgroundMohammad, N., Ahmad, R., Kurniawan, A., & Yusof, M. Y. P. M. (2022) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2022.1049584 · Cited by 77
“successfully implemented in closed-world environments such as video games (Silver et al., 2018) and applies to multi-agent systems, such as electronic markets (Peters”
BackgroundVignon, C., Rabault, J., Vasanth, J., Alcántara-Ávila, F., Mortensen, M., & Vinuesa, R. (2023) · Physics of Fluids · doi:10.1063/5.0153181 · Cited by 80
Croll, H. C., Ikuma, K., Ong, S. K., & Sarkar, S. (2023) · Environmental Science & Technology · doi:10.1021/acs.est.3c00353 · Cited by 70
Park: An Open Platform for Learning-Augmented Computer Systems
Mao, H., Negi, P., Narayan, A., Wang, H., Yang, J., Wang, H., Marcus, R., Addanki, R., Shirkoohi, M. K., He, S., et al. (2019) · DSpace@MIT (Massachusetts Institute of Technology) · Cited by 114
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 93
Tactical Decision-Making in Autonomous Driving by Reinforcement Learning with Uncertainty Estimation
Hoel, C.-J., Wolff, K., & Laine, L. (2020) · IEEE Intelligent Vehicles Symposium · doi:10.1109/iv47402.2020.9304614 · Cited by 81
“Reinforcement learning (RL) techniques have been successful in various domains during the last decade [10], [11], [12].”
BackgroundOff-the-shelf deep learning is not enough, and requires parsimony, Bayesianity, and causality
Vasudevan, R. K., Ziatdinov, M. A., Vlček, L., & Kalinin, S. V. (2021) · npj Computational Materials · doi:10.1038/s41524-020-00487-0 · Cited by 61
A visual path-following learning approach for industrial robots using DRL
Maldonado-Ramirez, A., Rìos-Cabrera, R., & López-Juárez, I. (2021) · Robotics and Computer-Integrated Manufacturing · doi:10.1016/j.rcim.2021.102130 · Cited by 61
Applying reinforcement learning and tree search to the unit commitment problem
de Mars, P., & O'Sullivan, A. (2021) · Applied Energy · doi:10.1016/j.apenergy.2021.117519 · Cited by 60
“Inspired by similar approaches in other domains [9,6], in this paper we use a novel RL-aided tree search algorithm, ‘guided tree search’, to solve the UC problem with uncertain demand and wind generation.”
BackgroundMethodVasudeva, A., Sheikh, N. A., & Sahu, S. (2021) · Journal of Family Medicine and Primary Care · doi:10.4103/jfmpc.jfmpc_692_21 · Cited by 56
“[6] It will require the requisite hardware of various dimensions to accommodate the diversity that exists amongst human beings.”
BackgroundMulti-type data fusion framework based on deep reinforcement learning for algorithmic trading
Liu, P., Zhang, Y., Bao, F., Yao, X., & Zhang, C. (2022) · Applied Intelligence · doi:10.1007/s10489-022-03321-w · Cited by 61
Flowsheet generation through hierarchical reinforcement learning and graph neural networks
Stops, L., Leenhouts, R. J., Gao, Q., & Schweidtmann, A. M. (2022) · AIChE Journal · doi:10.1002/aic.17938 · Cited by 76
van Steenbergen, R., Mes, M., & van Heeswijk, W. (2023) · Transportation Research Part C Emerging Technologies · doi:10.1016/j.trc.2023.104401 · Cited by 70
Kell, D. B., Samanta, S., & Swainston, N. (2020) · Biochemical Journal · doi:10.1042/bcj20200781 · Cited by 65
Review of Deep Reinforcement Learning Approaches for Conflict Resolution in Air Traffic Control
Wang, Z., Pan, W., Li, H., Wang, X., & Zuo, Q. (2022) · Aerospace · doi:10.3390/aerospace9060294 · Cited by 69
“DRL has been utilized in many fields and obtained great achievements of human-level or superhuman performance [7–9].”
MethodDeep reinforcement learning for the optimal placement of cryptocurrency limit orders
Schnaubelt, M. (2021) · European Journal of Operational Research · doi:10.1016/j.ejor.2021.04.050 · Cited by 61
“Inspired by recent successes of deep RL in solving complex tasks such as playing Go or controlling a robot’s hand to solve the Rubik’s Cube (see, for example, Silver et al., 2018; Akkaya et al., 2019), there is rising interest in both academia and practice to apply RL to optimal execution problems.”
MethodAvoiding 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
In-process quality improvement: Concepts, methodologies, and applications
Shi, J. (2022) · IISE Transactions · doi:10.1080/24725854.2022.2059725 · Cited by 64
“(1) Silver et al., 2018).”
The signature-testing approach to mapping biological and artificial intelligences
Taylor, A. H., Bastos, A. P. M., Brown, R. L., & Allen, C. (2022) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2022.06.002 · Cited by 67
Quantum Enhancements for Deep Reinforcement Learning in Large Spaces
Jerbi, S., Trenkwalder, L. M., Nautrup, H. P., Briegel, H. J., & Dunjko, V. (2021) · PRX Quantum · doi:10.1103/prxquantum.2.010328 · Cited by 111
Big data, machine learning and artificial intelligence: a neurologist’s guide
Auger, S. D., Jacobs, B. M., Dobson, R., Marshall, C. R., & Noyce, A. J. (2020) · Practical Neurology · doi:10.1136/practneurol-2020-002688 · Cited by 65
PrefixRL: Optimization of Parallel Prefix Circuits using Deep Reinforcement Learning
Roy, R., Raiman, J., Kant, N., Elkin, I., Kirby, R., Siu, M., Oberman, S. F., Godil, S., & Catanzaro, B. (2021) · ACM/IEEE Design Automation Conference (DAC) · doi:10.1109/dac18074.2021.9586094 · Cited by 56
“RL agents have outperformed humans in complex games [12] and produced novel solutions to search and design problems [13].”
BackgroundMethodSleep-like unsupervised replay reduces catastrophic forgetting in artificial neural networks
Tadros, T., Krishnan, G. P., Ramyaa, R., & Bazhenov, M. (2022) · Nature Communications · doi:10.1038/s41467-022-34938-7 · Cited by 94
“Although deep neural networks 9 can achieve supra-human level of performance on tasks ranging from complex games 10 to image recognition 11 , they lie at a suboptimal point on the stability-plasticity spectrum.”
BackgroundLopez-Martinez, D., Eschenfeldt, P., Ostvar, S., Ingram, M., Hur, C., & Picard, R. (2019) · 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) · doi:10.1109/embc.2019.8857295 · Cited by 53
“Combining RL with deep neural networks has achieved tremendous success in complex problems such as the Atari, Go, and chess games [6], [7].”
Artificial Intelligence for Social Good: A Survey
Shi, Z. R., Wang, C., & Fang, F. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2001.01818 · Cited by 53
“AI has also greatly advanced human capabilities in complex decision-making processes ranging from determining how to allocate security resources to protect airports [4] to games such as poker [5] and Go [6].”
BackgroundWang, R., Lehman, J., Rawal, A., Zhi, J., Li, Y., Clune, J., & Stanley, K. O. (2020) · arXiv · doi:10.48550/arxiv.2003.08536 · Cited by 53
“Other longstanding threads of research into self-play (Balduzzi et al., 2019; Bansal et al., 2018; OpenAI et al., 2019; Samuel, 1967; Silver et al., 2018) and Generative Adversarial Networks (GANs) (Goodfellow et al.”
BackgroundRisks from Learned Optimization in Advanced Machine Learning Systems
Hubinger, E., van Merwijk, C., Mikulik, V., Skalse, J., & Garrabrant, S. (2019) · arXiv · doi:10.48550/arxiv.1906.01820 · Cited by 53
“For example, Go, Chess, and Shogi have a very large domain of possible board states, but admit a single high-level strategy for play— heuristic-guided tree search—that performs well across all board states.[10] On the other hand, a classifier trained on random noise is unlikely to benefit from compression at all.”
BackgroundMethodStecuła, A., Hussain, M. S., & Viola, R. E. (2020) · Journal of Medicinal Chemistry · doi:10.1021/acs.jmedchem.0c00473 · Cited by 55
A Deep Reinforcement Learning Approach to Dynamic Loading Strategy of Repairable Multistate Systems
Chen, Y., Liu, Y., & Xiahou, T. (2021) · IEEE Transactions on Reliability · doi:10.1109/tr.2020.3044596 · Cited by 55
Black Boxes or Unflattering Mirrors? Comparative Bias in the Science of Machine Behaviour
Buckner, C. J. (2021) · The British Journal for the Philosophy of Science · doi:10.1086/714960 · Cited by 73
A Leap among Quantum Computing and Quantum Neural Networks: A Survey
Massoli, F. V., Vadicamo, L., Amato, G. M., & Falchi, F. (2022) · ACM Computing Surveys · doi:10.1145/3529756 · Cited by 94
“Thanks to their representation power and the development of new technologies and training algorithms, DL models obtained astonishing results in the past two decades, achieving superhuman performance on certain tasks [174].”
BackgroundSigned Neuron with Memory: Towards Simple, Accurate and High-Efficient ANN-SNN Conversion
Wang, Y., Zhang, M., Chen, Y., & Qu, H. (2022) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2022/347 · Cited by 71
Development of residual learning in deep neural networks for computer vision: A survey
Xu, G., Wang, X., Wu, X., Leng, X., & Xu, Y. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2024.109890 · Cited by 114
“3) Residual correction in reinforcement learning: Following the success of AlphaGo [157] and AlphaZero [158], reinforcement learning (RL) has demonstrated a significant improvement in interactive environments, such as gaming and self-control in robots.”
A reinforcement learning approach for quantum state engineering
Mackeprang, J., Dasari, D. B. R., & Wrachtrup, J. (2020) · Quantum Machine Intelligence · doi:10.1007/s42484-020-00016-8 · Cited by 56
INVIGORATE: Interactive Visual Grounding and Grasping in Clutter
Zhang*, H., Lu*, Y., Yu, C., Hsu, D., Lan, X., & Zheng, N. (2021) · Robotics: Science and Systems XVII · doi:10.15607/rss.2021.xvii.020 · Cited by 52
“To scale up complex decision making, both Silver et al. [32, 33, 34] and Cai et al. [2] use planning in the short term and use learning for long-term prediction.”
BackgroundOnline Continual Learning from Imbalanced Data
Chrysakis, A., & Moens, M. (2020) · Lirias (KU Leuven) · Cited by 209
Wehrle, M., Lechler, S., von der Gracht, H. A., & Hartmann, E. (2020) · Journal of Business Logistics · doi:10.1111/jbl.12259 · Cited by 61
AI for Everyone? Critical Perspectives
(2021) · University of Westminster Press eBooks · doi:10.16997/book55 · Cited by 51
Yang, J., Xi, M., Wen, J., Li, Y., & Song, H. (2022) · Digital Communications and Networks · doi:10.1016/j.dcan.2022.05.005 · Cited by 60
Peng, J., Ren, T., Chen, Z., Chen, W., Wu, C., & Ma, C. (2024) · Journal of Cleaner Production · doi:10.1016/j.jclepro.2024.141360 · Cited by 56
Effects of Communication Directionality and AI Agent Differences in Human-AI Interaction
Ashktorab, Z., Dugan, C., Johnson, J., Pan, Q., Zhang, W., Kumaravel, S., & Campbell, M. (2021) · Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3411764.3445256 · Cited by 50
“Games are frequently used as a test bed for state-of-the-art AI algorithms and an important application domain of AI [13, 22, 58, 64].”
BackgroundModel-based reinforcement learning for ultrasound-driven autonomous microrobots
Medany, M., Piglia, L., Achenbach, L., Mukkavilli, S. K., & Ahmed, D. (2025) · Nature Machine Intelligence · doi:10.1038/s42256-025-01054-2 · Cited by 56
“Although RL has the potential to surpass human capabilities in tasks like object manipulation and strategic gameplay 41,42 , its reliance on extensive interactions for stable training introduces unique challenges for microrobots, as the experimental conditions are highly uncontrolled and variable.”
Machine Learning-Based Unbalance Detection of a Rotating Shaft Using Vibration Data
Mey, O., Neudeck, W., Schneider, A. F., & Enge‐Rosenblatt, O. (2020) · IEEE International Conference on Emerging Technologies and Factory Automation (ETFA) · doi:10.1109/etfa46521.2020.9212000 · Cited by 71
“Progress in the field of machine learning has led to impressive results in recent years, for example in the areas of image recognition [1]–[4], natural language processing [5]–[9] or reinforcement learning [10]–[13].”
BackgroundKumar, V., Singh, D., Kaur, M., & Damaševičius, R. (2021) · PeerJ Computer Science · doi:10.7717/peerj-cs.564 · Cited by 54
Abbass, M. A. B., & Kang, H. (2023) · Drones · doi:10.3390/drones7040225 · Cited by 55
Khadivi, 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 103
“Since then, DeepMind has continued to excite the community with many other big developments, including Double-DQNs (Hasselt et al., 2016), AlphaGo (Silver et al., 2016), AlphaZero (Silver et al., 2018), and MuZero (Schrittwieser et al., 2020).”
BackgroundV-MPO: On-Policy Maximum a Posteriori Policy Optimization for Discrete and Continuous Control
Song, H., Abdolmaleki, A., Springenberg, J. T., Clark, A., Soyer, H., Rae, J. W., Noury, S., Ahuja, A., Liu, S., Tirumala, D., et al. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1909.12238 · Cited by 50
“Deep reinforcement learning (RL) with neural network function approximators has achieved superhu-man performance in several challenging domains (Mnih et al., 2015; Silver et al., 2016; 2018).”
BackgroundHow much intelligence is there in artificial intelligence? A 2020 update
van der Maas, H. L. J., Snoek, L., & Stevenson, C. E. (2021) · Intelligence · doi:10.1016/j.intell.2021.101548 · Cited by 67
“…could scale up reinforcement learning models that in turn led to the development of systems capable of (super) human performance on a range of tasks, including playing board games (Silver et al., 2018) and video games (Mnih et al., 2015), as well as autonomous driving (Kiran et al., 2020).”
Symphony: Learning Realistic and Diverse Agents for Autonomous Driving Simulation
Igl, M., Kim, D. W., Kuefler, A., Mougin, P., Shah, P., Shiarlis, K., Anguelov, D., Palatucci, M., White, B., & Whiteson, S. (2022) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra46639.2022.9811990 · Cited by 100
Enhancing Differential-Neural Cryptanalysis
Bao, Z., Guo, J., Liu, M., et al. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-22963-3_11 · Cited by 49
Lupo, A., Picco, E., Zajnulina, M., & Massar, S. (2023) · Optica · doi:10.1364/optica.489501 · Cited by 64
“Recent advances include, for instance, image classification and analysis [31], game playing [34], protein structure prediction[3, 18], chatbots that simulate human conversation such as ChatGPT and Bing[29, 5], and more.”
BackgroundAccelerating Self-Play Learning in GoPreprint
Wu, D. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1902.10565 · Cited by 122
Metadata
- Authors
- Silver, D., Hubert, T., Schrittwieser, J., Antonoglou, I., Lai, M., Guez, A., Lanctot, M., Sifre, L., Kumaran, D., Graepel, T., Lillicrap, T., Simonyan, K., & Hassabis, D.
- Affiliations
- Google DeepMind (United Kingdom) (GB)
- Corresponding author
- Silver, David
- Hubert, Thomas
- Schrittwieser, Julian
- Hassabis, Demis
- ORCID
- Venue
- Science
- Published
- 2018-12-07
- Volume
- 362
- Issue
- 6419
- Pages
- 1140-1144
- ISSN
- 0036-8075 (ISSN-L)
- 1095-9203
- OpenAlex
- W2902907165
- Cited by
- 3,562
- References
- 30
- Citations by year
- 2026: 257 · 2025: 427 · 2024: 512 · 2023: 514 · 2022: 569 · 2021: 546 · 2020: 462 · 2019: 256 · 2018: 13
- FWCI
- 179.89
- Type
- journal article
- Language
- en
- Open access
- bronze
- doi.org
- Locations
- Science (publishedVersion)
- PubMed (publishedVersion)
- UCL Discovery (University College London) (submittedVersion)
- Funders
- DeepMind
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Artificial Intelligence in Games
- Physical Sciences › Computer Science › Artificial Intelligence › Reinforcement Learning in Robotics
- Physical Sciences › Computer Science › Artificial Intelligence › Artificial Intelligence Applications
- Keywords
- AlphaZero, chess, shogi, Go, evaluation function learning, game tree search
- MeSH terms
- Algorithms, Artificial Intelligence, Humans, Reinforcement, Psychology, Software, 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.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
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.