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 1,501–1,600 of 3,633
Learning a Behavioral Repertoire from Demonstrations
Justesen, N., González-Duque, M., Cabarcas, D., Mouret, J.-B., & Risi, S. (2020) · IEEE Conference on Games (CoG) · doi:10.1109/cog47356.2020.9231897 · Cited by 7
“Deep reinforcement learning has shown impressive results, especially for board games [1] and video games [2].”
BackgroundOn Hard Exploration for Reinforcement Learning: A Case Study in Pommerman
Gao, C., Kartal, B., Hernández-Leal, P., & Taylor, M. E. (2019) · Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment · doi:10.1609/aiide.v15i1.5220 · Cited by 12
Fast Adaptation to New Environments via Policy-Dynamics Value Functions
Răileanu, R., Goldstein, M. A., & Szlam, A. D. (2020) · International Conference on Machine Learning · Cited by 31
Improving the Performance of MCTS-Based µRTS Agents Through Move Pruning
Ouessai, A., Salem, M., & Mora, A. M. (2020) · IEEE Conference on Games (CoG) · doi:10.1109/cog47356.2020.9231715 · Cited by 7
“Later work on AlphaZero [26] fully discarded expert knowledge.”
BackgroundModel-Free Deep Reinforcement Learning—Algorithms and Applications
Otto, F. (2021) · Studies in computational intelligence · doi:10.1007/978-3-030-41188-6_10 · Cited by 11
Kong, W., Zhou, D., Zhang, K., Yang, Z. C., & Yang, W. (2020) · International Conference on Machine Learning and Applications · doi:10.1109/icmla51294.2020.00238 · Cited by 7
“Moreover, reinforcement learning (RL) has made great strides in the field of adversarial games in recent years [3]–[5].”
BackgroundMethodFine-tuning Deep RL with Gradient-Free Optimization
de Bruin, T., Kober, J., Tuyls, K., & Babuška, R. (2020) · IFAC-PapersOnLine · doi:10.1016/j.ifacol.2020.12.2240 · Cited by 11
“Deep Reinforcement Learning (DRL) is a learning framework that has recently enjoyed significant successes (Silver et al., 2018; Mnih et al., 2015; Bansal et al., 2017; Andrychowicz et al., 2018).”
BackgroundSelf‐Programming Synaptic Resistor Circuit for Intelligent Systems
Shaffer, C. M., Deo, A., Tudor, A., Shenoy, R., Danesh, C., Nathan, D., Gamble, L. L., Inman, D. J., & Chen, Y. (2021) · Advanced Intelligent Systems · doi:10.1002/aisy.202100016 · Cited by 7
The Cerebral Cortex: A Delay-Coupled Recurrent Oscillator Network?
Singer, W. (2021) · Natural computing series · doi:10.1007/978-981-13-1687-6_1 · Cited by 8
Minds, Machines, and Molecules
Brunet, T. D. P., & Halina, M. (2020) · Philosophical Topics · doi:10.5840/philtopics202048111 · Cited by 24
“Recent advances in AI have resulted in computer programs with surprisingly sophisticated behaviours, such the ability to drive a car or play chess and Go at superhuman levels (Badue et al., 2019; Silver et al., 2018).”
Janković, B. R., & Gojobori, T. (2022) · Human Genomics · doi:10.1186/s40246-022-00376-1 · Cited by 8
InferNet for Delayed Reinforcement Tasks: Addressing the Temporal Credit Assignment Problem
Ausin, M. S., Azizsoltani, H., Ju, S., Kim, Y. J., & Chi, M. (2021) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata52589.2021.9671827 · Cited by 10
“For example, when applying DRL for games such as Chess, Shogi and Go, the final rewards are determined by the outcomes of the game: −1/0/+1 for loss/draw/win respectively; and for each state, Monte Carlo Tree Search (MCTS) was used to learn the likelihood of each outcome [21, 20].”
BackgroundMethodModel-Free Reinforcement Learning for Financial Portfolios: A Brief SurveyPreprint
Sato, Y. (2019) · RePEc: Research Papers in Economics · doi:10.48550/arxiv.1904.04973 · Cited by 46
“In recent years, a hybrid DRL algorithm called AlphaGo [4] defeated a human world champion in the game of Go, and its more advanced and generalized version, AlphaZero [5], convincingly defeated world champion programs in chess, Go, and Shogi, given no domain knowledge other than the underlying rules.”
MethodUnpredictability of AIPreprint
Yampolskiy, R. V. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1905.13053 · Cited by 16
I'm sorry Dave, I'm afraid I can't do that, Deep Q-learning from forbidden actionPreprint
Seurin, M., Preux, P., & Pietquin, O. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1910.02078 · Cited by 7
“Despite the success of Reinforcement Learning (RL) (Sutton and Barto, 2018) in different domains: Games (Silver et al., 2018), Ressources Management (Mao et al., 2016), Chemical reaction (Zhou et al., 2017), many reasons are deterring industrial from using reinforcement learning in the real world.”
BackgroundPettit, J. F., Glatt, R., Donadee, J., & Petersen, B. K. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1912.03408 · Cited by 17
“DRL has demonstrated superhuman performance on computer games [8], was used to beat the world’s strongest players in Go [13], and has made progress towards knowledge sharing [6] and multiagent learning [12].”
Divide-and-Conquer Monte Carlo Tree Search For Goal-Directed PlanningPreprint
Parascandolo, G., Buesing, L., Merel, J., Hasenclever, L., Aslanides, J., Hamrick, J. B., Heess, N., Neitz, A., & Weber, T. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2004.11410 · Cited by 29
“Note that we did not use visit counts as targets for the prior network — as done in AlphaGo and AlphaZero for example (Silver et al., 2016; 2018)— since for small search budgets visit counts tend to be noisy and require significant fine-tuning to avoid collapse (Hamrick et al., 2020).”
MethodJoint Policy Search for Multi-agent Collaboration with Imperfect InformationPreprint
Tian, Y., Gong, Q., & Jiang, T. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2008.06495 · Cited by 12
“, selfplay [35, 20]) work with guarantees, in multi-agent cooperative setting they often converge to sub-optimal Nash equilibrium.”
BackgroundThe Explanation Game: A Formal Framework for Interpretable Machine LearningPreprint
Watson, D., & Floridi, L. (2020) · SSRN Electronic Journal · doi:10.2139/ssrn.3509737 · Cited by 9
Scaling Scaling Laws with Board Games
Jones, A. L. (2021) · arXiv · doi:10.48550/arxiv.2104.03113 · Cited by 7
“Using AlphaZero [11], we train many different models on many different board sizes.”
MethodSkill Preferences: Learning to Extract and Execute Robotic Skills from Human FeedbackPreprint
Wang, X., Lee, K., Hakhamaneshi, K., Abbeel, P., & Laskin, M. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2108.05382 · Cited by 28
“Deep reinforcement learning (RL) is a framework for solving temporally extended tasks that has resulted in a number of breakthroughs in autonomous control including mastery of the game of Go [1, 2], learning to play video games [3, 4, 5], and learning basic robotic control [6, 7].”
BackgroundWhen Can We Learn General-Sum Markov Games with a Large Number of Players Sample-Efficiently?Preprint
Song, Z., Song, M., & Bai, Y. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2110.04184 · Cited by 80
“Multi-agent reinforcement learning (RL) has achieved substantial recent successes in solving artificial intelligence challenges such as GO (Silver et al., 2016, 2018), multi-player games with team play such as Starcraft (Vinyals et al., 2019) and Dota2 (Berner et al., 2019), behavior learning in…”
BackgroundReinforcement Learning in Quantitative Trading: A SurveyPreprint
Alameer, A., Saleh, H., & Alshehri, K. A. (2022) · doi:10.36227/techrxiv.19303853 · Cited by 14
Keerthika, K., Kannan, M. R., & Saravanan, T. (2024) · Wiley eBooks · doi:10.1002/9781394272587.ch7 · Cited by 8
Human strategy learning-based multi-agent deep reinforcement learning for online team sports game
Lee, S., Lee, G., Kim, W., Kim, J., Park, J., & Cho, K. (2025) · IEEE Access · doi:10.1109/access.2025.3531435 · Cited by 7
Improving Policies via Search in Cooperative Partially Observable Games
Lerer, A. K., Hu, H., Foerster, J., & Brown, N. (2020) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v34i05.6208 · Cited by 50
“…in reaching professional-level performance in zero-sum games, including backgammon (Tesauro 1994), chess (Campbell, Hoane Jr, and Hsu 2002), Go (Silver et al. 2016; Silver et al. 2017; Silver et al. 2018), and poker (Morav ˇ c ´ ık et al. 2017; Brown and Sandholm 2017; Brown and Sandholm 2019).”
BackgroundMethodThe Fittest Wins: A Multistage Framework Achieving New SOTA in ViZDoom Competition
Li, S., Xu, J., Dong, H., Yu, Y., Yuan, C., Sun, P., & Han, L. (2023) · IEEE Transactions on Games · doi:10.1109/tg.2023.3259724 · Cited by 7
“Multiagent RL: To obtain intelligent and superior AI agents in competitive multiagent games, self-play (SP) [16] is often necessary to generate high-quality competition.”
BackgroundSurrogate-assisted Monte Carlo Tree Search for real-time video games
Kim, M.-J., Lee, D., Kim, J. S., & Ahn, C. W. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2024.108152 · Cited by 8
Prediction of hepatic metastasis in esophageal cancer based on machine learning
Wan, J., & Zeng, Y. (2024) · Scientific Reports · doi:10.1038/s41598-024-63213-6 · Cited by 7
Analog Circuit Design Automation via Sequential RL Agents and Gm/ID Methodology
Hong, S., Tae, Y., Lee, D.-J., Park, G., Lim, J., Cho, K., Jeong, C., Park, M.-J., Hong, S., & Han, J. (2024) · IEEE Access · doi:10.1109/access.2024.3435331 · Cited by 10
A Machine Learning Approach to Solve the E-commerce Box-Sizing Problem
Kandula, S., Roy, D., & Akartunalı, K. (2024) · Production and Operations Management · doi:10.1177/10591478241282249 · Cited by 8
“RL agents have achieved superhuman performance in games such as Chess, Shogi, and Go (Silver et al., 2018).”
Understanding Visual Perception
Purves, D. (2025) · Journal of Cognitive Neuroscience · doi:10.1162/jocn_a_02292 · Cited by 6
Bridging the Gap: The Rise of Neurosymbolic Artificial Intelligence in Advanced Computing
Ganguly, P., & Mukherjee, I. (2025) · IT Professional · doi:10.1109/mitp.2025.3532388 · Cited by 7
Yan, S. L., Zhu, Y., Chen, W., et al. (2025) · Applied Sciences · doi:10.3390/app15084496 · Cited by 6
Shapley value-driven multi-modal deep reinforcement learning for complex decision-making
Zhang, J., Bao, B., Wang, C., & Zhu, F. (2025) · Neural Networks · doi:10.1016/j.neunet.2025.107650 · Cited by 6
Artificial intelligence in dentistry: insights and expectations from Swiss dental professionals
Donati, J., Beyer, M., Brokmeier, J., Neuhaus, K. W., Thieringer, F. M., & Berg, B.-I. (2025) · BMC Medical Informatics and Decision Making · doi:10.1186/s12911-025-03066-9 · Cited by 11
Separable Physics-Informed Neural Networks for Solving Elasticity Problems
Es’kin, V. A., Davydov, D. V., Gur'eva, J. V., Malkhanov, A. O., & Smorkalov, M. E. (2025) · Computational Mathematics and Mathematical Physics · doi:10.1134/s0965542525701118 · Cited by 17
Alrubyli, Y., & Bonarini, A. (2022) · IEEE International Conference on Mechatronics and Automation (ICMA) · doi:10.1109/icma54519.2022.9856292 · Cited by 8
“In the case of Go, an intelligent agent named AlphaGo beat the world champion [5].”
BackgroundFujita, K. (2022) · PeerJ Computer Science · doi:10.7717/peerj-cs.1123 · Cited by 10
“The weight decay and momentum are the same as those in the pseudocode of AlphaZero (Silver et al., 2018).”
BackgroundMethodSpaceGym: Discrete and Differential Games in Non-Cooperative Space Operations
Allen, R. E., Rachlin, Y., Ruprecht, J., Loughran, S., Varey, J., & Viggh, H. E. M. (2023) · IEEE Aerospace Conference (AERO) · doi:10.1109/aero55745.2023.10115968 · Cited by 18
“Atari [19], Go [57], [62], [63], and StarCraft [20]).”
BackgroundAI and atoms: How artificial intelligence is revolutionizing nuclear material production
He, J., & Degtyarev, N. (2023) · Bulletin of the Atomic Scientists · doi:10.1080/00963402.2023.2245251 · Cited by 6
Offline reinforcement learning in high-dimensional stochastic environments
Hêche, F., Barakat, O., Desmettre, T., Marx, T., & Robert, S. (2023) · Neural Computing and Applications · doi:10.1007/s00521-023-09029-3 · Cited by 10
“RL has achieved impressive results in a number of challenging areas, including games [2, 3], robotic control [4, 5] and even health care [6, 7].”
Matarese, M., Cocchella, F., Rea, F., & Sciutti, A. (2023) · IEEE RO-MAN · doi:10.1109/ro-man57019.2023.10309636 · Cited by 12
“On the other hand, we made iCub play via the AlphaZero deep neural network (DNN) [38] to obtain semi-perfect performance against the MC-TS agent.”
MethodPredicting Chess Player Rating Based on a Single Game
Tijhuis, T., Blom, P. M., & Spronck, P. H. M. (2023) · IEEE Conference on Games (CoG) · doi:10.1109/cog57401.2023.10333133 · Cited by 9
Yang, J., & Zhang, J. (2024) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2024.3379972 · Cited by 9
“DRLleveragesthecapacityofdeeplearningtorepresentfeaturesinconjunctionwithreinforcementlearning’sdecision-making capacity[1],therebyenablingnumerousgroundbreakingac-complishmentsinmanyfields,suchasboardgames[2],[3], trafficcontrol[4],[5]andautonomousdriving[6],[7].”
Kumar, D. (2024) · Neural Computing and Applications · doi:10.1007/s00521-024-09678-y · Cited by 10
“In this way, combining RL and IL techniques with the RIL algorithm can transcend limitations of each one, and results in better performance than what any of them is going to provide alone [3].”
InfluentialSolving optimal predictor-feedback control using approximate dynamic programming
Wang, H., Zhao, F., Zhang, Z., Xu, J., & Li, X. (2024) · Automatica · doi:10.1016/j.automatica.2024.111848 · Cited by 6
Learning Autonomous Navigation in Unmapped and Unknown Environments
He, N., Yang, Z. X., Bu, C., Fan, X., Wu, J., Sui, Y., & Que, W. (2024) · Sensors · doi:10.3390/s24185925 · Cited by 7
Learning to Move Like Professional Counter‐Strike Players
Durst, D., Xie, F., Sarukkai, V., Shacklett, B., Frosio, I., Tessler, C., Kim, J., Taylor, C., Bernstein, G., Choudhury, S., et al. (2024) · Computer Graphics Forum · doi:10.1111/cgf.15173 · Cited by 7
Real-Time Policy Optimization for UAV Swarms Based on Evolution Strategies
Chen, Z., Liu, H., & Liu, G. (2024) · Drones · doi:10.3390/drones8110619 · Cited by 7
Sequential memory improves sample and memory efficiency in episodic control
Freire, I. T., Amil, A. F., & Verschure, P. F. M. J. (2024) · Nature Machine Intelligence · doi:10.1038/s42256-024-00950-3 · Cited by 14
“…years by its ability to reach human-level performance in several domains that were traditionally considered hallmarks of human intelligence: for instance, beating world champions in games like chess and Go 1 or, more recently, in complex real-time multiplayer games like Starcraft 2 and DOTA 3 .”
Schütt, A., Huber, T., & André, E. (2024) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata62323.2024.10826087 · Cited by 6
“The first part is the position, which we encode as a binary tensor of size 18x8x8, with each of the 18 channels representing different piece types or board status, similar to [26].”
Chess Rating Estimation from Moves and Clock Times Using a CNN-LSTM
Omori, M., & Tadepalli, P. V. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-86585-5_1 · Cited by 14
“Maia [12] adapted AlphaZero’s residual Convolu-tional Neural Network (CNN) network [20], removing the search algorithm to predict moves based on rating bins, improving accuracy over a depth-limited Stockfish.”
MethodPrecision Systems Medicine: A Control Discovery Problem
An, G. C., & Day, J. (2019) · Elsevier eBooks · doi:10.1016/b978-0-12-801238-3.11513-2 · Cited by 7
Learning key steps to attack deep reinforcement learning agents
Yu, C.-M., Chen, M.-H., & Lin, H.-T. (2023) · Machine Learning · doi:10.1007/s10994-023-06318-9 · Cited by 7
“Deep RL agents have shown great success across various domains, such as achieving superhuman performance on games (Mnih et al., 2015; Silver et al., 2016, 2018), completing complex robotic tasks (Levine et al., 2016), optimizing patient treatments (Escandell-Montero et al., 2014; Raghu et al.,…”
BackgroundOptimal Planning of Emergency Communication Network Using Deep Reinforcement Learning
Yin, C., Yang, R., Zhu, W., Zou, X., & Zhang, J. (2020) · IEICE Transactions on Communications · doi:10.1587/transcom.2020ebp3061 · Cited by 6
Learning and Cognition in Financial Markets: A Paradigm Shift for Agent-Based Models
Lussange, J., Belianin, A. V., Bourgeois‐Gironde, S., & Gutkin, B. S. (2020) · Advances in intelligent systems and computing · doi:10.1007/978-3-030-55190-2_19 · Cited by 13
Kheradvar, A., Jafarkhani, H., Guy, T. S., & Finn, J. P. (2020) · Future Cardiology · doi:10.2217/fca-2020-0128 · Cited by 5
Manipulating the Distributions of Experience used for Self-Play Learning in Expert Iteration
Soemers, D. J. N. J., Piette, É., Stephenson, M., & Browne, C. (2020) · IEEE Conference on Games (CoG) · doi:10.1109/cog47356.2020.9231589 · Cited by 9
“Generating data (experience) as described above is the most common procedure, and has produced state-of-the-art results empirically [1], [3], but it is not certain that the optimal loss function is one that weights states s by p(s) as in Equation (9).”
BackgroundMethodInfluentialPlaying Catan with Cross-Dimensional Neural Network
Gendre, Q., & Kaneko, T. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-63833-7_49 · Cited by 6
“ation of a game, as 1, 0, -1 for win, draw, loss, respectively. Given that agents are not enhanced by game-specic knowledge, the agent as well as its opponent must start as random players. AlphaZero [9] begins by gathering game records of random players, then gradually updates the agent by their experiences and period…”
BackgroundA dual-memory architecture for reinforcement learning on neuromorphic platforms
Olin-Ammentorp, W., Sokolov, Y., & Bazhenov, M. (2021) · Neuromorphic Computing and Engineering · doi:10.1088/2634-4386/ac1a64 · Cited by 8
“With the addition of modern deep-learning techniques, RL has surpassed human performance on a variety of problems, including the classic games Chess et al [27].”
BackgroundXu, S., & Zou, W. (2021) · Chinese Optics Letters · doi:10.3788/col202119.082501 · Cited by 6
Combining Tree Search and Action Prediction for State-of-the-Art Performance in DouDiZhu
Zhang, Y., Yan, D., Shi, B., Fu, H., Fu, Q., Su, H., Zhu, J., & Chen, N. (2021) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2021/470 · Cited by 6
“AlphaZero achieves dramatic performance improvement in perfect-information games by combining MCTS and policy-value networks [Silver et al. , 2018].”
MethodAchieving Safe Deep Reinforcement Learning via Environment Comprehension Mechanism
Pai, P., Fei, Z., Liu, Q., Zhao, P., & Wu, W. (2021) · Chinese Journal of Electronics · doi:10.1049/cje.2021.07.025 · Cited by 6
RLCFR: Minimize counterfactual regret by deep reinforcement learning
Li, H., Wang, X., Jia, F., Wu, Y., Zhang, J., & Qi, S. (2021) · Expert Systems with Applications · doi:10.1016/j.eswa.2021.115953 · Cited by 9
Continuous Transition: Improving Sample Efficiency for Continuous Control Problems via MixUp
Lin, J., Huang, Z., Wang, K., Liang, X., Chen, W., & Lin, L. (2021) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48506.2021.9561137 · Cited by 16
“Deep reinforcement learning (RL), with high-capacity deep neural networks, has been applied to solve various complex decision-making tasks, including chess games [1], [2], video games [3], [4], etc.”
Selfish optimization and collective learning in populations
McAvoy, A., Mori, Y., & Plotkin, J. B. (2022) · Physica D Nonlinear Phenomena · doi:10.1016/j.physd.2022.133426 · Cited by 7
“Recent applications of artificial intelligence to games of Go [21] and chess [22], which are trained for competition against humans and other game engines using self-play, are striking examples of this program.”
BackgroundCazenave, T. (2021) · Communications in computer and information science · doi:10.1007/978-3-030-89453-5_5 · Cited by 6
“It is also a difficult game and its is more difficult for MCTS algorithms [Cazenave, 2016].”
BackgroundInfluentialPath-Analysis-Based Reinforcement Learning Algorithm for Imitation Filming
Dang, Y., Huang, C. X., Chen, P., Liang, R., Yang, X. Y., & Cheng, K.-T. T. (2022) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2022.3151463 · Cited by 12
“…tasks that are substantially difficult to be hard-coded or collect frame-level annotation, such as playing computer games [12], chess [13] and learning the autonomous driving policies [14] or navigate an embodied agent to carry out natural language instructions inside real 3D…”
BackgroundHow Do People Rank Multiple Mutant Agents?
Dodge, J. E., Anderson, A. A., Olson, M., Dikkala, R., & Burnett, M. (2022) · ACM International Conference on Intelligent User Interfaces (IUI) · doi:10.1145/3490099.3511115 · Cited by 10
“For example, AI researchers have long studied domains like Chess, Shougi, and Go, recently reaching performance exceeding the best humans [85, 86].”
BackgroundA Novel Approach to Solving Goal-Achieving Problems for Board Games
Shih, C.-C., Wu, T.-R., Wei, T.-H., et al. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i9.21278 · Cited by 6
Cekic, M., Bakiskan, C., & Madhow, U. (2022) · Software Impacts · doi:10.1016/j.simpa.2022.100347 · Cited by 6
MCTS With Refinement for Proposals Selection Games in Scene Understanding
Stekovic, S., Rad, M., Moradi, A., Fraundorfer, F., & Lepetit, V. (2022) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2022.3203729 · Cited by 6
“MCTSisastochasticalgorithmthatefficiently exploressearchtreesandhasbeenusedforexampleinAlphaGo andAlphaZerotoselectmoveswhenplayingGoorothergames withhighcombinatorials[9].”
Exploiting Abstract Symmetries in Reinforcement Learning for Complex Environments
Gupta, K., & Najjaran, H. (2022) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra46639.2022.9811652 · Cited by 8
Göttl, Q., Grimm, D. G., & Burger, J. (2022) · Computer-aided chemical engineering/Computer aided chemical engineering · doi:10.1016/b978-0-323-85159-6.50259-1 · Cited by 6
An Exoatmospheric Homing Guidance Law Based on Deep Q Network
Tang, J., Bai, Z., Liang, Y., Fan, Z., & Li, K. (2022) · International Journal of Aerospace Engineering · doi:10.1155/2022/1544670 · Cited by 5
Physics aware analytics for accurate state prediction of dynamical systems
Mandal, A., Tiwari, Y., Panigrahi, P. K., et al. (2022) · Chaos Solitons & Fractals · doi:10.1016/j.chaos.2022.112670 · Cited by 6
Ausin, M. S., Maniktala, M., Barnes, T. M., & Chi, M. (2022) · International Journal of Artificial Intelligence in Education · doi:10.1007/s40593-022-00312-3 · Cited by 8
“Since then, DRL has achieved success in various complex tasks, such as robotic control (Andrychowicz et al., 2018) and playing games, including Go (Silver et al., 2016), Chess/Shogi (Silver et al., 2018), and Starcraft II (Vinyals et al., 2019).”
Over-the-Horizon Air Combat Environment Modeling and Deep Reinforcement Learning Application
Wang, A., Zhao, S., Shi, Z., et al. (2022) · International Conference on Data-driven Optimization of Complex Systems (DOCS) · doi:10.1109/docs55193.2022.9967482 · Cited by 6
“[5] In our self-play method, games are generated by the agent of best performance in the past training and the latest agent.”
MethodThe Modest State of Learning, Sampling, and Verifying Strategies
Hartmanns, A., & Klauck, M. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19759-8_25 · Cited by 8
“Deep neural networks (NNs) are responsible for astounding advances across applications as diverse as image classification [76], natural language processing [67], and playing games [98].”
BackgroundChess Position Evaluation Using Radial Basis Function Neural Networks
Kagkas, D., Karamichailidou, D., & Alexandridis, A. (2023) · Complexity · doi:10.1155/2023/7143943 · Cited by 9
Warm-Starting Nested Rollout Policy Adaptation with Optimal Stopping
Dang, C., Bazgan, C., Cazenave, T., Chopin, M., & Wuillemin, P.-H. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i10.26459 · Cited by 8
“When combined with deep learning it surpasses the level of the best human players in games such as Go, Chess and Shogi (Silver et al. 2018).”
BackgroundInfluentialGros, T. P., Groß, J., Höller, D., Hoffmann, J., Klauck, M., Meerkamp, H., Müller, N. J., Schaller, L., & Wolf, V. (2023) · ACM Transactions on Modeling and Computer Simulation · doi:10.1145/3607198 · Cited by 9
“1 INTRODUCTION In recent years, neural networks (NN), especially deep neural networks, have accomplished major successes across many computer science domains, like image classiication [36], natural language processing [31], and gameplaying [55].”
Searching for answers: expert pattern recognition and planning
Gobet, F. R., & Waters, A. J. (2023) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2023.07.006 · Cited by 6
When Do Humans Heed AI Agents’ Advice? When Should They?
Dunning, R. E., Fischhoff, B., & Davis, A. L. (2023) · Human Factors The Journal of the Human Factors and Ergonomics Society · doi:10.1177/00187208231190459 · Cited by 16
“One potential resource for increasing trans-parency is the probability of success for possible actions that some algorithms calculate as part of their computational framework (Silver et al., 2018).”
BackgroundUnveiling Concepts Learned by a World-Class Chess-Playing Agent
Pálsson, A., & Björnsson, Y. (2023) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2023/541 · Cited by 15
Bio-Inspired Information Pathways: From Neuroscience to Neurotronics
Ziegler, M., Mussenbrock, T., & Kohlstedt, H. (2023) · Springer series on bio- and neurosystems · doi:10.1007/978-3-031-36705-2 · Cited by 10
Multiagent Rollout with Reshuffling for Warehouse Robots Path Planning
Emanuelsson, W., Riveiros, A. P., Li, Y., Johansson, K. H., & Mårtensson, J. (2023) · IFAC-PapersOnLine · doi:10.1016/j.ifacol.2023.10.1430 · Cited by 10
Respect the Difference: Reinforcement Learning for Heterogeneous FPGA Placement
Mahmoudi, F., Elgammal, M. A., Shahrouz, S. G., Murray, K. E., & Betz, V. (2023) · International Conference on Field-Programmable Technology (FPT) · doi:10.1109/icfpt59805.2023.00022 · Cited by 10
“Hence, it can optimize complex decision-making problems such as recommending websites [2], playing chess [3] and Go [4].”
BackgroundArtificial Intelligence for Climate Smart Forestry: A Forward Looking Vision
Luo, F., Liu, L., Wang, G. G., Kumar, V., Ashton, M. S., Abernethy, J., Afghah, F., Browning, M. H. E. M., Coyle, D. R., Dames, P., et al. (2023) · IEEE International Conference on Cognitive Machine Intelligence (CogMI) · doi:10.1109/cogmi58952.2023.00011 · Cited by 10
“To solve this complex, stochastic optimization problem, a deep reinforcement learning (DRL) [112, 113] based method will be useful to learn an optimal harvesting policy using information such as tree diversity and structure, harvest timing and rotation policy, tree growth modeling, carbon credit…”
MethodVision Transformers for Computer Go
Sagri, A., Cazenave, T., Arjonilla, J., & Saffidine, A. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-56855-8_23 · Cited by 6
Lussange, J., Vrizzi, S., Palminteri, S., & Gutkin, B. S. (2024) · PLoS ONE · doi:10.1371/journal.pone.0301141 · Cited by 10
Retrosynthesis Zero: Self-Improving Global Synthesis Planning Using Reinforcement Learning
Guo, J., Yu, C., Li, K., Zhang, Y., Wang, G., Li, S., & Dong, H. (2024) · Journal of Chemical Theory and Computation · doi:10.1021/acs.jctc.4c00071 · Cited by 11
Integrating White and Black Box Techniques for Interpretable Machine Learning
Vernon, E. M., Masuyama, N., & Nojima, Y. (2024) · Lecture notes in networks and systems · doi:10.1007/978-981-97-3562-4_51 · Cited by 10
Kotseruba, I., & Tsotsos, J. K. (2025) · Oxford University Press eBooks · doi:10.1093/oso/9780192844835.003.0005 · Cited by 6
A Deep Learning Approach for Probabilistic Security in Multi-Robot Teams
Wehbe, R., & Williams, R. K. (2019) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2019.2931238 · Cited by 6
“finding approximate solutions for notoriously difficult problems, such as the games of Go and Chess [14], and the traveling salesman problem [15].”
BackgroundFrom Reinforcement Learning Towards Artificial General Intelligence
Rocha, F. M., Costa, V. S., & Reis, L. P. (2020) · Advances in intelligent systems and computing · doi:10.1007/978-3-030-45691-7_37 · Cited by 7
Some new trends of intelligent simulation optimization and scheduling in intelligent manufacturing
Li, X., & Zhang, C. (2020) · Service Oriented Computing and Applications · doi:10.1007/s11761-020-00299-6 · Cited by 6
Considerations for Comparing Video Game AI Agents with Humans
Madan, C. R. (2020) · Challenges · doi:10.3390/challe11020018 · Cited by 8
A Study on Fight Against COVID-19 from Latest Technological Intervention
Arulprakash, E., & Aruldoss, M. (2020) · SN Computer Science · doi:10.1007/s42979-020-00301-0 · Cited by 7
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.