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 2,501–2,600 of 3,633
Fujita, N., & Watanabe, H. (2023) · Journal of the Physical Society of Japan · doi:10.7566/jpsj.92.034802 · Cited by 1
Application of Deep Reinforcement Learning in Guandan Game
Pan, J. H., Zhang, Z., Shen, H., Zeng, Y., & Wu, L. (2022) · Chinese Control and Decision Conference (CCDC) · doi:10.1109/ccdc55256.2022.10033565 · Cited by 1
The phenomenon of decision oscillation: A new consequence of pathology in game trees
Levene, M., & Fenner, T. I. (2023) · Computational Intelligence · doi:10.1111/coin.12570 · Cited by 1
“It is also worth mentioning that the newer Monte Carlo tree search (MCTS) algorithm 21,22 has proved to be competitive with the minimax algorithm when combined with reinforcement learning methods.”
MethodpH stablizing control of a neutral leaching process based on proximal policy optimization
Li, J., Li, Y., Sun, B., et al. (2022) · China Automation Congress (CAC) · doi:10.1109/cac57257.2022.10055264 · Cited by 1
“With the rapid development of artificial intelligence technology, deep reinforcement learning (DRL) has been widely concerned by industry and academia because of its success in solving some complex problems [9], [10].”
BackgroundDeep Reinforcement Learning for $$5 \times 5$$ Multiplayer Go
Driss, B., Arjonilla, J., Wang, H., Saffidine, A., & Cazenave, T. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-30229-9_48 · Cited by 1
“This great success has been achievable thanks to a combination of two key elements: Search (Monte Carlo Tree Search [1]) and Learning (Reinforcement Learning) methods [15,16,17].”
BackgroundMethodWhat medicine can learn from game-playing artificial intelligence: The 1 in 10,000 move
Chen, J. H., Dhaliwal, G., & Yang, D. S. (2023) · The American Journal of the Medical Sciences · doi:10.1016/j.amjms.2023.04.016 · Cited by 2
Improving Strategic Decisions in Sequential Games by Exploiting Positional Similarity
Evans, S., & Turrini, P. (2023) · Games · doi:10.3390/g14030036 · Cited by 2
“, [5,6]), have shown how limited human knowledge is when it comes to decision making in complex extensive games of perfect information.”
BackgroundArtificial Intelligence in Microbiology
Juhas, M. (2023) · Springer eBooks · doi:10.1007/978-3-031-29544-7_8 · Cited by 3
Rethinking the Physical Symbol Systems Hypothesis
Rosenbloom, P. S. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-33469-6_21 · Cited by 2
Delou, P. D. A., Bernardino, L. F., Capron, B. D. O., & Secchi, A. R. (2023) · Brazilian Journal of Chemical Engineering · doi:10.1007/s43153-023-00351-w · Cited by 1
Pigouvian algorithmic platform design
Norman, T. W. L. (2023) · Journal of Economic Behavior & Organization · doi:10.1016/j.jebo.2023.05.019 · Cited by 1
Żulicki, R. (2022) · Wydawnictwo Uniwersytetu Łódzkiego eBooks · doi:10.18778/8331-110-4 · Cited by 7
DM²: Decentralized Multi-Agent Reinforcement Learning via Distribution Matching
Wang, C., Durugkar, I., Liebman, E., & Stone, P. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i10.26382 · Cited by 2
“Multi-agent reinforcement learning (MARL) (Littman 1994) is a paradigm for learning agent policies that may interact with each other in cooperative or competitive settings (Silver et al. 2017, 2018; Barrett and Stone 2012; Leibo et al. 2017).”
BackgroundFunction Approximation for Solving Stackelberg Equilibrium in Large Perfect Information Games
Ling, C. K., Kolter, J. Z., & Fang, F. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i5.25715 · Cited by 1
Algorithm structure for real-time general game playing agents
Ge, C. (2023) · Applied and Computational Engineering · doi:10.54254/2755-2721/6/20230913 · Cited by 1
Fedorov, E., Nechyporenko, О., Chychuzhko, M., Chychuzhko, V., & Leshchenko, R. (2023) · RADIOELECTRONIC AND COMPUTER SYSTEMS · doi:10.32620/reks.2023.2.05 · Cited by 3
“The main directions of single-agent learning with reinforcement are: – dynamic programming [5, 6]; – adaptive dynamic programming [7, 8]; – Monte-Carlo [9, 10]; – Temporal-Difference Learning [11, 12]; – policy-based methods [13, 14]; – actor-critiс methods [15, 16].”
BackgroundHeterogeneous-graph Attention Reinforcement Learning for Football Matches
Wang, S., Pan, Y. F., Pu, Z., Yi, J., Liang, Y., & Zhang, D. (2023) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn54540.2023.10191648 · Cited by 1
Soft Ethics and the Governance of AI
Floridi, L. (2023) · Oxford University Press eBooks · doi:10.1093/oso/9780198883098.003.0006 · Cited by 1
Human Intelligence (HI) Versus Artificial Intelligence (AI) and Intelligence Augmentation (IA)
Makridakis, S., & Polemitis, A. (2023) · Palgrave advances in the economics of innovation and technology · doi:10.1007/978-3-031-35879-1_1 · Cited by 4
Goddard, Z. C., Rajasekar, R., Mocharla, M., Manaster, G., Williams, K., & Mazumdar, A. (2023) · Journal of Aerospace Information Systems · doi:10.2514/1.i011283 · Cited by 3
Parameter Identification for Fictitious Play Algorithm in Repeated Games
Dong, H., & Mu, Y. (2023) · Communications in computer and information science · doi:10.1007/978-981-99-7590-7_22 · Cited by 1
Self-Attention for Visual Reinforcement Learning
Fernandes, Z., Joseph, E., Vogel, D., & Si, M. (2023) · IEEE Conference on Games (CoG) · doi:10.1109/cog57401.2023.10333243 · Cited by 2
“Deep reinforcement learning (RL) has emerged as a popular approach for Game AI and has exhibited remarkable proficiency in several complex games, such as Dota 2 [1], StarCraft II [2], Stratego [3], Gran Turismo [4], and Chess [5].”
BackgroundInterpretation Using Classified Gradient-Based Saliency Maps for Two-Player Board Games
Nakasone, G., & Sugawara, T. (2023) · IEEE Conference on Games (CoG) · doi:10.1109/cog57401.2023.10333188 · Cited by 2
HiveMind: Learning to Play the Cooperative Chess Variant Bughouse with DNNs and MCTS
Woo, B., Sweetser, P., & Aitchison, M. (2023) · IEEE Conference on Games (CoG) · doi:10.1109/cog57401.2023.10333210 · Cited by 1
“As a comparison, CrazyAra was initially trained on 569,537 games from the free open-source chess server lichess 2 and AlphaZero was trained on 44 million games for 700,00 training steps [11].”
MethodAn Open-Source Robotic Chinese Chess Player
An, S., Che, G., Guo, J., Xu, Y., Wang, G., Tsintotas, K. A., Zhang, F., Ye, J., Fu, C., Zhu, H., et al. (2023) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros55552.2023.10341697 · Cited by 2
Akhtar, M. H., & Ramkumar, J. (2023) · Springer eBooks · doi:10.1007/978-981-99-6897-8_2 · Cited by 1
Kajizuka, Y., Motegi, K., Shiraishi, Y., et al. (2023) · Annual Conference of the Society of Instrument and Control Engineers (SICE) · doi:10.23919/sice59929.2023.10354172 · Cited by 1
Ship Path Planning Based on AlphaZero Algorithm
Liu, L., Wang, C., Teng, F., & Li, T. (2023) · International Conference on Information Science and Technology · doi:10.1109/icist59754.2023.10367178 · Cited by 1
Iida, H., Iida, H., Schaeffer, J., et al. (2024) · ICGA Journal · doi:10.3233/icg-230240 · Cited by 1
Understanding the Limits of Artificial Intelligence for Warfighters: Volume 4, Wargames
Geist, E., Frank, A., & Menthe, L. (2024) · RAND Corporation eBooks · doi:10.7249/rra1722-4 · Cited by 9
Reinforcement Learning for Turn-Based Strategy Game
Jonathan, G., & Maulidevi, N. U. (2023) · International Conference on Advanced Informatics: Concepts, Theory and Applications (ICAICTA) · doi:10.1109/icaicta59291.2023.10390086 · Cited by 1
Hu, H., Wu, W., Song, Y., et al. (2024) · Complex & Intelligent Systems · doi:10.1007/s40747-023-01312-z · Cited by 1
“…the subsequently developed deep reinforcement learning (DRL) have natural advantages in decision-making scenarios, and they have been widely investigated in fields like robot path planning [10, 11], unmanned aerial vehicle planning and control [12–14], games [15–17] and mobile edge computing [18].”
Understanding the Limits of Artificial Intelligence for Warfighters: Volume 5, Mission Planning
Scholl, K., Briggs, G. J., Zhang, L. A., & Salmon, J. L. (2024) · RAND Corporation eBooks · doi:10.7249/rra1722-5 · Cited by 1
Yuan, M., & Freris, N. M. (2023) · International Conference on Big Data Computing and Communications (BigCom) · doi:10.1109/bigcom61073.2023.00042 · Cited by 4
“Other examples include Shogi [6] and fighting game [7].”
BackgroundMaking Superhuman AI More Human in Chess
Barrish, D., Kroon, S., & van der Merwe, B. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-54968-7_1 · Cited by 1
UNeC: Unsupervised Exploring In Controllable Space
Xiong, X., Meng, L., Ruan, J., Xu, S., & Xu, B. (2024) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp48485.2024.10447360 · Cited by 1
“It has profoundly influenced the evolution of artificial intelligence, with notable applications spanning gaming [1, 2], robotic control [3], and autonomous driving [4].”
BackgroundSusanto, F., & Hamzah, S. (2022) · International Journal of Educational Research & Social Sciences · doi:10.51601/ijersc.v3i6.544 · Cited by 1
Human divergent exploration capacity for material design: A comparison with artificial intelligence
Sakai, H., Matsuda, K., Kikkawa, N., & Kajita, S. (2024) · Computers in Human Behavior Artificial Humans · doi:10.1016/j.chbah.2024.100064 · Cited by 4
Toward General Function Approximation in Nonstationary Reinforcement Learning
Feng, S., Yin, M., Huang, R., Wang, Y.-X., Yang, J., & Liang, Y. (2024) · IEEE Journal on Selected Areas in Information Theory · doi:10.1109/jsait.2024.3381818 · Cited by 6
“RL has witnessed a great success in practical applications, including robotics [2], [3], games [4], [5], [6], [7], and autonomous driving [8].”
BackgroundMisconduct in Post-Selections and Deep Learning
Weng, J. (2024) · International Conference on Control, Robotics and Cybernetics (CRC) · doi:10.1109/crc60659.2023.10488526 · Cited by 4
“Currently, well-known Deep Learning networks include Convolution Neural Networks (CNNs), such as AlexNet [12], AlphaGo Zero [23], AlphaZero [22], AlphaFold [19], MuZero [18], AlphaDev [14], those in IBM Debater [24] among others [25], [1], [13].”
BackgroundComparing Reinforcement Learning and Human Learning With the Game of Hidden Rules
Pulick, E., Meñkov, V., Mintz, Y. D., Kantor, P. B., & Bier, V. M. (2024) · IEEE Access · doi:10.1109/access.2024.3395249 · Cited by 4
“(*,*,*,[19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36],Nearby) \ (*,*,*,[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18],[0,1,2,3])”
BackgroundReinforcement learning for VR table tennis
Seah, H. S., Jiang, D., Tandianus, B., Sui, Y., & Wang, H. (2024) · International Workshop on Advanced Imaging Technology (IWAIT) · doi:10.1117/12.3018124 · Cited by 2
“However, the number of video games utilizing machine learning is limited as most video games with AI trained with machine learning are in board games such as chess and Go [1, 2].”
BackgroundAdvance Chess Engine: an use of ML Approach
Srivastava, K., Pandey, T. N., Dash, B. B., et al. (2024) · doi:10.1109/inocon60754.2024.10511742 · Cited by 1
“Authors in [6] introduces AlphaZero program by reinforcement learning from self-play.”
BackgroundA MZI-based optical neural network for image classification
Zhang, Y., Zhang, S., Zhang, D., Wang, R., Su, Y., Yi, J., Wang, P., Luo, G., Zhou, X., & Pan, J. (2024) · Optica Applicata · doi:10.37190/oa240108 · Cited by 1
Deep Reinforcement Learning Applied in Distribution Network Control and Optimization
Han, X. H., Li, Y., Li, J., Zhang, B., & Ma, Z. (2023) · IEEE Conference on Energy Internet and Energy System Integration (EI2) · doi:10.1109/ei259745.2023.10512736 · Cited by 2
“The search method is mainly completed through Monte Carlo Tree Search (MCTS), with classic examples being the success of AlphaGo and AlphaZero in chess game [32] .”
MethodA Brief Overview of Machine Learning
Škrlj, B. (2024) · SpringerBriefs in computer science · doi:10.1007/978-3-031-57016-2_1 · Cited by 1
Working with machines in mathematics
Davies, A. (2024) · Bulletin of the American Mathematical Society · doi:10.1090/bull/1843 · Cited by 3
UNSAT Solver Synthesis via Monte Carlo Forest Search
Cameron, C., Hartford, J., Lundy, T., Truong, T., Milligan, A., Chen, R., & Leyton‐Brown, K. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-60597-0_12 · Cited by 6
“Silver et al. [66, 67] took the world by storm when their AlphaGo system beat world champion Lee Sedol at Go, marking the first time a computer program had achieved su-perhuman performance on a game with such a large action space.”
BackgroundLiu, R., & Lei, H. (2023) · International Conference on Big Data, Artificial Intelligence and Risk Management (ICBAR) · doi:10.1145/3656766.3656909 · Cited by 1
Thorstad, D. (2024) · Oxford University Press eBooks · doi:10.1093/oso/9780198886143.003.0003 · Cited by 2
Hierarchical deep reinforcement learning for hydrogen supply chain management
Song, G., Khaligh, V., Liu, J. J., et al. (2024) · Computer-aided chemical engineering/Computer aided chemical engineering · doi:10.1016/b978-0-443-28824-1.50485-3 · Cited by 1
New Lines of Business Development: Artificial Intelligence in Business
Gandía, J. A. G., González‐Tejero, C. B., & Miguel, Á. J. Á. (2024) · Contributions to management science · doi:10.1007/978-3-031-58704-7_1 · Cited by 1
IoT and machine learning for management of diabetes mellitus
Priyadarshini, A., & Yogesh, Y. (2024) · Elsevier eBooks · doi:10.1016/b978-0-323-95686-4.00007-1 · Cited by 1
ELO-Rated Sequence Rewards: Advancing Reinforcement Learning Models
Ju, Q., Hei, F., Fang, Z., & Luo, Y. (2024) · IEEE Data Driven Control and Learning Systems Conference (DDCLS) · doi:10.1109/ddcls61622.2024.10606606 · Cited by 2
“Deep Reinforcement Learning (RL) has achieved remarkable progress in various fields, including Atari games [7], Go [2, 5], poker [12], and video gaming [4, 6].”
A Data-Driven Optimization Method for Strongly Non-Separable Mixed-Integer Problems
Satô, T. (2024) · IEEE Congress on Evolutionary Computation (CEC) · doi:10.1109/cec60901.2024.10611978 · Cited by 1
“Using a tree structure, we propose a new data-driven optimization method combining reinforcement learning for tree structures [10], [11] and evolutionary algorithms.”
MethodPerfect Information Monte Carlo with Postponing Reasoning
Arjonilla, J., Saffidine, A., & Cazenave, T. (2024) · IEEE Conference on Games (CoG) · doi:10.1109/cog60054.2024.10645574 · Cited by 1
“Perfect information games, like Chess and Go, where all information is available, have been extensively studied, allowing algorithms to surpass human professionals [1], [5], [6].”
BackgroundMethodAlphaZero como ferramenta de playtest
Malosto, C. G. D. A., Knop, I. D. O., & Campos, L. C. D. (2023) · Revista ComInG - Communications and Innovations Gazette · doi:10.5902/2448190485269 · Cited by 1
Initial state diversification for efficient AlphaZero-style training
Demura, Y., & Kaneko, T. (2024) · ICGA Journal · doi:10.3233/icg-240255 · Cited by 2
“AlphaZero (Silver et al., 2018), a model-based RL algorithm, has outperformed state-of-the-art computer programs in chess, shogi, and the game of Go.”
InfluentialA model-adaptive random search actor critic: convergence analysis and inventory-control case studies
Luo, Y., Hu, J., & Gosavi, A. (2024) · Annals of Operations Research · doi:10.1007/s10479-024-06284-y · Cited by 2
When Rewards Deceive: Counteracting Reward Poisoning on Online Deep Reinforcement Learning
Bouhaddi, M., & Adi, K. (2024) · IEEE International Conference on Cyber Security and Resilience (CSR) · doi:10.1109/csr61664.2024.10679438 · Cited by 1
“This technique has successfully powered applications like playing Go [18] and mastering Atari games [13].”
MethodA Simplified Query-Only Attention for Encoder-Based Transformer Models
Yeom, H. G., & An, K. (2024) · Applied Sciences · doi:10.3390/app14198646 · Cited by 2
Artificial Intelligence and Quantum Computing
Wirtz, B. W. (2024) · Springer texts in business and economics · doi:10.1007/978-3-658-44675-8_13 · Cited by 1
De Novo Drug Design by Multi-Objective Path Consistency Learning With Beam A* Search
Zhao, D., Zhou, J., Tu, S., & Xu, L. (2024) · IEEE Transactions on Computational Biology and Bioinformatics · doi:10.1109/tcbb.2024.3477592 · Cited by 2
“PCZero [75], [76] has demonstrated the effectiveness of incorporating PC as a supplementary loss in AlphaZero [55] and MuZero [53], [72].”
Enhancing Reinforcement Learning Through Guided Search
Arjonilla, J., Saffidine, A., & Cazenave, T. (2024) · Frontiers in artificial intelligence and applications · doi:10.3233/faia240696 · Cited by 1
“In recent years, MCTS has been integrated with offline neural networks to boost its performance [41–43], however, in most scenarios, neural networks are employed to predict the outcomes generated by MCTS. Toourknowledge, no prior work has attempted to utilize MCTS as a guiding expert while…”
BackgroundMethodModel-based reinforcement learning experimental study for mobile robot navigation
Dobriborsci, D., Chichkanov, I., Zashchitin, R., et al. (2024) · International Conference on Control, Decision and Information Technologies (CoDIT) · doi:10.1109/codit62066.2024.10708080 · Cited by 1
“Over the past decade, reinforcement learning (RL) has made significant strides across a spectrum of tasks and competitions, ranging from video games [1], [2].”
BackgroundZhu, H., Zhong, H., Yan, B., & Li, X. (2024) · International Conference on Advanced Robotics and Mechatronics (ICARM) · doi:10.1109/icarm62033.2024.10715990 · Cited by 1
“In order to acquire a comprehensive defensive strategy, a technique called self-play [25], [26] is used by introducing an offensive agent to help the training of defensive strategy.”
MethodEnhancing Safety and Efficiency through Explainable Transfer Learning
Guo, Y., & Sycara, K. P. (2024) · Artificial intelligence · doi:10.5772/intechopen.115536 · Cited by 1
Variation and evolution analysis of SARS-CoV-2 using self-game sequence optimization
Liu, Z., Shen, Y., Jiang, Y., Zhu, H., Hu, H., Kang, Y., Chen, M., & Li, Z. (2024) · Frontiers in Microbiology · doi:10.3389/fmicb.2024.1485748 · Cited by 2
Safe Reinforcement Learning Through Regret and State Restorations in Evaluation Stages
Gros, T. P., Müller, N. J., Höller, D., & Wolf, V. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-75778-5_2 · Cited by 4
“Notably, it has demonstrated exceptional success in domains such as board games, including chess and go (Silver et al. 2016, 2018, 2017).”
BackgroundArtificial Intelligence and Inequality
Liso, N. D. (2024) · Springer eBooks · doi:10.1007/978-3-031-32152-8_49 · Cited by 2
Künstliche Intelligenz in der Proteinforschung: Nobelpreis für Chemie
Welter, K. (2024) · Chemie in unserer Zeit · doi:10.1002/ciuz.202400038 · Cited by 1
Unveiling Computer Chess Evolution: Can Machine Learning Detect Historical Trends?
Borghesi, A., Ciancarini, P., Iorio, A. D., et al. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-74353-5_3 · Cited by 1
Solving Min-Max MTSP in a Reinforcement Learning context
Buinoschi-Tirpescu, A., & Breaban, M. E. (2024) · Procedia Computer Science · doi:10.1016/j.procs.2024.09.519 · Cited by 2
Gao, Y., Kong, D., Dai, H., et al. (2024) · World Wide Web · doi:10.1007/s11280-024-01325-9 · Cited by 1
Santina, C. D., Hernández, C., Şişman, B., Leiva, L. A., Arapakis, I., Vakalellis, M., Vanderdonckt, J. M., D’Haro, L. F., Manzi, G., Becchio, C., et al. (2024) · Springer proceedings in advanced robotics · doi:10.1007/978-3-031-76424-0_20 · Cited by 7
An efficient heuristic search algorithm for discovering large Condorcet domains
Zhou, B., & Riis, S. K. (2025) · 4OR · doi:10.1007/s10288-025-00583-1 · Cited by 1
“…AlphaZero-style algorithms, one of the deep reinforcement learning algorithms, has achieved remarkable success on board games like Chess and Go by utilizing a value function 2 that evaluates the current board position without waiting for the game’s completion, as described by Silver et al. (2018).”
Deep Reinforcement Learning-based Training for Virtual Table Tennis Agent
Jiang, D., Seah, H. S., Tandianus, B., Sui, Y., & Wang, H. (2024) · International Conference on Virtual Reality · doi:10.1109/icvr62393.2024.10867925 · Cited by 1
“In the field of table tennis, recognized for its competitive multi-agent nature, the training of two agents employs a self-play algorithm [13].”
Wang, S., Li, K., Zhang, T., Zhang, Z., & Hu, Z. (2024) · China Automation Congress (CAC) · doi:10.1109/cac63892.2024.10864643 · Cited by 1
A Novel Decision-Making Model for Playing Board Game Combining Planning and Opponent Behaviors
Zhang, J., Jiang, J., Li, L., et al. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10889240 · Cited by 1
Reinforcement learning model, algorithms, and its application
Kamble, A. R., Jiet, M. M., Weerarathna, I. N., Dhande, S., Izankar, S. V., & Puri, C. (2025) · AIP conference proceedings · doi:10.1063/5.0241925 · Cited by 3
Doya, K. (2025) · Elsevier eBooks · doi:10.1016/b978-0-443-15754-7.00088-2 · Cited by 1
Data-efficient reinforcement learning by generalized value estimation
Zhou, J., Tian, Y., & Ren, M. (2025) · Machine Learning · doi:10.1007/s10994-025-06763-8 · Cited by 2
“Reinforcement Learning (RL) (Sutton & Barto, 2018) has made empirical successes in the virtual world, from games (Mnih et al., 2015), chess (Silver et al., 2018), to simulated robotic control (Schulman et al., 2015; Fujimoto et al., 2018; Haarnoja et al., 2018).”
ASP-Driven Emergency Planning for Norm Violations in Reinforcement Learning
Adam, S., & Eiter, T. (2025) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v39i14.33619 · Cited by 2
A Variational Inequality Approach to Independent Learning in Static Mean-Field Games
Yardim, B., Çaycı, S., & He, N. (2025) · ACM / IMS Journal of Data Science · doi:10.1145/3731553 · Cited by 1
“Multi-agent reinforcement learning ( MARL ) has been an active area of research, with a vast range of successful applications in games such as Chess, Shogi [63], Go [64], Stratego [51], as well as real-world applications for instance in robotics [41].”
Methods for Identifying Implicit Communities and Analyzing Social Networks Using AI
Komardin, M. A., Britvin, R. M., & Panilov, P. A. (2025) · International Russian Smart Industry Conference (SmartIndustryCon) · doi:10.1109/smartindustrycon65166.2025.10985967 · Cited by 1
Reinforcement learning in path-dependent, non-ergodic contexts
Verbruggen, B., Vanhoyweghen, A., & Ginis, V. (2025) · Proceedings of the Royal Society A Mathematical Physical and Engineering Sciences · doi:10.1098/rspa.2024.0415 · Cited by 1
A Reinforcement Learning Based Approach to Solve Voltage Issues in Distribution Networks
Çakır, M. T., Nayır, H., Demіr, A., Kaya, H., & Ceylan, O. (2025) · IEEE International Conference on Compatibility, Power Electronics and Power Engineering · doi:10.1109/cpe-powereng63314.2025.11027244 · Cited by 1
“In gaming, RL has been pivotal in developing AI agents capable of mastering complex games such as Chess, Go, and video games, often exceeding human performance [14].”
Bus Loop Scheduling with Dueling Double Deep Q Network
Pradana, A., & Chew, L. Y. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-97626-1_3 · Cited by 1
Algorithm Configuration in Sequential Decision-Making
Begnardi, L., von Meijenfeldt, B., Zhang, Y., et al. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-95973-8_6 · Cited by 1
GPI-tree search: algorithms for decision-time planning with the general policy improvement theorem
Bagot, L., D’eer, L., Latré, S., Schepper, T. D., & Mets, K. (2025) · Neural Computing and Applications · doi:10.1007/s00521-025-11304-4 · Cited by 1
Schwartz, J., Newbury, R., Kulić, D., & Kurniawati, H. (2025) · Autonomous Agents and Multi-Agent Systems · doi:10.1007/s10458-025-09716-6 · Cited by 3
“This finding aligns with previous research on combining search and RL in 2-player, zero-sum games [15, 16, 18] and cooperative games [19, 89].”
InfluentialEvolving Macro-Actions for MCTS Video Game Agents
Gourley, A., Dann, M., Li, X., et al. (2025) · Genetic and Evolutionary Computation Conference Companion (GECCO Companion) · doi:10.1145/3712255.3726551 · Cited by 1
“Tree search based methods, particularly Monte Carlo Tree Search (MCTS) [2, 6], have been incredibly successful when applied to certain problems, notably mastering Chess and Go [13].”
MARLadona - Towards Cooperative Team Play Using Multi-Agent Reinforcement Learning
Li, Z., Bjelonic, F., Klemm, V., & Hutter, M. (2025) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra55743.2025.11128256 · Cited by 11
“Self-play is another approach commonly applied in competitive RL problems [29], [7], [8], [9], [30].”
de Carvalho, G. J. A., de Carvalho, J. A., Roesler, H., & Borges, H. R. (2025) · Caderno Pedagógico · doi:10.54033/cadpedv22n12-136 · Cited by 1
The Cost of Uncertainty in Self-Play Reinforcement Learning and Search
Oscarsson, M., Brynielsson, J., Cohen, M., Kamrani, F., & Limér, C. (2025) · IEEE International Conference on Intelligence and Security Informatics (ISI) · doi:10.1109/isi65680.2025.11201174 · Cited by 1
“While these milestones in poker were based on look-ahead search alone, the strongest AIs for perfect information games use a combination of reinforcement learning and look-ahead search introduced in AlphaGo and later generalized in AlphaZero [11].”
Multi-Source Feature Fusion and Neural Embedding for Predicting Chess Puzzle Difficulty
Xiao, H., Yu, D., Wen, X., Chen, L., & Fu, K. (2025) · Annals of Computer Science and Information Systems · doi:10.15439/2025f2456 · Cited by 1
“From Deep Blue’s brute-force victory over Kasparov [1] to AlphaZero’s self-taught superhuman play [2], each algorithmic milestone has advanced a deeper goal: understanding how machines can model, anticipate, and ultimately support human cognitive behavior.”
Artificial intelligence transforming materials research: possible road ahead for India
Sen, P. (2025) · Proceedings of the Indian National Science Academy · doi:10.1007/s43538-025-00641-6 · Cited by 1
Pursuit-Evasion for Car-like Robots with Sensor Constraints
Gonultas, B. M., & Isler, V. (2025) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros60139.2025.11246244 · Cited by 5
Optimal Use of Experience in First Person Shooter Environments
Aitchison, M. (2019) · IEEE Conference on Games (CoG) · doi:10.1109/cig.2019.8848049 · Cited by 1
“Reinforcement learning (RL) in games has gained a lot of attention recently due to some high-profile successes such as super-human performance in the Atari games [1], StarCraft [2], and the game of Go [3]. However, these algorithms require an extraordinary amount of experience to train the agents. AlphaZero, for exampl…”
BackgroundRankNet for evaluation functions of the game of Go
Mandai, Y., & Kaneko, T. (2019) · ICGA Journal · doi:10.3233/icg-190108 · Cited by 1
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.