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 1,201–1,300 of 3,633
Optimal Control in Soft and Active Matter
Alvarado, J., Teich, E. G., Sivak, D. A., & Bechhoefer, J. (2026) · Annual Review of Condensed Matter Physics · doi:10.1146/annurev-conmatphys-031324-031350 · Cited by 42
Hügle, T. (2022) · Annals of the Rheumatic Diseases · doi:10.1136/annrheumdis-2022-222141 · Cited by 10
Sample Efficient Reinforcement Learning Using Graph-Based Memory Reconstruction
Kang, Y., Zhao, E., Zang, Y., Li, L., Li, K., Tao, P., & Xing, J. (2023) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2023.3268612 · Cited by 16
“R EINFORCEMENT learning (RL) agents have achieved superhuman performances in various game environments, such as Atari 2600 games [1], [2], Go [3], [4], [5], and Star-Craft [6].”
Reinforcement learning based agents for improving layouts of automotive crash structures
Trilling, J., Schumacher, A., & Zhou, M. (2024) · Applied Intelligence · doi:10.1007/s10489-024-05276-6 · Cited by 13
A Recursively Recurrent Neural Network (R2N2) Architecture for Learning Iterative Algorithms
Doncevic, D. T., Mitsos, A., Guo, Y., Li, Q., Dietrich, F., Dahmen, M., & Kevrekidis, I. G. (2024) · SIAM Journal on Scientific Computing · doi:10.1137/22m1535310 · Cited by 22
Developing Rapport between Humans and Machines: Emotionally Intelligent AI Assistants
Agrawal, R., & Pandey, N. (2024) · International Journal for Research in Applied Science and Engineering Technology · doi:10.22214/ijraset.2024.59015 · Cited by 11
Tang, Y. F., Guo, S., Liu, J., Wan, B., An, L., & Liu, J. K. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.111736 · Cited by 11
Application of reinforcement learning to deduce nuclear power plant severe accident scenario
Song, S. H., Lee, Y., Bae, J. Y., Song, K. S., Seo, M. R., Kim, S. J., & Lee, J. I. (2024) · Annals of Nuclear Energy · doi:10.1016/j.anucene.2024.110605 · Cited by 11
Evolutionary reinforcement learning with action sequence search for imperfect information games
Wu, X., Zhu, Q., Chen, W., Lin, Q., Li, J., & Coello, C. A. C. (2024) · Information Sciences · doi:10.1016/j.ins.2024.120804 · Cited by 10
Efficient Deep Q-Learning for Industrial Equipment Calibration in Elevator Manufacturing
Evangelidis, A., Dimitriou, N., Charalampous, P., Mastos, T. D., & Tzovaras, D. K. (2024) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2024.3417254 · Cited by 13
“Even though it has been successfully applied to solve closed-environment problems with finite state spaces, such as chess [3], and go [4], there still exist several challenges when approaching more complex real-world problems that need to be addressed.”
BackgroundOptimizing ZX-diagrams with deep reinforcement learning
Nägele, M., & Marquardt, F. (2024) · Machine Learning Science and Technology · doi:10.1088/2632-2153/ad76f7 · Cited by 25
“RL has been successfully applied to various domains such as game-playing [12, 13], robotics [14, 15], quantum chemistry [16, 17], and problems in quantum computing like quantum error correction [18–20], quantum control [21, 22], and circuit optimization [23, 24].”
MethodXu, L., Li, C., Gao, S., Zhao, L., Guan, C., Shen, X., Zhu, Z., Guo, C., Zhang, L., Yang, C., et al. (2024) · JMIR Medical Informatics · doi:10.2196/52837 · Cited by 13
“Recent, noteworthy strides in machine learning (ML) have given rise to remarkable breakthroughs, encompassing fields like autonomous driving, recommending products, and surpassing human expertise in intricate games such as chess [10-12].”
BackgroundCatalytic evolution of cooperation in a population with behavioral bimodality
Sheng, A., Zhang, J., Zheng, G., Zhang, J., Cai, W., & Chen, L. (2024) · Chaos An Interdisciplinary Journal of Nonlinear Science · doi:10.1063/5.0231772 · Cited by 11
Towards Understanding AI Delegation: The Role of Self-Efficacy and Visual Processing Ability
Westphal, M., Hemmer, P., Vössing, M., Schemmer, M., Vetter, S., & Satzger, G. (2024) · ACM Transactions on Interactive Intelligent Systems · doi:10.1145/3696423 · Cited by 16
“Nowadays, the performance of AI models is similar to, and in certain application areas even exceeds the performance of human experts [16, 41, 91].”
Jang, K., Lichtlé, N., Vinitsky, E., Shah, A., Bunting, M., Nice, M. W., Piccoli, B., Seibold, B., Work, D. B., Monache, M. L. D., et al. (2025) · IEEE Control Systems · doi:10.1109/mcs.2024.3503372 · Cited by 20
Three systems of circuit formation: assembly, updating and tuning
Barabási, D. L., Castro, A. F., & Engert, F. (2025) · Nature reviews. Neuroscience · doi:10.1038/s41583-025-00910-9 · Cited by 25
“The success of artificial intelligence, particularly in its ability to surpass human experts at complex games such as chess and Go 92 , has reinforced the idea that learning from experience alone can lead to sophisticated behaviour in models with minimal initial structure.”
Achieving Fairness Generalizability for Learning-based Congestion Control with Jury
Han, T., Liao, X., Sun, D., Zeng, C., Jin, Y., Zhang, J., Wan, X., Wang, Z., Wang, Y. J., & Chen, K. (2025) · European Conference on Computer Systems · doi:10.1145/3689031.3696065 · Cited by 13
Deep Learning Analysis for Split-Manufactured Layouts With Routing Perturbation
Li, H., Patnaik, S., Ashraf, M., et al. (2020) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2020.3037297 · Cited by 10
Self-controlling photonic-on-chip networks with deep reinforcement learning
Do, N., Truong, D., Nguyen, D. H. N., Hoai, M., & Pham, C. (2021) · Scientific Reports · doi:10.1038/s41598-021-02583-7 · Cited by 15
Robot arm navigation using deep deterministic policy gradient algorithms
Farag, W. (2022) · Journal of Experimental & Theoretical Artificial Intelligence · doi:10.1080/0952813x.2021.1960640 · Cited by 11
“Moreover, reinforcement learning has been developed further to beat human-level performance in board games like Go (Silver et al., 2016) and Chess (Silver, et al., 2018).”
BackgroundMachine learning in scanning transmission electron microscopy
(2022) · Nature Reviews Methods Primers · doi:10.1038/s43586-022-00107-9 · Cited by 11
A MADDPG-based multi-agent antagonistic algorithm for sea battlefield confrontation
Chen, W., & Nie, J. (2022) · Multimedia Systems · doi:10.1007/s00530-022-00922-w · Cited by 13
“In recent years, reinforcement learning has conquered complete information games such as go [3, 4], chess [5] and Gobang [6], as well as incomplete information games such as poker.”
BackgroundPathfinding in stochastic environments: learning vs planning
Skrynnik, A., Андреевич, А. А., Yakovlev, K., & Panov, A. I. (2022) · PeerJ Computer Science · doi:10.7717/peerj-cs.1056 · Cited by 13
An SMDP approach for Reinforcement Learning in HPC cluster schedulers
Cunha, R. L. F., & Chaimowicz, L. (2022) · Future Generation Computer Systems · doi:10.1016/j.future.2022.09.025 · Cited by 10
Artificial intelligence development in Islamic System of Governance: a literature review
Shamdi, W., Lai, D. T. C., Aziz, A. A., & Anshari, M. (2022) · Contemporary Islam · doi:10.1007/s11562-022-00504-7 · Cited by 21
“Following the success of AlphaGo, DeepMind built another programme called AlphaZero that can play different types of board games, indicating the other significance of DeepMind that demonstrates the potential generalisability of the DL programme (Silver et al., 2018).”
MethodDeep Reinforcement Learning for Preparation of Thermal and Prethermal Quantum States
Baba, S., Yoshioka, N., Ashida, Y., & Sagawa, T. (2023) · Physical Review Applied · doi:10.1103/physrevapplied.19.014068 · Cited by 11
Adversarial agent-learning for cybersecurity: a comparison of algorithms
Shashkov, A., Hemberg, E., Tulla, M., & O’Reilly, U.-M. (2023) · The Knowledge Engineering Review · doi:10.1017/s0269888923000012 · Cited by 14
“We evaluate AlphaZero, an algorithm that trains its policy with DRL using tree search.”
BackgroundMethodInfluentialLearning Audio and Video Bitrate Selection Strategies via Explicit Requirements
Li, W., Huang, J., Liu, J., Jiang, W., & Wang, J. (2023) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2023.3265380 · Cited by 10
“(cid:2) To effectively integrate audio and video bitrate adaption across diverse network environments, we leverage self-play reinforcement learning [17], [18], [19] to obtain the bitrate selection strategy via learning from the explicit rule rather than a fixed-weight linear QoE function.”
BackgroundCatellani, P., Biella, M., Carfora, V., Nardone, A., Brischigiaro, L., Manera, M. R., & Piastra, M. (2023) · Frontiers in Psychology · doi:10.3389/fpsyg.2023.1200304 · Cited by 13
On-the-fly Raman microscopy guaranteeing the accuracy of discrimination
Tabata, K., Kawagoe, H., Taylor, J. N., Mochizuki, K., Kubo, T., Clément, J.-E., Kumamoto, Y., Harada, Y., Nakamura, A., Fujita, K., et al. (2024) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2304866121 · Cited by 9
Comparing search algorithms on the retrosynthesis problem
Roucairol, M., & Cazenave, T. (2024) · Molecular Informatics · doi:10.1002/minf.202300259 · Cited by 12
Discovery of optimal quantum codes via reinforcement learning
Su, V. P., Cao, C., Hu, H.-Y., Yanay, Y., Tahan, C., & Swingle, B. G. (2025) · Physical Review Applied · doi:10.1103/physrevapplied.23.034048 · Cited by 43
Exploring Deep Reinforcement Learning for Autonomous Powerline Tracking
Pienroj, P., Schönborn, S., & Birke, R. (2019) · IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS) · doi:10.1109/infcomw.2019.8845212 · Cited by 12
Engineered self-organization for resilient robot self-assembly with minimal surprise
Kaiser, T. K., & Hamann, H. (2019) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2019.103293 · Cited by 12
“However, the automatic generation of robot controllers for complex tasks and even more for collective robot systems is still challenging [6], despite recent improvements in machine learning [7, 8].”
BackgroundYuan, K., McGreavy, C., Yang, C., Wolfslag, W. J., & Li, Z. (2020) · IEEE Robotics & Automation Magazine · doi:10.1109/mra.2020.2980547 · Cited by 11
“The idea of using AI-generated policies as inspiration for solving hard problems has taken flight in other fields, most notably in games such as Go, Chess, and Shogi [9] for which the AI policy achieved superhuman-level performance.”
BackgroundData Fusion-Link Prediction for Evolutionary Network with Deep Reinforcement Learning
Lim, M., Jhanjhi, N. Z., & Abdullah, A. B. (2020) · International Journal of Advanced Computer Science and Applications · doi:10.14569/ijacsa.2020.0110644 · Cited by 10
Eswaran, S., Sachdeva, M., Vimal, V., Seth, D., Kalpam, S., Agarwal, S., Mukherjee, T., & DattaGupta, S. (2020) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3394486.3403316 · Cited by 12
“Most of the literature in the game AI domain focuses on building agents that can learn and predict next moves and effectively play the game as skillfully as or better than a human, utilizing techniques such as Monte Carlo Tree Search and deep reinforcement learning for deterministic games such as Chess, Go, Atari, and…”
MethodZha, Z., Wang, B. Y., & Tang, X. (2021) · Neural Computing and Applications · doi:10.1007/s00521-020-05663-3 · Cited by 11
Lutter, M., & Weidner, L. (2021) · Poetics · doi:10.1016/j.poetic.2021.101535 · Cited by 17
A Preliminary Study on the Application of Reinforcement Learning for Predictive Process Monitoring
Chiorrini, A., Diamantini, C., Mircoli, A., & Potena, D. (2021) · Lecture notes in business information processing · doi:10.1007/978-3-030-72693-5_10 · Cited by 10
Collective intelligence evolution using ant colony optimization and neural networks
Qi, X., Gan, Z., Liu, C., Xu, Z., Zhang, X., Li, W., & Ouyang, C. (2021) · Neural Computing and Applications · doi:10.1007/s00521-021-05918-7 · Cited by 11
Leveraging Granularity: Hierarchical Reinforcement Learning for Pedagogical Policy Induction
Zhou, G., Azizsoltani, H., Ausin, M. S., Barnes, T. M., & Chi, M. (2021) · International Journal of Artificial Intelligence in Education · doi:10.1007/s40593-021-00269-9 · Cited by 15
Yapay Zekâ ve Özgür İrade: Yapay Özgür İradenin İmkânı
DOĞAN, M. (2021) · TRT Akademi · doi:10.37679/trta.969912 · Cited by 20
Maximum-Entropy Progressive State Aggregation for Reinforcement Learning
Mavridis, C. N., Suriyarachchi, N., & Baras, J. S. (2021) · Proceedings of the IEEE Conference on Decision & Control, including the Symposium on Adaptive Processes · doi:10.1109/cdc45484.2021.9682927 · Cited by 10
“Reinforcement learning algorithms are being extensively studied, not only due to their effectiveness in numerous applications [1], [2], but also due to their promise to solve difficult optimal control problems in an online and data-driven fashion.”
BackgroundHuman in the Loop Fuzzy Pattern Tree Evolution
Murphy, A., Murphy, G., Dias, D. M., Amaral, J. L. M., Naredo, E., & Ryan, C. (2022) · SN Computer Science · doi:10.1007/s42979-022-01044-w · Cited by 10
“This has produced AI success stories in chess and Go [39].”
BackgroundControl with adaptive Q-learning: A comparison for two classical control problems
Araújo, J. P., Figueiredo, M. A. T., & Botto, M. A. (2022) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2022.104797 · Cited by 16
Backprop-Free Reinforcement Learning with Active Neural Generative Coding
Ororbia, A. G., & Mali, A. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i1.19876 · Cited by 25
Conformer‐RL : A deep reinforcement learning library for conformer generation
Jiang, R., Gogineni, T., Kammeraad, J. A., et al. (2022) · Journal of Computational Chemistry · doi:10.1002/jcc.26984 · Cited by 10
Saini, G. S., Pournazari, P., Ashok, P., & van Oort, E. (2022) · Energies · doi:10.3390/en15155749 · Cited by 12
Photonic Bayesian Neural Network Using Programmed Optical Noises
Wu, C., Yang, X., Chen, Y., & Li, M. (2022) · IEEE Journal of Selected Topics in Quantum Electronics · doi:10.1109/jstqe.2022.3217819 · Cited by 18
Bhatia, A., Svegliato, J., Nashed, S. B., & Zilberstein, S. (2022) · Proceedings of the International Conference on Automated Planning and Scheduling · doi:10.1609/icaps.v32i1.19842 · Cited by 10
“Generally, deep RL has been effective across a variety of applications, such as Atari (Mnih et al. 2015), chess (Silver et al. 2018), and StarCraft (Vinyals et al. 2019).”
BackgroundKorecki, M., Dailisan, D. N., & Helbing, D. (2023) · IEEE Access · doi:10.1109/access.2023.3266644 · Cited by 18
“This has shown the ability to exceed human performance in complex tasks [39], [40], [41] and in traffic control [15], [16], [17], [18].”
BackgroundLearning General Policies with Policy Gradient Methods
Ståhlberg, S., Bonet, B., & Geffner, H. (2023) · Principles of Knowledge Representation and Reasoning · doi:10.24963/kr.2023/63 · Cited by 37
“Reinforcement learning (RL) has delivered remarkable re-sults in a number of settings like game playing and robotics (Mnih et al. 2015; Levine et al. 2016; Silver et al. 2018).”
WordCraft: An Environment for Benchmarking Commonsense AgentsPreprint
Jiang, M., Luketina, J., Nardelli, N., Minervini, P., Torr, P. H. S., Whiteson, S., & Rocktäschel, T. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2007.09185 · Cited by 20
“Rule-based game environments such as Go or Chess (Silver et al., 2017; 2018), while very hard to master, are based on simple transition dynamics and largely disconnected from entities and concepts encountered in the real world.”
BackgroundUnlocking the Potential of Deep Counterfactual Value NetworksPreprint
Zarick, R., Pellegrino, B., Brown, N., & Banister, C. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2007.10442 · Cited by 15
“Lookahead search has been a key component of superhuman AIs in perfect-information games such as backgammon [49], chess [18], and Go [43, 45, 44].”
BackgroundConceptual Challenges for Interpretable Machine LearningPreprint
Watson, D. (2020) · SSRN Electronic Journal · doi:10.2139/ssrn.3668444 · Cited by 10
Learning to Synthesize Programs as Interpretable and Generalizable PoliciesPreprint
Trivedi, D., Zhang, J. M., Sun, S.-H., & Lim, J. J. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2108.13643 · Cited by 64
“[3] David Silver, Thomas Hubert, Julian Schrittwieser, Ioannis Antonoglou, Matthew Lai, Arthur Guez, Marc Lanctot, Laurent Sifre, Dharshan Kumaran, Thore Graepel, et al.”
BackgroundKey requirements for advancing machine learning approaches in single entity electrochemistry
(2024) · doi:10.1016/j.coelec.2024.101526 · Cited by 10
Mahmud, S., Sayed, A. N., Himeur, Y., Nhlabatsi, A., & Bensaali, F. (2025) · Renewable and Sustainable Energy Reviews · doi:10.1016/j.rser.2025.116481 · Cited by 12
Reinforcement learning in robotic systems : A review on sim-to-real transfer
Tiwari, R., Khapre, S. P., & Singh, A. (2026) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2025.105327 · Cited by 21
Kang, M., Choi, S., Fu, K., Liu, X., Wei, Z., Jin, L., Wang, H., Martin, O. J. F., Yang, J. K. W., So, S., et al. (2026) · Opto-Electronic Advances · doi:10.29026/oea.2026.250263 · Cited by 13
Optimal production ramp‐up in the smartphone manufacturing industry
Wang, L., Ai, W., Deng, T., Shen, Z. M., & Hong, C. (2020) · Naval Research Logistics (NRL) · doi:10.1002/nav.21886 · Cited by 9
“After that, AlphaZero achieves superhuman performance in not only Go but also chess and shogi (Silver et al., 2018).”
BackgroundMethodSun, T., Gao, Z., Chang, Z., & Zhao, K. (2021) · Journal of Bionic Engineering · doi:10.1007/s42235-021-00113-9 · Cited by 10
Toward the Next Generation of Neural Iontronic Interfaces
Forró, C., Musall, S., Montes, V. R., Linkhorst, J., Walter, P., Weßling, M., Offenhäusser, A., Ingebrandt, S., Weber, Y. G., Lampert, A., et al. (2023) · Advanced Healthcare Materials · doi:10.1002/adhm.202301055 · Cited by 8
Artificial Intelligence in Surgery, Surgical Subspecialties, and Related Disciplines
Lee, R. Y., Ziehm, A. I., Ullrich, L. A., & Stawicki, S. P. (2023) · Artificial intelligence · doi:10.5772/intechopen.112691 · Cited by 11
Organic synaptic transistors: Biocompatible neuromorphic devices for in-vivo applications
Jang, H., Biswas, S., Lang, P., Bae, J., & Kim, H. (2024) · Organic Electronics · doi:10.1016/j.orgel.2024.107014 · Cited by 9
A Review of Symbolic, Subsymbolic and Hybrid Methods for Sequential Decision Making
Núñez-Molina, C., Mesejo, P., & Fernández-Olivares, J. (2024) · ACM Computing Surveys · doi:10.1145/3663366 · Cited by 14
“We will use Dyna [146] and AlphaZero [141] as illustrative examples of the different existing model-based RL algorithms.”
InfluentialScalable policies for the dynamic traveling multi-maintainer problem with alerts
Verleijsdonk, P., van Jaarsveld, W. L., & Kapodistria, S. (2024) · European Journal of Operational Research · doi:10.1016/j.ejor.2024.05.049 · Cited by 9
“DRL has achieved state-of-the-art performance, for instance, in Atari video games (Mnih et al., 2013) and chess (Silver et al., 2018).”
BackgroundMethodCGGNet: Compiler-Guided Generation Network for Smart Contract Data Augmentation
Hwang, S.-J., Ju, S. H., & Choi, Y.-H. (2024) · IEEE Access · doi:10.1109/access.2024.3427829 · Cited by 9
Quantum search with prior knowledge
He, X., Sun, X., & Zhang, J. (2024) · Science China Information Sciences · doi:10.1007/s11432-023-3972-y · Cited by 17
Two-step dynamic obstacle avoidance
Hart, F., Waltz, M., & Okhrin, O. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.112402 · Cited by 12
Niazmand, V. R., Raheb, M. A., Eqra, N., Vatankhah, R., & Farrokhi, A. (2024) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2024.109041 · Cited by 12
Quantum 2-Player Games and Realizations with Circuits
Zhang, J., Chen, T., Deng, W., Tong, X., & Zhang, X. (2024) · Research · doi:10.34133/research.0480 · Cited by 8
“This convergence has resulted in the development of high-performance computer programs capable of playing at a superhuman level [ 6 – 13 ].”
BackgroundA Stock Prediction Method Based on Deep Reinforcement Learning and Sentiment Analysis
Du, S., & Shen, H. (2024) · Applied Sciences · doi:10.3390/app14198747 · Cited by 15
Symbolic equation solving via reinforcement learning
Dabelow, L., & Ueda, M. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.128732 · Cited by 16
May, D. C., Taylor, M. E., & Musílek, P. (2024) · Energy and AI · doi:10.1016/j.egyai.2024.100446 · Cited by 16
AlphaRouter: Quantum Circuit Routing with Reinforcement Learning and Tree Search
Tang, W., Duan, Y., Kharkov, Y., Fakoor, R., Kessler, E. M., & Shi, Y. (2024) · IEEE International Conference on Quantum Computing and Engineering (QCE) · doi:10.1109/qce60285.2024.00112 · Cited by 35
“Given the recent impressive results of Deep RL [4], [15]–[17], [19], [26], [30], [34], [37], [48], [49], adapting Deep RL frameworks to this problem offers several advantages.”
Can artificial intelligence embody moral values?
Swoboda, T., & Lauwaert, L. (2025) · AI and Ethics · doi:10.1007/s43681-025-00662-3 · Cited by 15
“It has also allowed systems to achieve above human-level gameplay by discovering highly skilled behaviors in games like chess and Go [ 35 , 36 ].”
The impact of artificial intelligence: from cognitive costs to global inequality
Paić, G., & Serkin, L. (2025) · The European Physical Journal Special Topics · doi:10.1140/epjs/s11734-025-01561-8 · Cited by 19
“From sophisticated reinforcement learning algorithms that master chess and other games [2] to AI-driven coding assistants [3] and large language models (LLMs) like ChatGPT [4] or DeepSeek [5], these innovations not only empower individuals to learn and innovate but also help build a more inclusive,…”
Lin, H., Zhu, C., Shang, T., Zhu, N., Lin, K., Zhang, C., Shao, X., Wang, X., & Duan, H. (2025) · Journal of Medicinal Chemistry · doi:10.1021/acs.jmedchem.5c00896 · Cited by 10
Al-Sharman, M., Edes, L., Sun, B., Jayakumar, V., Tahir, H., Daoud, M. A., Emran, B. J., Rayside, D., & Melek, W. (2025) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2025.3646982 · Cited by 28
“Researchers have been motivated to develop these algorithms, owing to their capabilities in handling partially-observable environments by training its data-driven models based on mapping the environmental observations into actions [39].”
Future Vision 2020 and Beyond—5 Critical Trends in Eye Research
Huang, S. S. (2020) · Asia-Pacific Journal of Ophthalmology · doi:10.1097/apo.0000000000000299 · Cited by 10
“Clinical Research Collapse data set to eliminate patients who are unlikely to progress or to respond to treatment Will help direct patient management by making recommendations on follow-up This daunting challenge would seem to be on a distant horizon but consider the story of Stockfish and AlphaZero.(48) From its earl…”
BackgroundInfluentialPick the Moment: Identifying Critical Pedagogical Decisions Using Long-Short Term Rewards.
Ju, S. H., Zhou, G., Barnes, T. M., & Chi, M. (2020) · Educational Data Mining · Cited by 18
Big Data Analytics for Understanding and Fighting COVID-19
Verma, S., & Gazara, R. K. (2020) · Studies in computational intelligence · doi:10.1007/978-981-15-8534-0_17 · Cited by 11
Certificates of quantum many-body properties assisted by machine learning
Requena, B., Muñoz-Gil, G., Lewenstein, M., Dunjko, V., & Tura, J. (2023) · Physical Review Research · doi:10.1103/physrevresearch.5.013097 · Cited by 18
Asher, D. E., Basak, A., Fernandez, R., Sharma, P., Zaroukian, E. G., Hsu, C. D., Dorothy, M. R., Mahre, T., Galindo, G., Frerichs, L., et al. (2022) · The Journal of Defence Modeling and Simulation: Applications, Methodology, Technology · doi:10.1177/15485129221104096 · Cited by 7
(A)I Will Teach You to Play Gomoku: Exploring the Use of Game AI to Teach People
Zhang, Q., & MacLellan, C. J. (2022) · ACM Conference on Learning @ Scale (L@S) · doi:10.1145/3491140.3528331 · Cited by 9
“Fortunately, emerging Game AI systems, such as AlphaZero [6], specifically address the problem of authoring an expert model for these Games.”
BackgroundExact solving scheduling problems accelerated by graph neural networks
Juroš, J., Brčić, M., Končić, M., & Kovač, M. (2022) · International Convention on Information and Communication Technology, Electronics and Microelectronics · doi:10.23919/mipro55190.2022.9803345 · Cited by 12
Feng, X., Chen, H., Wu, G., Zhang, A., & Zhao, Z. (2022) · Applied Sciences · doi:10.3390/app12199869 · Cited by 10
Gao, X., Peng, D., Kui, G., Pan, J., Zuo, X., & Li, F. (2022) · Computers & Chemical Engineering · doi:10.1016/j.compchemeng.2022.108131 · Cited by 9
Yang, S., Barlow, M., Townsend, T., Liu, X., Samarasinghe, D., Lakshika, E., Moy, G., Lynar, T., & Turnbull, B. (2023) · IEEE Access · doi:10.1109/access.2023.3287100 · Cited by 13
“MCTS has shown advantages in large, fully observable games [47], [48], [49], [50], such as chess [51] and Go [52], [53].”
BackgroundFrustratingly Easy Regularization on Representation Can Boost Deep Reinforcement Learning
He, Q., Su, H., Zhang, J., & Hou, X. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.01936 · Cited by 18
“Deep reinforcement learning (DRL) leverages the function approximation abilities of deep neural networks (DNN) and the credit assignment capabilities of RL to enable agents to perform complex control tasks using high-dimensional observations such as image pixels and sensor information [1, 2, 3, 4].”
Multi-sensor based strategy learning with deep reinforcement learning for unmanned ground vehicle
Luo, M. (2023) · International Journal of Intelligent Networks · doi:10.1016/j.ijin.2023.11.003 · Cited by 12
Scalable imaginary time evolution with neural network quantum states
Ledinauskas, E., & Anisimovas, E. (2023) · SciPost Physics · doi:10.21468/scipostphys.15.6.229 · Cited by 14
Semi-Infinitely Constrained Markov Decision Processes and Provably Efficient Reinforcement Learning
Zhang, L., Yang, P., Yang, W., & Zhang, Z. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2023.3348460 · Cited by 11
Shen, J., Lu, H., Lyu, S., & Lu, Y. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.123393 · Cited by 9
Recovery from Adversarial Attacks in Cyber-physical Systems: Shallow, Deep, and Exploratory Works
Lu, P., Zhang, L., Liu, M., Sridhar, K., Sokolsky, O. V., Kong, F., & Lee, I. (2024) · ACM Computing Surveys · doi:10.1145/3653974 · Cited by 18
“Another application considers exoskeleton in healthcare industry [8].”
BackgroundDynamic control of self-assembly of quasicrystalline structures through reinforcement learning
Lieu, U. T., & Yoshinaga, N. (2025) · Soft Matter · doi:10.1039/d4sm01038h · Cited by 11
A Middle Game Search Algorithm Applicable to Low-Cost Personal Computer for Go
Li, X., Lv, Z., Wang, S., Wei, Z., Zhang, X., & Wu, L. (2019) · IEEE Access · doi:10.1109/access.2019.2937943 · Cited by 10
“A total of 5000 first generation tensor processing units (TPUs) were used to generate the self-play games, and 16 second-generation TPUs were used to train the neural networks of AlphaZero [3].”
BackgroundMethodHughes, J. G. (2019) · Yearbook of international humanitarian law · doi:10.1007/978-94-6265-343-6_4 · Cited by 13
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.