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 1,801–1,900 of 3,633
Proposing and solving olympiad geometry with guided tree search
Zhang, C., Song, J., Li, S., Liang, Y., Ma, Y., Wang, W., Zhu, Y. B., & Zhu, S. C. (2026) · Nature Machine Intelligence · doi:10.1038/s42256-025-01164-x · Cited by 30
“Inspired by the success of board-game reinforcement learning agents 32–34 , we trained both a policy model and a value model.”
Grid-to-Graph: Flexible Spatial Relational Inductive Biases for Reinforcement Learning
Jiang, Z., Minervini, P., Jiang, M., & Rocktäschel, T. (2021) · Adaptive Agents and Multi-Agents Systems · doi:10.65109/lmmu1902 · Cited by 12
An, G. C., Cockrell, C., Zamora, R., & Yoram Vodovotz, B. (2021) · Elsevier eBooks · doi:10.1016/b978-0-12-821440-4.00009-8 · Cited by 5
How to Make AlphaGo’s Children Explainable
Park, W. (2022) · Philosophies · doi:10.3390/philosophies7030055 · Cited by 8
Reinforcement learning to reduce failures in SOT-MRAM switching
Ender, J., de Orio, R. L., Fiorentini, S., et al. (2022) · Microelectronics Reliability · doi:10.1016/j.microrel.2022.114570 · Cited by 4
Learning Chess with Language Models and Transformers
DeLeo, M., & Guven, E. (2022) · International Conference on Data Science and Machine Learning (DSML) · doi:10.5121/csit.2022.121515 · Cited by 13
Metaverse: A Solution to the Multi-Agent Value Alignment Problem
Yampolskiy, R. V. (2022) · Journal of Artificial Intelligence and Consciousness · doi:10.1142/s2705078522500072 · Cited by 8
Artificial General Intelligence and the Common Sense Argument
Häggström, O. (2022) · Studies in applied philosophy, epistemology and rational ethics · doi:10.1007/978-3-031-09153-7_12 · Cited by 5
Semi-model-Based Reinforcement Learning in Organic Computing Systems
von Pilchau, W. P., Stein, A., & Hähner, J. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-21867-5_16 · Cited by 3
Importance of prefrontal meta control in human-like reinforcement learning
Lee, J. H., Leibo, J. Z., An, S.-J., & Lee, S. W. (2022) · Frontiers in Computational Neuroscience · doi:10.3389/fncom.2022.1060101 · Cited by 4
Extracting tactics learned from self-play in general games
Soemers, D. J. N. J., Samothrakis, S., Piette, É., & Stephenson, M. (2022) · Information Sciences · doi:10.1016/j.ins.2022.12.080 · Cited by 10
“Stateof-the-art results in terms of game playing strength [1,2] are, in recent years, essentially always obtained using Deep Neural Networks (DNNs) [3] as function approximators for policies and/or value functions [4].”
InfluentialCurrent trend in control of artificial intelligence for health robotic manipulator
Suwarno, I., Çakan, A., Raharja, N. M., Baballe, M. A., & Mahmoud, M. S. (2022) · Journal of Soft Computing Exploration · doi:10.52465/joscex.v4i1.96 · Cited by 7
Tensor Implementation of Monte-Carlo Tree Search for Model-Based Reinforcement Learning
Baláž, M., & Tarábek, P. (2023) · Applied Sciences · doi:10.3390/app13031406 · Cited by 4
AlphaZe∗∗: AlphaZero-like baselines for imperfect information games are surprisingly strong
Blüml, J., Czech, J., & Kersting, K. (2023) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2023.1014561 · Cited by 12
“In 2017, DeepMind introduced AlphaZero (Silver, 2017, 2018), a system that teaches itself games from scratch and can beat current grandmasters or world champions in games like chess, shogi, and Go.”
BackgroundDeep Reinforcement Learning in Financial Markets Context: Review and Open Challenges
Boutyour, Y., & Idrissi, A. (2023) · Studies in computational intelligence · doi:10.1007/978-3-031-33309-5_5 · Cited by 4
User Cold-Start Learning in Recommender Systems using Monte Carlo Tree Search
Rajapakse, D. C., & Leith, D. J. (2023) · ACM Transactions on Recommender Systems · doi:10.1145/3618002 · Cited by 5
“Since we have a budget of t max items, the maximum lookahead d at step t is equal to t max − t .”
Weighting Information Sets with Siamese Neural Networks in Reconnaissance Blind Chess
Bertram, T., Fürnkranz, J., & Müller, M. (2023) · IEEE Conference on Games (CoG) · doi:10.1109/cog57401.2023.10333170 · Cited by 4
“The majority of this encoding is taken up by specifying the last-requested move by the player [18].”
BackgroundA Computational Bibliometric Analysis of Game Advertising using VOSviewer
Ginting, Y. R. (2023) · International Journal of Informatics Information System and Computer Engineering (INJIISCOM) · doi:10.34010/injiiscom.v4i1.9571 · Cited by 3
Quantum neural networks successfully calibrate language models
Neto, J. S., Ardila, L. Q., Nogueira, T. N., Albuquerque, F., Papa, J. P., Guido, R. C., & Fanchini, F. F. (2024) · Quantum Machine Intelligence · doi:10.1007/s42484-024-00139-2 · Cited by 7
“For example, in computer vision (He et al. 2015; Szegedy et al. 2014), natural language processing (Vaswani et al. 2017; Devlin et al. 2018), deep reinforcement learning (Silver et al. 2018; Open et al. 2019) and also automatic speech recognition (Hinton et al. 2012).”
A survey on medical image analysis using deep reinforcement learning approach
Shaikh, K. A.-H. (2024) · World Journal of Advanced Research and Reviews · doi:10.30574/wjarr.2024.21.2.0294 · Cited by 4
Reinforcement Learning Guided Symbolic Execution for Ethereum Smart Contracts
Wang, M., Fei, W., Wang, M., & Cui, J. F. (2023) · Asia-Pacific Software Engineering Conference (APSEC) · doi:10.1109/apsec60848.2023.00019 · Cited by 6
“For example, the famous Go program AlphaGo Zero achieved superhuman performance in the game of Go through the training of reinforcement learning algorithms [25].”
MethodNetwork modeling: The explanatory power of activity flow models of brain function
Cole, M. W. (2024) · Elsevier eBooks · doi:10.1016/b978-0-443-13480-7.00004-1 · Cited by 7
Intelligent Recognition of Tool Wear with Artificial Intelligence Agent
Gao, J., Qiao, H., & Zhang, Y. (2024) · Coatings · doi:10.3390/coatings14070827 · Cited by 5
Difficulties and approaches in enabling learning-in-memory using crossbar arrays of memristors
Wang, W. J., Li, Y., & Wang, M. (2024) · Neuromorphic Computing and Engineering · doi:10.1088/2634-4386/ad6732 · Cited by 6
“AlpahGoZero [24] and AlphaZero [25] developed by DeepMind increase the computation demands, i.e. the cost in dollars, by roughly 3–4 orders of magnitude compared with ResNet-50 [26, 27].”
Learning With Generalised Card Representations for “Magic: The Gathering”
Bertram, T., Fürnkranz, J., & Müller, M. (2024) · IEEE Conference on Games (CoG) · doi:10.1109/cog60054.2024.10645602 · Cited by 7
“While super-human play has been achieved in some games such as Go, chess, DOTA and poker [1], [6], [13], many others still provide numerous obstacles.”
BackgroundGodø, J. M. K., Steen, S., Barrett, D. S., & Triantafyllou, M. S. (2024) · Ocean Engineering · doi:10.1016/j.oceaneng.2024.119236 · Cited by 4
Neural network approaches for parameterized optimal control
Verma, D., Winovich, N., Ruthotto, L., & Waanders, B. V. B. (2024) · Foundations of Data Science · doi:10.3934/fods.2024042 · Cited by 16
“Reinforcement learning (RL) has emerged as a potent data-driven approach and shown success in diverse applications such as robotics, autonomous driving, games, and recommender systems [24, 30, 36, 31, 7, 14, 4].”
BackgroundZhang, K., Zhu, D., Xu, Q., et al. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-72356-8_18 · Cited by 4
Nagel, J. (2024) · Oxford University Press eBooks · doi:10.1093/9780191976766.003.0007 · Cited by 4
Disentangled representations for causal cognition
Torresan, F., & Baltieri, M. (2024) · Physics of Life Reviews · doi:10.1016/j.plrev.2024.10.003 · Cited by 10
ExSPIN: Explicit Feedback-Based Self-Play Fine-Tuning for Text-to-SQL Parsing
Yan, L., Su, J., Liu, C., Duan, S., Zhang, Y., Li, J., Han, P., & Liu, Y. (2025) · Entropy · doi:10.3390/e27030235 · Cited by 4
Mean-Field Aided QMIX: A Scalable and Flexible Q-Learning Approach for Large-Scale Agent Groups
Zhang, E., Tang, H., Zhang, X.-P. S., & Ding, W. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10887881 · Cited by 4
“To learn effective behaviors in these scenarios based on agents’ experiences, multi-agent reinforcement learning (MARL) is viewed as a potentially practical method [6, 7].”
Retracted: The Ethics of AI in Literature: Reflections On Representation and Responsibility
Saddhono, K., Nugrahani, F., Syamsuri, A. S., Ermanto, Rahim, A., & Sarimanah, E. (2024) · International Conference on IoT, Communication and Automation Technology (ICICAT) · doi:10.1109/icicat62666.2024.10923247 · Cited by 7
RetractedSwami, A. K., Verma, D. K., Soni, R., & Goswami, D. (2025) · Materials Today Communications · doi:10.1016/j.mtcomm.2025.112553 · Cited by 12
Simple modification of the upper confidence bound algorithm by generalized weighted averages
Manome, N., Shinohara, S., & Chung, U. (2025) · PLoS ONE · doi:10.1371/journal.pone.0322757 · Cited by 16
“UCB-based models such as Q-learning models with UCB (Jin et al. 2018) and reinforcement learning models with a Monte Carlo tree search (Silver et al. 2017; Silver et al. 2018; Schrittwieser et al. 2020) may be easily extended.”
Role play: Learning adaptive role-specific strategies in multi-agent interactions
Long, W., Wen, W., Zhai, P., & Zhang, L. (2025) · Knowledge-Based Systems · doi:10.1016/j.knosys.2025.113811 · Cited by 10
A review of simulation optimization with connection to artificial intelligence
Peng, Y., Chen, C.-H., & Fu, M. C. (2025) · Fundamental Research · doi:10.1016/j.fmre.2024.12.020 · Cited by 4
Looking deeper into the algorithms underlying human planning
Kuperwajs, I., Russek, E. M., Mattar, M. G., Ma, W. J., & Griffiths, T. L. (2025) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2025.06.006 · Cited by 7
“AlphaGo was improved upon with variants that do not rely on human data [80] or domain-speci fi c knowledge [81].”
The Computational Evolution of Cognitive Architectures
Kotseruba, I., & Tsotsos, J. K. (2025) · doi:10.1093/oso/9780192844835.001.0001 · Cited by 7
Bank, T., Madsen, F. W., Mirshekali, H., et al. (2025) · Journal of Energy Storage · doi:10.1016/j.est.2025.119338 · Cited by 4
Deep Learning Competition Framework on Othello for Education
Lin, C.-N., Chen, J.-C., & Yen, S.-J. (2019) · IEEE Transactions on Games · doi:10.1109/tg.2019.2931153 · Cited by 4
“AlphaZero used a ResNet model and 17 input planes [7], [8].”
MethodMulti-Armed Bandits for Minesweeper: Profiting From Exploration–Exploitation Synergy
Lordeiro, I. Q., Haddad, D. B., & Cardoso, D. O. (2021) · IEEE Transactions on Games · doi:10.1109/tg.2021.3082909 · Cited by 4
“[9] were able to achieve groundbreaking results in Chess, Shogi and Go with AlphaZero.”
BackgroundResultNeural Embeddings of Financial Time-Series Data
Sokolov, A., Mostovoy, J., Parker, B., & Seco, L. (2020) · The Journal of Financial Data Science · doi:10.3905/jfds.2020.1.041 · Cited by 4
Itoh, H., Nakano, H., Tokushima, R., Fukumoto, H., & Wakuya, H. (2020) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2020.3034428 · Cited by 5
“Parallelizing MCTS [44] and combining MCTS with deep neural networks [45] can sig-nificantly speed up the planning.”
MethodDeep Reinforcement Learning for Supply Chain Synchronization
Jackson, I. (2022) · Hawaii International Conference on System Sciences (HICSS) · doi:10.24251/hicss.2022.246 · Cited by 5
Combining M-MCTS and Deep Reinforcement Learning for General Game Playing
Liang, S., Jiang, G., & Zhang, Y. (2022) · Lecture notes in computer science · doi:10.1007/978-3-030-94662-3_14 · Cited by 4
Deployment of Differential Privacy for Application in Artificial Intelligence
Senekane, M. (2021) · International Conference on Electrical, Computer and Energy Technologies (ICECET) · doi:10.1109/icecet52533.2021.9698473 · Cited by 6
Deng, H., Yuan, X., Tian, Y., & Hu, J. (2022) · IEEJ Transactions on Electrical and Electronic Engineering · doi:10.1002/tee.23556 · Cited by 5
“In general, the architecture of the PV-Net is analogous to the one in AlphaZero [11].”
BackgroundResultBeckman, L., & Rosenberg, J. H. (2022) · Philosophy & Technology · doi:10.1007/s13347-022-00525-3 · Cited by 8
Jacobson, S. H., & Jokela, J. A. (2022) · Cambridge University Press eBooks · doi:10.1017/9781108872188 · Cited by 5
Machine Learning for Cognitive Robotics
Ogata, T., Takahashi, K., Yamada, T., Murata, S., & Sasaki, K. (2022) · The MIT Press eBooks · doi:10.7551/mitpress/13780.003.0014 · Cited by 6
Specification Aware Multi-Agent Reinforcement Learning
Ritz, F., Phan, T., Müller, R., Gabor, T., Sedlmeier, A., Zeller, M., Wieghardt, J., Schmid, R., Sauer, H., Klein, C., et al. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-10161-8_1 · Cited by 6
Fast and Accurate User Cold-Start Learning Using Monte Carlo Tree Search
Rajapakse, D. C., & Leith, D. J. (2022) · ACM Conference on Recommender Systems (RecSys) · doi:10.1145/3523227.3546786 · Cited by 4
“MCTS was used by AlphaGo [16] to defeat the world champion in the game of Go, and also in the later AlphaZero [17] game playing engine.”
MethodA Training Model of Wargaming Based on Imitation Learning and Deep Reinforcement Learning
Wu, K., Liu, M., Cui, P., et al. (2022) · Lecture notes in electrical engineering · doi:10.1007/978-981-19-6203-5_78 · Cited by 4
Szczepańska, A., & Kaźmierczak, R. (2022) · International Journal of Environmental Research and Public Health · doi:10.3390/ijerph191912353 · Cited by 5
Hierarchical Architecture for Multi-Agent Reinforcement Learning in Intelligent Game
Li, B. (2022) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn55064.2022.9892666 · Cited by 5
“AlphaGo Zero [22] followed AlphaGo and defeated it thoroughly.”
BackgroundWagerWin: An Efficient Reinforcement Learning Framework for Gambling Games
Wang, H., Wu, H.-J., & Lai, G. (2023) · IEEE Transactions on Games · doi:10.1109/tg.2022.3226526 · Cited by 6
“The fast-developing deep RL has led to great success in various types of games, e.g., AlphaGo [1] and AlphaZero [2] in board games; AlphaStar [3], OpenAI Five [4], and JueWu [5] in large video games.”
BackgroundMethodInfluentialPark, S., Kim, J., Jeong, H., Kim, T.-K., & Yoo, J. (2023) · Sensors · doi:10.3390/s23104946 · Cited by 5
“Notable successes include learning to play various games from raw images (board games [2] and video games [3,4]), controlling a car from a camera frame in the virtual environment [5], solving complicated problems from camera observations [6–8], and picking up objects in the real world [9].”
BackgroundHeaton, J., & Givigi, S. N. (2023) · IEEE International Systems Conference (SysCon) · doi:10.1109/syscon53073.2023.10131174 · Cited by 6
“It would be very hard to programmatically tell a robot how it should move its joints to walk [13], or what moves it should make in a game of chess [14], and using RL is a way to get the agent to teach itself.”
BackgroundEx(plainable) Machina: how social-implicit XAI affects complex human-robot teaming tasks
Matarese, M., Cocchella, F., Rea, F., & Sciutti, A. (2023) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48891.2023.10160839 · Cited by 9
“We chose the AlphaZero architecture [38] trained to play the Connect 4 game because it could easily reach perfect performance against the MC-TS agent.”
MethodDeep reinforcement learning applied to Monte Carlo power system reliability analysis
Solheim, Ø. R., Høverstad, B. A., & Korpås, M. (2023) · IEEE PowerTech Conference · doi:10.1109/powertech55446.2023.10202826 · Cited by 7
Present: AI as a New Form of Agency, Not Intelligence
Floridi, L. (2023) · Oxford University Press eBooks · doi:10.1093/oso/9780198883098.003.0002 · Cited by 6
Zheng, X., Chen, D., Lin, Z., et al. (2023) · The Journal of Supercomputing · doi:10.1007/s11227-023-05621-5 · Cited by 4
“Under the supervision conditions of setting the rules of Go and determining the winners and losers of Go, AlphaZero played much self-play, then improved its power by filtering the conditions [21].”
IxDRL: A Novel Explainable Deep Reinforcement Learning Toolkit Based on Analyses of Interestingness
Sequeira, P., & Gervasio, M. T. (2023) · Communications in computer and information science · doi:10.1007/978-3-031-44064-9_20 · Cited by 6
Using machine learning to find exact analytic solutions to analytically posed physics problems
Ashhab, S. (2024) · Heliyon · doi:10.1016/j.heliyon.2024.e28124 · Cited by 8
Reinforcement Learning and Data-Generation for Syntax-Guided Synthesis
Parsert, J., & Polgreen, E. (2024) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v38i9.28938 · Cited by 12
“The algorithm that we adapt to enumerative synthesis is based on AlphaZero (Silver et al. 2018) and has successfully been applied to first-order logic theorem proving (Kaliszyk et al. 2018) and combinator synthesis (Gauthier 2020).”
MethodInfluentialRetracted: AI-Infused Research and Development in Universities: Accelerating Scientific Discovery
Badrus, Salim, M. A., Sari, I. N., Azha, S., Zulfikri, M., & Sutantri (2024) · International Conference on Advance Computing and Innovative Technologies in Engineering (ICACITE) · doi:10.1109/icacite60783.2024.10616647 · Cited by 5
RetractedComparison Training for Computer Chinese Chess
Chen, J.-C., Tseng, W.-J., Wu, I., & Wei, T.-H. (2019) · IEEE Transactions on Games · doi:10.1109/tg.2019.2893430 · Cited by 4
“In contrast, although deep reinforcement learning [20], [21] recently made significant success on Go, chess, and shogi, the required computing power (5000 first-generation tensor processing units (TPUs) as mentioned in [21]) is too costly for many developers.”
BackgroundControl Approach Combining Reinforcement Learning and Model-Based Control
Okawa, Y., Sasaki, T., & Iwane, H. (2019) · Cited by 4
Li, X., He, S., Wei, Z., & Wu, L. (2019) · IEEE Access · doi:10.1109/access.2019.2938568 · Cited by 4
“AlphaGo’s [3] descendant AlphaGo Zero [4] and the comprehensive AlphaZero [5] achieved superhuman performance in Go, Chess, and Shogi without using any human knowledge and experience, thereby indicating that deep reinforcement learning combined with MCTS [6] achieves good results for”
BackgroundInfluentialArtificial Intelligence Accidentally Learned Ecology through Video Games
Barbe, L., Mony, C., & Abbott, B. W. (2020) · Trends in Ecology & Evolution · doi:10.1016/j.tree.2020.04.006 · Cited by 6
Therapeutics as Control: Model-Based Control Discovery for Sepsis
An, G. C., Cockrell, C., & Day, J. (2020) · Springer eBooks · doi:10.1007/978-3-030-56510-7_5 · Cited by 5
Interactive Robot for Playing Russian Checkers
Kopets, E. E., Karimov, A. I., Kolev, G. Y., Scalera, L., & Бутусов, Д. (2020) · Robotics · doi:10.3390/robotics9040107 · Cited by 4
“We developed an original game algorithm using basic features of the renowned AlphaZero, which proved to be e ffi cient in the game Go and then was extended to chess and other games [22].”
MethodA Tutorial Introduction to Monte Carlo Tree Search
Fu, M. C. (2020) · Winter Simulation Conference · doi:10.1109/wsc48552.2020.9384090 · Cited by 5
“(Silver et al. 2018, p.3)”
Tschirhart, P., & Segall, K. (2021) · Frontiers in Neuroscience · doi:10.3389/fnins.2021.750748 · Cited by 5
Split Moves for Monte-Carlo Tree Search
Kowalski, J., Mika, M., Pawlik, W., et al. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i9.21265 · Cited by 4
“More recently, ALPHAZERO, showed how to utilize a single technique to play Go, Chess, and Shogi on a level above all other compared AI agents (Silver et al. 2018).”
MethodPlanning Rational Behavior of Cognitive Semiotic Agents in a Dynamic Environment
Osipov, G. S., & Panov, A. I. (2021) · Scientific and Technical Information Processing · doi:10.3103/s0147688221060113 · Cited by 5
The Games We Play: Critical Complexity Improves Machine Learning
Birhane, A., & Sumpter, D. J. T. (2022) · Frontiers in artificial intelligence and applications · doi:10.3233/faia220186 · Cited by 9
Curriculum based Reinforcement Learning for traffic simulations
Makri, S., & Charalambous, P. (2023) · Computers & Graphics · doi:10.1016/j.cag.2023.04.009 · Cited by 4
Are Talkative AI Agents More Likely to Win the Werewolf Game?
Tellols, D. (2019) · doi:10.18653/v1/w19-8302 · Cited by 4
“Proposals for games that do not require social interaction between the players, like chess, shogi and go, are achieving promising results (Silver et al., 2016, 2018).”
BackgroundDeep reinforcement learning for feedback control in a collective flashing ratchet
Kim, D.-K., & Jeong, H. (2021) · Physical Review Research · doi:10.1103/physrevresearch.3.l022002 · Cited by 5
On Strongly Solving Chinese Checkers
Sturtevant, N. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-65883-0_13 · Cited by 4
“As Chinese Checkers falls into this category, we expect that the AlphaZero approach is able to learn to play the game.”
BackgroundAgent-Based Modeling of Systemic Inflammation: A Pathway Toward Controlling Sepsis
An, G. C., & Cockrell, R. C. (2021) · Methods in molecular biology · doi:10.1007/978-1-0716-1488-4_20 · Cited by 5
Gobang Game Algorithm Based on Reinforcement Learning
Li, X., Zhang, W., Chen, J., et al. (2022) · Communications in computer and information science · doi:10.1007/978-981-16-9247-5_36 · Cited by 4
DeepPursuit: Uniting Classical Wisdom and Deep RL for Sparse Recovery
Chen, Z., Zhong, S., Chen, J., & Zhao, Y. (2021) · Asilomar Conference on Signals, Systems and Computers (ACSSC) · doi:10.1109/ieeeconf53345.2021.9723110 · Cited by 4
“Since the MDP transition is deterministic and known, we are able to use such model-based information to perform planning with MCTS in RL training [24].”
MethodInfluentialDynamic Prediction of Origin-Destination Flows Using Fusion Line Graph Convolutional Networks.Preprint
Xiong, X., Özbay, K. M. A., Li, J., et al. (2019) · arXiv (Cornell University) · Cited by 4
Symmetrical Gaussian Error Linear Units (SGELUs)Preprint
Yu, C., & Su, Z. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1911.03925 · Cited by 4
“Artificial intelligent (AI) has shown its power in changing the world in a smart way, such as the well-known AlphaZero [Silver et al., 2018].”
BackgroundNorm-based generalisation bounds for multi-class convolutional neural networksPreprint
Ledent, A., Mustafa, W. A. A., Lei, Y., et al. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1905.12430 · Cited by 4
“Deep learning has enjoyed an enormous amount of success in a variety of engineering applications in the last decade [1, 2, 3, 4].”
BackgroundCoordination in Adversarial Sequential Team Games via Multi-Agent Deep Reinforcement LearningPreprint
Celli, A., Ciccone, M., Bongo, R., & Gatti, N. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1912.07712 · Cited by 20
“Thanks to the recent successes of deep RL in single-agent environments [40, 54, 55], MARL is recently experiencing a new wave of interest and some old ideas have been adapted to leverage the power of function approximators [46, 47].”
BackgroundLearning from Learning Machines: Optimisation, Rules, and Social NormsPreprint
LaCroix, T., & Bengio, Y. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2001.00006 · Cited by 7
“…2016; Vaswani et al., 2018), bioinformatics (Chen et al., 2016; Min et al., 2017; Li et al., 2019), drug de-sign (Gawehn et al., 2016; Jing et al., 2018), medical image analysis (Litjens et al., 2017; Shen et al., 2017), game playing (Mnih et al., 2013, 2015; Silver et al., 2016, 2017, 2018), etc.”
BackgroundTowards Neural-Guided Program Synthesis for Linear Temporal Logic SpecificationsPreprint
Camacho, A., & McIlraith, S. A. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1912.13430 · Cited by 4
EdgeSlice: Slicing Wireless Edge Computing Network with Decentralized Deep Reinforcement LearningPreprint
Liu, Q., Han, T., & Moges, E. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2003.12911 · Cited by 4
“Although DRL techniques have been extensively studied in many areas such as robotic control [16], traffic control [17], and chess games [18], the existing DRL models are not appropriate to solve problem P3 for two reasons.”
BackgroundOptimizing Quantum Annealing Schedules: From Monte Carlo Tree Search to QuantumZeroPreprint
Chen, Y., Chen, Y., Lee, C., Zhang, S., & Hsieh, C. (2020) · arXiv (Cornell University) · Cited by 4
Selective Dyna-style Planning Under Limited Model CapacityPreprint
Abbas, Z., Sokota, S., Talvitie, E. J., & White, M. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2007.02418 · Cited by 9
“And in Chess, Shogi, Go, and Poker, superhuman performance can be reached by means of decision-time planning on exact transition models (Silver et al., 2018; Moravčı́k et al., 2017; Brown & Sandholm, 2017).”
BackgroundDocument-editing Assistants and Model-based Reinforcement Learning as a Path to Conversational AIPreprint
Kudashkina, K., Pilarski, P. M., & Sutton, R. S. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2008.12095 · Cited by 6
Deep Reinforcement Learning for Process SynthesisPreprint
Midgley, L. I. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2009.13265 · Cited by 16
“RL has had many recent successful applications, including mastering games such as chess and Go [1].”
BackgroundDisentangled Planning and Control in Vision Based Robotics via Reward MachinesPreprint
Camacho, A., Varley, J., Jain, D., Işçen, A., & Kalashnikov, D. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2012.14464 · Cited by 8
“g, playing Atari games from pixels [23], playing board games [28], and doing control from raw pixel observations in robotics applications [17]).”
BackgroundTime Matters in Using Data Augmentation for Vision-based Deep Reinforcement Learning.Preprint
Ko, B., & Ok, J. (2021) · arXiv (Cornell University) · Cited by 4
Automated and Autonomous Experiment in Electron and Scanning Probe MicroscopyPreprint
Kalinin, S. V., Ziatdinov, M., Hinkle, J. D., Jesse, S., Ghosh, A., Kelley, K. P., Lupini, A., Sumpter, B., & Vasudevan, R. (2021) · arXiv.org · doi:10.48550/arxiv.2103.12165 · Cited by 4
“In many of these cases, the strategies learned by the agents are remarkably different from those of top players.(86) They have also shown steady progression in the area of robotics, for example with learning dexterous hand manipulations.”
BackgroundLevel generation and style enhancement -- deep learning for game development overviewPreprint
Migdał, P., Olechno, B., & Podgórski, B. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2107.07397 · Cited by 6
“Recent progress in reinforcement learning resulted in creating networks surpassing top human players in Go [11], reaching grandmaster level in real-time strategy StarCraft II [12], and playing at a decent level in a multiplayer battle arena Dota 2 [13].”
Background
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.