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 701–800 of 3,630
A machine learning method correlating pulse pressure wave data with pregnancy
Chen, J., Huang, H., Hao, W., & Xu, J. (2019) · International Journal for Numerical Methods in Biomedical Engineering · doi:10.1002/cnm.3272 · Cited by 25
ANN modelling of CO2 refrigerant cooling system COP in a smart warehouse
Opalic, S. M., Goodwin, M., Jiao, L., Nielsen, H. K., Pardiñas, Á. Á., Hafner, A., & Kolhe, M. L. (2020) · Journal of Cleaner Production · doi:10.1016/j.jclepro.2020.120887 · Cited by 28
The AI Triad and What It Means for National Security Strategy
Buchanan, B. (2020) · doi:10.51593/20200021 · Cited by 28
Global Convergence of Multi-Agent Policy Gradient in Markov Potential GamesPreprint
Leonardos, S., Overman, W., Panageas, I., & Piliouras, G. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2106.01969 · Cited by 108
“…driver of numerous recent advances in Artificial Intelligence (AI) applications that range from super-human performance in competitive game-playing (Silver et al., 2016; 2018; Brown and Sandholm, 2019) and strategic decision-making in multiple tasks (Mnih et al., 2015; OpenAI, 2018; Vinyals et…”
BackgroundBots and Beasts: What Makes Machines, Animals, and People Smart?
Thagard, P. R. (2021) · The MIT Press eBooks · doi:10.7551/mitpress/14102.001.0001 · Cited by 31
Yao, H., Tan, P., Yang, T., & Zhou, F. (2024) · Computer-Aided Civil and Infrastructure Engineering · doi:10.1111/mice.13159 · Cited by 25
Mansfield, D., & Montazeri, A. (2024) · Frontiers in Robotics and AI · doi:10.3389/frobt.2024.1336612 · Cited by 40
Machine learning for open-pit mining: a systematic review
Liu, S. Q., Liu, L., Kozan, E., et al. (2024) · International Journal of Mining Reclamation and Environment · doi:10.1080/17480930.2024.2362579 · Cited by 24
“In addition, ML has a wide range of applications and plays a major role in manufacturing [8,9], transportation [10–12], healthcare [13,14], environment [15,16], entertainment [17], chemical sciences [18,19], agriculture [20] and mining [21,22].”
Pre-training with non-expert human demonstration for deep reinforcement learning
de la Cruz, G. V., Du, Y., & Taylor, M. E. (2019) · The Knowledge Engineering Review · doi:10.1017/s0269888919000055 · Cited by 39
“Using a deep neural network as its function approximator, deep RL can learn state representations directly from raw input pixels and has achieved state-of-the-art results in various domains (Mnih et al., 2015; Silver et al., 2016; Kempka et al., 2016; Lillicrap et al., 2016; Duan et al., 2016; Silver et al., 2018).”
BackgroundKishimoto, A., Buesser, B., Chen, B., & Botea, A. (2019) · neural information processing systems · Cited by 27
Putting deep learning in perspective for pest management scientists
Clark, R. D. (2020) · Pest Management Science · doi:10.1002/ps.5820 · Cited by 28
Deep Q-learning for the selection of optimal isocratic scouting runs in liquid chromatography
Kensert, A., Collaerts, G., Efthymiadis, K., Desmet, G., & Cabooter, D. (2021) · Journal of Chromatography A · doi:10.1016/j.chroma.2021.461900 · Cited by 26
Fifty years of P vs. NP and the possibility of the impossible
Fortnow, L. (2021) · Communications of the ACM · doi:10.1145/3460351 · Cited by 63
Omnidirectional-Wheel Conveyor Path Planning and Sorting Using Reinforcement Learning Algorithms
Zaher, W., Youssef, A. W., Shihata, L. A., Azab, E., & Mashaly, M. E. (2022) · IEEE Access · doi:10.1109/access.2022.3156924 · Cited by 28
“It also can be used to play chess, shogi and Go [10-11] and to control and guide robots in indoor navigation [12-13].”
BackgroundAutomating reinforcement learning architecture design for code optimization
Wang, H., Tang, Z., Zhang, C., Zhao, J., Cummins, C., Leather, H., & Wang, Z. (2022) · ACM SIGPLAN International Conference on Compiler Construction (CC) · doi:10.1145/3497776.3517769 · Cited by 40
“In recent years, deep reinforcement learning (RL) has been shown to be effective for navigating a sizeable discrete space, outperforming traditional search techniques on a range of optimization tasks [33, 60, 79, 85, 87].”
BackgroundRetroGraph: Retrosynthetic Planning with Graph Search
Xie, S., Yan, R., Han, P., Xia, Y., Wu, L., Guo, C., Yang, B., & Qin, T. (2022) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3534678.3539446 · Cited by 56
“Reinforcement learning based methods have achieved significant results on many conventional games such as GO [31, 32], chess [31], poker [4], and mahjong [24].”
MethodEmergence of cooperation in two-agent repeated games with reinforcement learning
Ding, Z.-W., Zheng, G.-Z., Cai, C.-R., Cai, W., Chen, L., Zhang, J., & Wang, X. (2023) · Chaos Solitons & Fractals · doi:10.1016/j.chaos.2023.114032 · Cited by 35
“Reinforcement learning is originally designed to make optimal decision for maximizing the rewards for the given states through exploratory experimentation [38, 48, 49, 50, 51, 52, 53], and this setup exactly matches with the evolution of cooperation.”
BackgroundA survey of decision making in adversarial games
Li, X., Meng, M., Hong, Y., & Chen, J. (2024) · Science China Information Sciences · doi:10.1007/s11432-022-3777-y · Cited by 36
“…the game of Go has both perfect and complete information, since it has explicit game rules and all chess pieces’ positions are visible to both players at all times as well as the actions of the opponent, which has been well solved by well-known AI agents, such as AlphaGo and AlphaZero [10]– [12].”
BackgroundReinforcement learning for semi-autonomous approximate quantum eigensolver
Albarrán-Arriagada, F., Retamal, J. C., Solano, E., & Lamata, L. (2020) · Machine Learning Science and Technology · doi:10.1088/2632-2153/ab43b4 · Cited by 28
Automated synthesis of steady-state continuous processes using reinforcement learning
Göttl, Q., Grimm, D. G., & Burger, J. (2021) · Frontiers of Chemical Science and Engineering · doi:10.1007/s11705-021-2055-9 · Cited by 39
TorchBeast: A PyTorch Platform for Distributed RLPreprint
Küttler, H., Nardelli, N., Lavril, T., Selvatici, M., Sivakumar, V., Rocktäschel, T., & Grefenstette, E. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1910.03552 · Cited by 70
“Reinforcement learning has recently had a surge of interest thanks to the rise of deep learning and new GPU hardware, conquering important challenges such as chess, Go, and other board games [4, 5], demonstrating the ability to learn policies on visual inputs [6, 7], and tackle strategically complex environments [8, 9…”
BackgroundFast Adaptation Trajectory Prediction Method Based on Online Multisource Transfer Learning
Yang, B., Zhu, J., Yu, Z., Fan, F., Liu, X., & Ni, R. (2024) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2024.3362980 · Cited by 30
“Moreover, conventional deep learning models focus on specific tasks [44], leading to catastrophic forgetting [45] when encountering new tasks.”
Continuous control of structural vibrations using hybrid deep reinforcement learning policy
Panda, J., Chopra, M., Matsagar, V. A., & Chakraborty, S. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.124075 · Cited by 26
Effective Multi-Agent Deep Reinforcement Learning Control With Relative Entropy Regularization
Miao, C., Cui, Y., Li, H., & Wu, X. (2024) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2024.3398712 · Cited by 25
“Utilizing the power of deep neural networks [3] to adapt abstract features from high-dimensional input states, RL has demonstrated superior performances than humans in various complex scenarios including board games [4], video games [5], and robot control [6].”
BackgroundAcceleration for Deep Reinforcement Learning using Parallel and Distributed Computing: A Survey
Liu, Z., Xu, X., Qiao, P., & Li, D. (2024) · ACM Computing Surveys · doi:10.1145/3703453 · Cited by 39
“TPUs are designed specifically for machine learning workloads and used frequently in the latest Google findings in the DRL domains such as AlphaZero [3], AlphaStar [4], GATO [121], and so on.”
From Nodes to Roads: Surveying DRL Applications in MEC-Enhanced Terrestrial Wireless Networks
Ullah, S. A., Bibi, M., Hassan, S. A., Abou-Zeid, H., Qureshi, H. K., Jung, H., Mahmood, A., Gidlund, M., & Hossain, E. (2025) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2025.3568035 · Cited by 35
“Examples include AlphaZero [53], model-based RL with model-free fine-tuning (MBMF) [54], imagination-augmented agents (I2A) [55], and Monte Carlo tree search (MCTS) [56]. b. Model-Free Methods: Model-free DRL methods rely on direct interaction with the environment to learn optimal policies without…”
Can Machines Be Artists? A Deweyan Response in Theory and Practice
Still, A. W., & d’Inverno, M. (2019) · Arts · doi:10.3390/arts8010036 · Cited by 42
“It was used in the generation of the portrait Edmond de Belamy which fetched $432,550 at Christies in 2018, and in the development of AlphaZero, which rapidly learns a variety of games, including Chess, to a world-class level (Silver et al. 2018).”
MethodDistributed Fusion-Based Policy Search for Fast Robot Locomotion Learning
Cao, Z., Xiao, Q., & Zhou, M. (2019) · IEEE Computational Intelligence Magazine · doi:10.1109/mci.2019.2919364 · Cited by 23
“Moreover, recently the algorithm is improved to master chess, shogi, and Go completely from scratch without any human knowledge [20]–[22].”
Aoyama, Y., Maruko, I., Kawano, T., Yokoyama, T., Ogawa, Y., Maruko, R., & Iida, T. (2021) · PLoS ONE · doi:10.1371/journal.pone.0244469 · Cited by 24
Hierarchies of Planning and Reinforcement Learning for Robot Navigation
Wöhlke, J., Schmitt, F., & van Hoof, H. (2021) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48506.2021.9561151 · Cited by 46
Intelligent career planning via stochastic subsampling reinforcement learning
Guo, P., Xiao, K., Ye, Z., Zhu, H., & Zhu, W. (2022) · Scientific Reports · doi:10.1038/s41598-022-11872-8 · Cited by 44
Cosine Annealing, Mixnet and Swish Activation for Computer Go
Cazenave, T., Sentuc, J., & Videau, M. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-11488-5_5 · Cited by 23
“Typically, the dataset used for the supervised learning is constituted of superhuman games produced by Deep Reinforcement Learning agent like AlphaZero or Katago.”
MethodRethinking Closed-Loop Training for Autonomous Driving
Zhang, W., Guo, R., Zeng, W., Xiong, Y., Dai, B., Hu, R., Ren, M., & Urtasun, R. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19842-7_16 · Cited by 45
“It can also be used during inference for planning with multi-step lookaheads [48,26,15,69,64,51,25,70].”
BackgroundMulti-Label Continual Learning Using Augmented Graph Convolutional Network
Du, K., Lyu, F., Li, L., Hu, F., Feng, W., Xu, F., Xi, X., & Cheng, H. (2023) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2023.3305871 · Cited by 40
“M ACHINE learning approaches have achieved human-level performance on various tasks, such as Atari games [1] or object recognition [2].”
BackgroundAbbas, A. N., Chasparis, G. C., & Kelleher, J. D. (2023) · Data & Knowledge Engineering · doi:10.1016/j.datak.2023.102240 · Cited by 35
Ghadermazi, P., & Chan, S. H. J. (2024) · Bioinformatics · doi:10.1093/bioinformatics/btae003 · Cited by 29
“Reinforcement learning has shown great promise in solving very complex problems in the past decade (Mnih et al. 2013, Silver et al. 2018, Brown et al. 2020) and have been used with success in different fields of science and engineering (Mousavi et al. 2017, Treloar et al. 2020, Jebellat et al.…”
MethodSmoothing policies and safe policy gradients
Papini, M., Pirotta, M., & Restelli, M. (2022) · Machine Learning · doi:10.1007/s10994-022-06232-6 · Cited by 48
“1 3 1 Introduction Reinforcement Learning (RL) (Sutton & Barto, 2018) has achieved astounding successes in games (Mnih et al., 2015; Silver et al., 2018; OpenAI, 2018; Vinyals et al., 2019), matching or surpassing human performance in several occasions.”
BackgroundAdversarial Risk via Optimal Transport and Optimal Couplings
Pydi, M. S., & Jog, V. (2021) · IEEE Transactions on Information Theory · doi:10.1109/tit.2021.3100107 · Cited by 32
Adaptive FPGA Placement Optimization via Reinforcement Learning
Murray, K. E., & Betz, V. (2019) · ACM/IEEE Workshop on Machine Learning for CAD (MLCAD) · doi:10.1109/mlcad48534.2019.9142079 · Cited by 27
“RL techniques have also been shown to be effective for hard search and decision problems such as Chess, Go [6], and real time strategy games [7].”
MethodAutonomous Algorithmic Collusion: Q-Learning Under Sequential PricingPreprint
Klein, T. (2018) · SSRN Electronic Journal · doi:10.2139/ssrn.3195812 · Cited by 23
Online and Offline Reinforcement Learning by Planning with a Learned Model
Schrittwieser, J., Hubert, T., Mandhane, A., Barekatain, M., Antonoglou, I., & Silver, D. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2104.06294 · Cited by 23
Self-citationCausal reinforcement learning based on Bayesian networks applied to industrial settings
Valverde, G., Quesada, D., Larrañaga, P., & Bielza, C. (2023) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2023.106657 · Cited by 26
Robust and efficient task scheduling for robotics applications with reinforcement learning
Tejer, M., Szczepański, R., & Tarczewski, T. (2023) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2023.107300 · Cited by 26
Autonomous navigation of marine surface vessel in extreme encounter situation
Guan, W., Han, H., & Cui, Z. (2024) · Journal of Marine Science and Technology · doi:10.1007/s00773-023-00979-w · Cited by 23
“Notably, the AlphaGo program developed by the Google DeepMind team in 2016, achieved a significant breakthrough in decision-making by defeating Lee Sedol, a professional chess player from Korea [10].”
Akkerman, F., Mes, M., & van Jaarsveld, W. L. (2024) · Computers & Industrial Engineering · doi:10.1016/j.cie.2024.110747 · Cited by 24
“The adopted algorithm is a natural extension of the batch-AVI algorithm that has conceptual similarities to AlphaZero (Silver et al., 2018), see also van Jaarsveld (2020).”
Neda, O. M., Adabi, J., Gholinezhadomran, H., et al. (2025) · Sustainable Cities and Society · doi:10.1016/j.scs.2025.106744 · Cited by 22
Comparing Reinforcement Learning Methods for Real-Time Optimization of a Chemical Process
Quah, T., Machalek, D., & Powell, K. M. (2020) · Processes · doi:10.3390/pr8111497 · Cited by 27
Offline Reinforcement Learning as Anti-exploration
Rezaeifar, S., Dadashi, R., Vieillard, N., Hussenot, L., Bachem, O., Pietquin, O., & Geist, M. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i7.20783 · Cited by 77
Machine Learning and Artificial Intelligence for Surgical Decision Making
Byerly, S. E., Maurer, L. R., Mantero, A. M. A., Naar, L., An, G. C., & Kaafarani, H. M. A. (2021) · Surgical Infections · doi:10.1089/sur.2021.007 · Cited by 26
Ebrahimi, M. R., Pacheco, J. L., Li, W., Hu, J. L., & Chen, H. (2021) · IEEE Security and Privacy Workshops (SPW) · doi:10.1109/spw53761.2021.00021 · Cited by 27
“Deep RL has shown breakthrough performance in estimating π and Q using neural networks [27].”
BackgroundPlattfaut, R., & Koch, J. (2021) · Journal of Business Research · doi:10.1016/j.jbusres.2021.07.024 · Cited by 27
A reinforcement learning approach to the stochastic cutting stock problem
Neto, A. R. P., & Murta, A. H. F. (2022) · EURO Journal on Computational Optimization · doi:10.1016/j.ejco.2022.100027 · Cited by 22
“RL has recently shown impressive results in video games (Mnih et al., 2015), board games (Silver et al., 2018) and robotics (Gu et al., 2017).”
The application of reinforcement learning to NATM tunnel design
Soranzo, E., Guardiani, C., & Wu, W. (2022) · Underground Space · doi:10.1016/j.undsp.2022.01.005 · Cited by 25
Mimetic Models: Ethical Implications of AI that Acts Like You
McIlroy-Young, R., Kleinberg, J. M., Sen, S., Barocas, S., & Anderson, A. (2022) · Proceedings of the AAAI/ACM Conference on AI Ethics and Society · doi:10.1145/3514094.3534177 · Cited by 44
“…to build mimetic models of particular players when there are available records of people’s past game play [6, 23, 34, 43, 63, 64, 68, 71, 92, 97, 98, 100], whether we’re considering Chess [100], Go [68], Shogi [92], Hanabi [6], Diplo-macy [43], or other games with a finite set of legal moves.”
BackgroundReinforcement learning with guarantees: a review
Osinenko, P., Dobriborsci, D., & Aumer, W. (2022) · IFAC-PapersOnLine · doi:10.1016/j.ifacol.2022.07.619 · Cited by 26
Natural General Intelligence: How understanding the brain can help us build AI
Summerfield, C. (2022) · doi:10.1093/oso/9780192843883.001.0001 · Cited by 40
Demonstration of a programmable optical lattice atom interferometer
LeDesma, C., Mehling, K., Shao, J., Wilson, J. D., Axelrad, P., Nicotra, M. M., Anderson, D. Z., & Holland, M. (2024) · Physical Review Research · doi:10.1103/physrevresearch.6.043120 · Cited by 29
Crossing the Cleft: Communication Challenges Between Neuroscience and Artificial Intelligence
Chance, F. S., Aimone, J. B., Musuvathy, S., Smith, M. R., Vineyard, C. M., & Wang, F. (2020) · Frontiers in Computational Neuroscience · doi:10.3389/fncom.2020.00039 · Cited by 22
“Correspondingly, notable advances in artificial intelligence (AI), for example reinforcement learning [e.g., as used by AlphaZero, see (Silver et al., 2018)], the Transformer network (Vaswani et al., 2017), and deep convolutional networks (Krizhevsky et al., 2012), are based upon descriptions or…”
BackgroundMulti-component transfer metric learning for handling unrelated source domain samples
Yi, C., Xu, Y., Yu, H., Yan, Y., & Liu, Y. (2020) · Knowledge-Based Systems · doi:10.1016/j.knosys.2020.106132 · Cited by 23
The present and future role of artificial intelligence and machine learning in anesthesiology
Alexander, J. C., Romito, B. T., & Çobanoğlu, M. C. (2020) · International Anesthesiology Clinics · doi:10.1097/aia.0000000000000294 · Cited by 25
Navigating the landscape of multiplayer games
Omidshafiei, S., Tuyls, K., Czarnecki, W. M., Santos, F. C., Rowland, M., Connor, J., Hennes, D., Muller, P., Pérolat, J., Vylder, B. D., et al. (2020) · Nature Communications · doi:10.1038/s41467-020-19244-4 · Cited by 51
“[137] David Silver, Thomas Hubert, Julian Schrittwieser, Ioannis Antonoglou, Matthew Lai, Arthur Guez, Marc Lanctot, Laurent Sifre, Dharshan Kumaran, Thore Graepel, Timothy Lillicrap, Karen Simonyan, and Demis Hassabis.”
BackgroundLearn to Place: FPGA Placement Using Reinforcement Learning and Directed Moves
Elgamma, M. A., Murray, K. E., & Betz, V. (2020) · International Conference on Field-Programmable Technology (FPT) · doi:10.1109/icfpt51103.2020.00021 · Cited by 22
“Go and chess [19]).”
Gossip-based Actor-Learner Architectures for Deep Reinforcement LearningPreprint
Assran, M., Romoff, J., Ballas, N., Pineau, J., & Rabbat, M. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1906.04585 · Cited by 51
“Deep Reinforcement Learning (Deep RL) agents have reached superhuman performance in a few domains [Silver et al., 2016, 2018, Mnih et al., 2015, Vinyals et al., 2019], but this is typically at significant computational expense [Tian et al., 2019].”
BackgroundA Generalized Training Approach for Multiagent LearningPreprint
Müller, P. F., Omidshafiei, S., Rowland, M., Tuyls, K., Pérolat, J., Liu, S., Hennes, D., Marris, L., Lanctot, M., Hughes, E., et al. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1909.12823 · Cited by 90
“…results have been recently achieved in restricted settings (e.g., zero-sum, two-player games) using game-theoretic principles such as iterative best response computation (Lanctot et al., 2017), self-play (Silver et al., 2018), and evolution-based training (Jaderberg et al., 2019; Liu et al., 2019).”
BackgroundSelf-citationDeep Learning for Scene Classification: A SurveyPreprint
Zeng, D., Liao, M., Tavakolian, M., Guo, Y., Zhou, B., Hu, D., Pietikäinen, M., & Liu, L. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2101.10531 · Cited by 59
“Since then, deep learning has started to dominate various tasks (like computer vision [72], [118], [119], [120], speech recognition [121], autonomous driving [122], cancer detection [123], [124], machine translation [125], playing complex games [126], [127], [128], [129], earthquake forecasting [130], medicine discove…”
BackgroundRepresentation Matters: Offline Pretraining for Sequential Decision MakingPreprint
Yang, M., & Nachum, O. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2102.05815 · Cited by 75
“Several recent model-based RL methods use latent state representations as a way to simplify the model learning and policy rollout elements of model-based policy optimization (Oh et al., 2017; Silver et al., 2018; Hafner et al., 2020), with the rollout in latent space sometimes referred to as ‘imagination’ (Racanière e…”
BackgroundMethodHuman-Level Reinforcement Learning through Theory-Based Modeling, Exploration, and PlanningPreprint
Tsividis, P. A., Loula, J., Burga, J., Foss, N., Campero, A., Pouncy, T., Gershman, S. J., & Tenenbaum, J. B. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2107.12544 · Cited by 101
Deep Reinforcement Learning: A New Beacon for Intelligent Active Flow Control
Xie, F., Zheng, C., Ji, T., Zhang, X., Bi, R., Zhou, H., & Zheng, Y. (2023) · Aerospace Research Communications · doi:10.3389/arc.2023.11130 · Cited by 24
“Based on the such setting, the DRL is embedded as a state representation technology, which makes it possible to deal with high-dimensional complex problems, like Go, StarCraft, Robotics [21, 25 – 27].”
Khuat, T. T., Kedziora, D. J., & Gabryś, B. (2023) · Foundations and Trends® in Human–Computer Interaction · doi:10.1561/1100000091 · Cited by 47
“1The AlphaZero approach is shared by AutoML systems such as AlphaD3M (Drori et al., 2018), but, for the purpose of the chess analogy, the two are not equivalent; automation in chess is significantly more advanced than automation across an entire ML application.”
BackgroundInfluentialWang, Y., Jia, Y., Fan, S., & Xiao, J. (2024) · Artificial Intelligence Review · doi:10.1007/s10462-023-10697-9 · Cited by 27
“DRL has been widely used in various real-world fields (Moor et al. 2022; Fan et al. 2020; Liu et al. 2020; Patel et al. 2019; Silver et al. 2018), and the most popular model is the DQN model (Mnih et al. 2013, 2015).”
Model-based offline reinforcement learning framework for optimizing tunnel boring machine operation
Cao, Y., Luo, W., Xue, Y., Lin, W., & Zhang, F. (2024) · Underground Space · doi:10.1016/j.undsp.2024.01.008 · Cited by 22
Sun, L., Li, H., Nagel, J., & Yang, S. (2024) · Applied Sciences · doi:10.3390/app14177949 · Cited by 24
“Due to its versatility, RL has found successful applications across various domains, including autonomous vehicles [109] and board games; notably, it played a pivotal role in the development of AlphaGo and AlphaZero [87,110].”
Camillo, F. D., Grimaldi, D. A., Cattarinussi, G., Giorgio, A. D., Locatelli, C., Khuntia, A. T., Enrico, P., Brambilla, P., Koutsouleris, N., & Sambataro, F. (2024) · Psychiatry and Clinical Neurosciences · doi:10.1111/pcn.13736 · Cited by 32
“27 Notably, in the context of psychiatric disorders, these methods have demonstrated substantial promise, including in discerning nuanced patterns within neuroimaging data.”
Fire-Flyer AI-HPC: A Cost-Effective Software-Hardware Co-Design for Deep Learning
An, W., Bi, X., Chen, G.-T., Chen, S., Deng, C., Ding, H., Dong, K., Du, Q., Gao, W., Guan, K., et al. (2024) · International Conference for High Performance Computing, Networking, Storage and Analysis (SC) · doi:10.1109/sc41406.2024.00089 · Cited by 55
“With the advent of more complex models like Al-phaFold [24] and AlphaZero [25] highlighting the need for more computational power and memory, revealing the limitations of traditional FP64 / FP32 computing devices.”
CHAOPT: A Testbed for Evaluating Human-Autonomy Team Collaboration Using the Video Game Overcooked!2
Bishop, J., Burgess, J., Ramos, C., Driggs, J. B., Williams, T., Tossell, C. C., Phillips, E. K., Shaw, T. H., & de Visser, E. J. (2020) · Systems and Information Engineering Design Symposium · doi:10.1109/sieds49339.2020.9106686 · Cited by 24
“…teamwork and communication provides a challenging litmus test for an artificially intelligent agents’ capabilities, which have traditionally focused on competitively beating human players in games such as Chess, Go and Starcraft [5,6], as opposed to working collaborative with human players.”
BackgroundSelf-Play Reinforcement Learning for Fast Image Retargeting
Kajiura, N., Kosugi, S., Wang, X., et al. (2020) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3394171.3413857 · Cited by 21
“This method is mainly used for learning game strategies such as chess, shogi, Go, and Othello [23, 24, 28].”
MethodThe Game Is Not over Yet—Go in the Post-AlphaGo Era
Egri-Nagy, A., & Törmänen, A. (2020) · Philosophies · doi:10.3390/philosophies5040037 · Cited by 23
“The technical details are precisely described [13–15], but they do not sufficiently answer the question on a conceptual level.”
BackgroundMethodAddressing Hindsight Bias in Multigoal Reinforcement Learning
Bai, C., Wang, L., Wang, Z., Wang, Z., Zhao, R., Bai, C., & Liu, P. (2021) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2021.3107202 · Cited by 29
“Associated with recent advances in deep learning, deep RL (DRL) has achieved human-level performance in video games [2]–[4], Go [5], [6], and challenging robotic tasks [7], [8].”
BackgroundBergler, C., Gebhard, A., Towers, J. R., Butyrev, L., Sutton, G. J., Shaw, T. J. H., Maier, A., & Nöth, E. (2021) · Scientific Reports · doi:10.1038/s41598-021-02506-6 · Cited by 28
Scalability investigation of Double Deep Q Learning for factory layout planning
Klar, M., Hussong, M., Ruediger-Flore, P., Yi, L., Glatt, M. F., & Aurich, J. C. (2022) · Procedia CIRP · doi:10.1016/j.procir.2022.04.027 · Cited by 23
Pisarski, D., & Jankowski, Ł. (2022) · Computer-Aided Civil and Infrastructure Engineering · doi:10.1111/mice.12920 · Cited by 24
“…results in a variety of hard real-world control-like problems, ranging from a superhuman level of proficiency in the games of chess and Go (Silver et al., 2018), through the thermal soaring of gliders (Reddy et al., 2016) and swimming by body undulation (Jiao et al., 2021), to bus traffic…”
BackgroundArtificial intelligence can improve patients’ experience in decentralized clinical trials
Thomas, K. A., & Kidziński, Ł. (2022) · Nature Medicine · doi:10.1038/s41591-022-02034-4 · Cited by 21
Dynamic Self-Supervised Teacher-Student Network Learning
Ye, F., & Gerasimou, S. (2022) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2022.3220928 · Cited by 32
“In recent years, machine learning models have surpassed human-level performance in individual tasks [1], such as in Atari gaming [2] or in image recognition [3].”
BackgroundPrediction of bone metastasis in non-small cell lung cancer based on machine learning
Li, M.-P., Liu, W., Sun, B.-L., Zhong, N. S., Liu, Z., Huang, S.-H., Zhang, Z., & Liu, J. (2023) · Frontiers in Oncology · doi:10.3389/fonc.2022.1054300 · Cited by 25
“There already are many applications of this technology throughout science and society ranging from driverless cars to Board games to decision-making (12, 13).”
BackgroundA human-centered safe robot reinforcement learning framework with interactive behaviors
Gu, S., Kshirsagar, A., Du, Y., Chen, G., Peters, J., & Knoll, A. (2023) · Frontiers in Neurorobotics · doi:10.3389/fnbot.2023.1280341 · Cited by 43
“By leveraging deep learning, Reinforcement Learning (RL) has achieved remarkable successes in many scenarios and superhuman performance in some challenging tasks [25, 24], e.g., autonomous driving [23], recommender system [81], robotics [10], games [65, 16, 26], and finance [71].”
BackgroundStudent of Games: A unified learning algorithm for both perfect and imperfect information games
Schmid, M., Moravčík, M., BURCH, N. R., Kadlec, R., Davidson, J., Waugh, K., Bard, N., Timbers, F., Lanctot, M., Holland, G. Z., et al. (2023) · Science Advances · doi:10.1126/sciadv.adg3256 · Cited by 94
“Likewise, AlphaGo could not play chess; however, its successor AlphaZero ( 6 ) could, and did.”
Self-citationOn the computational complexity of ethics: moral tractability for minds and machines
Stenseke, J. (2024) · Artificial Intelligence Review · doi:10.1007/s10462-024-10732-3 · Cited by 37
Automated segmentation of dermal fillers in OCT images of mice using convolutional neural networks
Pfister, M., Schützenberger, K., Pfeiffenberger, U., Messner, A., Chen, Z., dos Santos, V. A., Puchner, S., Garhöfer, G., Schmetterer, L. F., Gröschl, M., et al. (2019) · Biomedical Optics Express · doi:10.1364/boe.10.001315 · Cited by 20
Combination and application of machine learning and computational mechanics
Li, X., Yan, Z., & Liu, Z. (2019) · Chinese Science Bulletin (Chinese Version) · doi:10.1360/n972019-00005 · Cited by 20
Linking Big Data and Prediction Strategies: Tools, Pitfalls, and Lessons Learned
Yang, S., Stansbury, L. G., Rock, P., Scalea, T. M., & Hu, P. (2019) · Critical Care Medicine · doi:10.1097/ccm.0000000000003739 · Cited by 24
Reinforcement learning of adaptive online rescheduling timing and computing time allocation
Ikonen, T. J., Heljanko, K., & Harjunkoski, I. (2020) · Computers & Chemical Engineering · doi:10.1016/j.compchemeng.2020.106994 · Cited by 21
“A recent example of the e(cid:11)ectiveness of RL is the AlphaZero algorithm by DeepMind that can achieve, via self-play, higher ELO rating 1 than Stock(cid:12)sh 8 in chess and a superhuman level in two other board games shogi and Go (Silver et al., 2018).”
MethodArchetypal landscapes for deep neural networks
Verpoort, P. C., Lee, A. A., & Wales, D. J. (2020) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1919995117 · Cited by 22
Induction and Exploitation of Subgoal Automata for Reinforcement Learning
Furelos-Blanco, D., Law, M., Jonsson, A., Broda, K. B., & Russo, A. M. (2021) · Journal of Artificial Intelligence Research · doi:10.1613/jair.1.12372 · Cited by 57
“, 2015) and mastering complex board games (Silver et al., 2018).”
BackgroundProjection Path Explorer: Exploring Visual Patterns in Projected Decision-making Paths
Hinterreiter, A., Steinparz, C., Schöfl, M., Stitz, H., & Streit, M. (2021) · ACM Transactions on Interactive Intelligent Systems · doi:10.1145/3387165 · Cited by 26
“removedEspecially, itIt would be particularly interesting to compare strategies used in classic chess computer programs with those learned by state-of-the art reinforcement learning algorithms such as AlphaZero [45].”
MethodVariational Dynamic for Self-Supervised Exploration in Deep Reinforcement Learning
Bai, C., Liu, P., Liu, K., Wang, L., Zhao, Y., Han, L., & Wang, Z. (2021) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2021.3129160 · Cited by 30
“ning. I. INTRODUCTION R EINFORCEMENT learning (RL) [1] achieves promising results in solving a wide range of problems, including human-level performance on Atari games [2]–[4], the board game Go [5], [6], the strategy game StarCraft II [7], and challenging robotic tasks [8]–[10]. Most of the successes in RL rely on a…”
BackgroundInformation Theory for Agents in Artificial Intelligence, Psychology, and Economics
Harré, M. (2021) · Entropy · doi:10.3390/e23030310 · Cited by 30
“The next candidate is AlphaGo [45] that beat the 9 dan Go champion Lee Sedol in 2016 and its decedent, AlphaZero [46], that learned to play Go, Chess, and Shogi entirely from self-play.”
Background(2019) · Springer eBooks · doi:10.1007/978-3-030-19734-6 · Cited by 23
Deep Counterfactual Regret MinimizationPreprint
Brown, N., Lerer, A. K., Gross, S., & Sandholm, T. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1811.00164 · Cited by 104
The Chess Transformer: Mastering Play using Generative Language Models
Noever, D. A., Ciolino, M., & Kalin, J. D. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2008.04057 · Cited by 20
“This approach further contrasts with the Monte Carlo Tree Search (MCTS) employed so effectively with Alpha-Go and reinforcement learning (Silver et al. 2018).”
ResultReplay-Guided Adversarial Environment Design
Jiang, M., Dennis, M., Parker-Holder, J., Foerster, J., Grefenstette, E., & Rocktäschel, T. (2021) · arXiv · doi:10.48550/arxiv.2110.02439 · Cited by 21
“While deep reinforcement learning (RL) approaches have led to many successful applications in challenging domains like Atari [21], Go [35], Chess [36], Dota [4], and StarCraft [40] in recent years, deep RL agents still prove to be brittle, often failing to transfer to environments only slightly different from those en…”
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.