Filters








767 Hits in 3.4 sec

Traditional Wisdom and Monte Carlo Tree Search Face-to-Face in the Card Game Scopone

Pier Luca Lanzi, Stefano Di Palma
2018 IEEE Transactions on Games  
We compare rule-based players using the most established strategies (one for beginners and two for advanced players) against players using Monte Carlo Tree Search (MCTS) and Information Set Monte Carlo  ...  Our results show that, as expected, the cheating MCTS outperforms all the other strategies; ISMCTS is stronger than all the rule-based players implementing well-known and most advanced strategies and it  ...  ACKNOWLEDGEMENT Stefano and Pier Luca wish to thank the three anonymous reviewers for their invaluable comments and the people who volunteered to play the game.  ... 
doi:10.1109/tg.2018.2834618 fatcat:7h2rfuhhubap3ek7xqxkj63mvq

ALTERNATIVE SELECTION FUNCTIONS FOR INFORMATION SET MONTE CARLO TREE SEARCH

Viliam Lisy
2014 Acta Polytechnica  
We evaluate the performance of various selection methods for the Monte Carlo Tree Search algorithm in two-player zero-sum extensive-form games with imperfect information.  ...  We show that UCT after initial fast convergence towards a Nash equilibrium computes increasingly worse strategies after some point in time.  ...  Access to the computing and storage facilities owned by parties and projects contributing to the National Grid Infrastructure MetaCentrum, provided under the "Projects of Large Infrastructure for Research  ... 
doi:10.14311/ap.2014.54.0333 fatcat:ticau6oz6rcgjga4pinjh2jm5y

Single-player Monte-Carlo tree search for SameGame

Maarten P.D. Schadd, Mark H.M. Winands, Mandy J.W. Tak, Jos W.H.M. Uiterwijk
2012 Knowledge-Based Systems  
Moreover, SP-MCTS makes use of randomized restarts. We tested IDA* and SP-MCTS on the puzzle SameGame and used the Cross-Entropy Method to tune the SP-MCTS parameters.  ...  Therefore, we propose a new MCTS variant, called Single-Player Monte-Carlo Tree Search (SP-MCTS). The selection and backpropagation strategy in SP-MCTS are different from standard MCTS.  ...  Acknowledgments This work is funded by the Dutch Organisation for Scientific Research (NWO) in the framework of the project Go4Nature, grant number 612.000.938.  ... 
doi:10.1016/j.knosys.2011.08.008 fatcat:lgpy6ib2tneqnmtvbqqf3ee4ym

Implementation and Evaluation of Information Set Monte Carlo Tree Search for Pokémon

Hiroyuki Ihara, Shunsuke Imai, Satoshi Oyama, Masahito Kurihara
2018 2018 IEEE International Conference on Systems, Man, and Cybernetics (SMC)  
In this study, we take Pokémon as an example of a complex imperfect information game and implement a simulator to evaluate the effectiveness of ISMCTS.  ...  Information set Monte Carlo tree search (ISMCTS) has been developed to reduce the effects of strategy fusion caused by determinization of the imperfect information and demonstrated advantages over the  ...  We equated the number of iterations of the algorithms including CMCTS with the total number of the iterations used for multiple search trees by CEMCTS to keep the fairness in comparing them.  ... 
doi:10.1109/smc.2018.00375 dblp:conf/smc/IharaIOK18 fatcat:i5ukma7largmfjmxb2jpeyd53m

Converging to a player model in Monte-Carlo Tree Search

Trevor Sarratt, David V. Pynadath, Arnav Jhala
2014 2014 IEEE Conference on Computational Intelligence and Games  
Evaluation of this approach is done in comparison with value iteration for an iterated version of the prisoner's dilemma problem.  ...  This paper investigates the integration of belief distributions over player models in the Monte-Carlo Tree Search (MCTS) algorithm.  ...  INTRODUCTION Evaluating a player's actions and infering a strategy or player type are useful capabilities for an agent within a game.  ... 
doi:10.1109/cig.2014.6932881 dblp:conf/cig/SarrattPJ14 fatcat:hfwez44z7zh2ng6bsmp4oohoau

Improving Hearthstone AI by Combining MCTS and Supervised Learning Algorithms [article]

Maciej Świechowski, Tomasz Tajmajer, Andrzej Janusz
2018 arXiv   pre-print
We show that even simple neural networks can be trained and successfully used for the evaluation of game states.  ...  We investigate the impact of supervised prediction models on the strength and efficiency of artificial agents that use the Monte-Carlo Tree Search (MCTS) algorithm to play a popular video game Hearthstone  ...  The deck used by the second player, has on the other hand, a lot of complex strategies and needs to be played carefully; yet used by a skillful player, it has a much greater winning potential compared  ... 
arXiv:1808.04794v1 fatcat:e5jyvgibirdyrndx3c5or4jzpu

Combinatorial Multi-armed Bandits for Real-Time Strategy Games [article]

Santiago Ontañón
2017 arXiv   pre-print
We then evaluate these strategies in the context of real-time strategy (RTS) games, a genre of computer games characterized by their very large branching factors.  ...  In this paper, we address this problem with a sampling strategy for Monte Carlo Tree Search (MCTS) algorithms called naïve sampling, based on a variant of the Multi-armed Bandit problem called Combinatorial  ...  For simplicity, however, in our experimental evaluation, we will use a standard MCTS algorithm. In this paper, we will use CMABs to model the decision process that a player faces in RTS games.  ... 
arXiv:1710.04805v1 fatcat:s5u36e5tprgv3g5cota2lbfva4

Combinatorial Multi-armed Bandits for Real-Time Strategy Games

Santiago Ontañón
2017 The Journal of Artificial Intelligence Research  
We then evaluate these strategies in the context of real-time strategy (RTS) games, a genre of computer games characterized by their very large branching factors.  ...  In this paper, we address this problem with a sampling strategy for Monte Carlo Tree Search (MCTS) algorithms called "naive sampling", based on a variant of the Multi-armed Bandit problem called "Combinatorial  ...  For simplicity, however, in our experimental evaluation, we will use a standard MCTS algorithm. In this paper, we will use CMABs to model the decision process that a player faces in RTS games.  ... 
doi:10.1613/jair.5398 fatcat:vfjocekemfe2nkh5zqhigmf5ta

Learning Self-Game-Play Agents for Combinatorial Optimization Problems [article]

Ruiyang Xu, Karl Lieberherr
2019 arXiv   pre-print
We try to leverage the computational power of neural MCTS to solve a class of combinatorial optimization problems.  ...  The ZG also provides a specially designed neural MCTS. We use a combinatorial planning problem for which the ground-truth policy is efficiently computable to demonstrate that ZG is promising.  ...  A neural MCTS algorithm, instead, uses a neural network V ϕ to predict the result evaluation so that the algorithm saves the time on rolling out. (4) BACKUP: This is the last phase of an iteration where  ... 
arXiv:1903.03674v2 fatcat:a4hkhku42ff63lohosmghdcyze

Multiplayer AlphaZero [article]

Nick Petosa, Tucker Balch
2019 arXiv   pre-print
Our work supports the use of AlphaZero in multiplayer games and suggests future research for more complex environments.  ...  In this work, we suggest novel modifications of the AlphaZero algorithm to support multiplayer environments, and evaluate the approach in two simple 3-player games.  ...  So when backpropagating value, MCTS uses the corresponding score in v for each player instead of flipping the sign of a scalar v. 4.  ... 
arXiv:1910.13012v3 fatcat:npceip42vzdffgihhf7qj5ispq

ExIt-OOS: Towards Learning from Planning in Imperfect Information Games [article]

Andy Kitchen, Michela Benedetti
2018 arXiv   pre-print
We use Online Outcome Sampling, an online search algorithm for imperfect information games in place of MCTS.  ...  While training online, our neural strategy is used to improve the accuracy of playouts in OOS, allowing a learning and planning feedback loop for imperfect information games.  ...  The authors would like to thank anonymous colleagues for implementing a large part of our testing infrastructure. GNU parallel was used for some experimental runs [Tange, 2018] .  ... 
arXiv:1808.10120v2 fatcat:vr7rmui74jf5tkkadzscc2zrb4

Monte Carlo tree search experiments in hearthstone

Andre Santos, Pedro A. Santos, Francisco S. Melo
2017 2017 IEEE Conference on Computational Intelligence and Games (CIG)  
We illustrate through extensive empirical validation the superior performance of our approach against vanilla MCTS and the current state-of-the art AI for Hearthstone.  ...  We argue that, in light of the challenges posed by the game (such as uncertainty and hidden information), Monte Carlo tree search offers an appealing alternative to existing AI players.  ...  The authors would like to thank the Metastone developers for assisting in the use of the platform, and particularly to the GitHub user @demilich1 for all the assistance throughout the development of this  ... 
doi:10.1109/cig.2017.8080446 dblp:conf/cig/SantosSM17 fatcat:an22rlghwjegdn3qvsl7ikabg4

Biasing MCTS with Features for General Games [article]

Dennis J. N. J. Soemers, Éric Piette, Cameron Browne
2019 arXiv   pre-print
We evaluate the playing strength of an MCTS player biased by learnt features against a standard upper confidence bounds for trees (UCT) player in multiple different board games, and demonstrate significantly  ...  This paper proposes using a linear function approximator, rather than a deep neural network (DNN), to bias a Monte Carlo tree search (MCTS) player for general games.  ...  A Biased MCTS player using this feature set throughout complete play-outs achieves a win percentage of 93% against UCT, despite a 30× reduction in the MCTS iteration count.  ... 
arXiv:1903.08942v1 fatcat:7thq7n6eibgvlp62asqzo55xbu

Monte Carlo Tree Search in Simultaneous Move Games with Applications to Goofspiel [chapter]

Marc Lanctot, Viliam Lisý, Mark H. M. Winands
2014 Communications in Computer and Information Science  
When actions are chosen simultaneously, players may need to mix between their strategies. In this paper, we discuss the extension of MCTS to simultaneous move games with and without chance events.  ...  We compare both head-to-head performance and exploitability of several MCTS variants in Goofspiel.  ...  This work is partially funded by the Netherlands Organisation for Scientific Research (NWO) in the framework of the project Go4Nature, grant number 612.000.938 and the Czech Science Foundation, grant no  ... 
doi:10.1007/978-3-319-05428-5_3 fatcat:y62cbw2sfjcd3bpvwh3nbkwryi

Monte Carlo Tree Search: A Review of Recent Modifications and Applications [article]

Maciej Świechowski, Konrad Godlewski, Bartosz Sawicki, Jacek Mańdziuk
2021 arXiv   pre-print
MCTS performs random sampling in the form of simulations and stores statistics of actions to make more educated choices in each subsequent iteration.  ...  The method has become a state-of-the-art technique for combinatorial games, however, in more complex games (e.g. those with high branching factor or real-time ones), as well as in various practical domains  ...  In StG each of the two players commits to a certain mixed strategy, i.e. a probability distribution of pure strategies.  ... 
arXiv:2103.04931v2 fatcat:zxkssxeqqjarzlgcglvmb476f4
« Previous Showing results 1 — 15 out of 767 results