Abstract
The architecture of neural networks in neural based computer game programs influences greatly the strength of the game playing programs. We present developments on the recently tested Mobile Network architecture that has good results for the game of Go. The three proposed improvements deal with the optimization process, the activation function and the convolution layers. These three modifications improve the accuracy of the policy and the error of the evaluation, as well as the playing strength of a computer Go program using the resulting networks.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Cazenave, T.: Residual networks for computer go. IEEE Trans. Games 10(1), 107–110 (2018)
Cazenave, T.: Batch Monte Carlo tree search. Arxiv (2021)
Cazenave, T.: Improving model and search for computer Go. In: IEEE Conference on Games (2021)
Cazenave, T.: Mobile networks for computer Go. IEEE Trans. Games. 28, 58–69 (2021)
Loshchilov, I., Hutter, F.: SGDR: stochastic gradient descent with warm restarts. arXiv preprint arXiv:1608.03983 (2016)
Ramachandran, P., Zoph, B., Le, Q.V.: Searching for activation functions. arXiv preprint arXiv:1710.05941 (2017)
Silver, D., et al.: A general reinforcement learning algorithm that masters chess, shogi, and go through self-play. Science 362(6419), 1140–1144 (2018)
Tan, M., Le, Q.V.: MixConv: mixed depthwise convolutional kernels. arXiv preprint arXiv:1907.09595 (2019)
Wu, D.J.: Accelerating self-play learning in go. CoRR abs/1902.10565 (2019)
Acknowledgment
This work was granted access to the HPC resources of IDRIS under the allocation 2021-AD011012539 made by GENCI.
This work was supported in part by the French government under management of “Agence Nationale de la Recherche” as part of the “Investissements d’avenir” program, reference ANR19-P3IA-0001 (PRAIRIE 3IA Institute).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 Springer Nature Switzerland AG
About this paper
Cite this paper
Cazenave, T., Sentuc, J., Videau, M. (2022). Cosine Annealing, Mixnet and Swish Activation for Computer Go. In: Browne, C., Kishimoto, A., Schaeffer, J. (eds) Advances in Computer Games. ACG 2021. Lecture Notes in Computer Science, vol 13262. Springer, Cham. https://doi.org/10.1007/978-3-031-11488-5_5
Download citation
DOI: https://doi.org/10.1007/978-3-031-11488-5_5
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-11487-8
Online ISBN: 978-3-031-11488-5
eBook Packages: Computer ScienceComputer Science (R0)