Learning in Nonzero-Sum Stochastic Games with Potentials

Authors: David H Mguni, Yutong Wu, Yali Du, Yaodong Yang, Ziyi Wang, Minne Li, Ying Wen, Joel Jennings, Jun Wang

ICML 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate SPot-AC in three popular multi-agent environments: the particle world (Lowe et al., 2017), a network routing game (Roughgarden, 2007) and a Cournot duopoly problem (Agliari et al., 2016).
Researcher Affiliation Collaboration 1Huawei R&D UK 2Institute of Automation, Chinese Academy of Sciences 3University College London, UK 4Shanghai Jiao Tong University.
Pseudocode Yes Algorithm 1 SPot Q: Stochastic POTential Q-Learning; Algorithm 2 SPot-AC: Stochastic POTential Actor-Critic
Open Source Code No The paper does not provide a specific link or explicit statement about the release of source code for the described methodology.
Open Datasets Yes We evaluate SPot-AC in three popular multi-agent environments: the particle world (Lowe et al., 2017), a network routing game (Roughgarden, 2007) and a Cournot duopoly problem (Agliari et al., 2016).
Dataset Splits No The paper does not provide specific train/validation/test dataset splits. It only states that experiments are repeated for 5 independent runs.
Hardware Specification No The paper does not explicitly describe the hardware used to run its experiments.
Software Dependencies No The paper does not provide specific ancillary software details with version numbers.
Experiment Setup No The paper states "Further details on the settings can be found in Appendix" but does not provide specific experimental setup details or hyperparameters in the main text.