Adversarial Policy Learning in Two-player Competitive Games

Authors: Wenbo Guo, Xian Wu, Sui Huang, Xinyu Xing

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

Reproducibility Variable Result LLM Response
Research Type Experimental In this section, we evaluate our proposed learning algorithm by using five selected games (i.e., four Mu Jo Co games and Star Craft II).
Researcher Affiliation Collaboration 1College of Information Sciences and Technology, The Pennsylvania State University, State College, PA, USA 2Netflix Inc., Los Gatos, CA, USA.
Pseudocode No The paper describes the learning algorithm in text and mathematical equations but does not include a structured pseudocode or algorithm block in the main paper.
Open Source Code Yes We released our source code to support future research.1 https://github.com/psuwuxian/rl_adv_ valuediff
Open Datasets Yes In this section, we evaluate our proposed learning algorithm by using five selected games (i.e., four Mu Jo Co games and Star Craft II). (Todorov et al., 2012) (Sun et al., 2018)
Dataset Splits No The paper describes training adversarial agents in game environments but does not provide specific train/validation/test dataset splits needed for reproduction.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory) used for running its experiments.
Software Dependencies No The paper does not provide specific ancillary software details with version numbers (e.g., library or solver names with version numbers) needed to replicate the experiment.
Experiment Setup Yes Due to space limit, we specify the implementation details and experiment setup (i.e., game and victim policy selection, evaluation metric, hyperparameters) in Supplementary Section S5.