Abstracting Imperfect Information Away from Two-Player Zero-Sum Games

Authors: Samuel Sokota, Ryan D’Orazio, Chun Kai Ling, David J Wu, J Zico Kolter, Noam Brown

ICML 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We perform two experiments in which we naively tabularly solve small Pu B-AMGs under regularized objectives using magnetic mirror descent (Sokota et al., 2023) to offer further evidence for our results. We show the results for perturbed rock-paper-scissors Figure 4 and include results for Kuhn poker, as well as the details of our solving procedures, in Section C.
Researcher Affiliation Collaboration Samuel Sokota 1 Ryan D Orazio 2 Chun Kai Ling 1 David J. Wu 3 J. Zico Kolter 1 4 Noam Brown 3 [...] 1Carnegie Mellon University 2Mila, Universit e de Montr eal 3Meta AI 4Bosch Center for AI.
Pseudocode Yes Algorithm 1 Correspondence Mapping Π
Open Source Code No The paper does not contain an explicit statement about providing open-source code for the methodology described in this paper, nor does it provide a direct link to a code repository.
Open Datasets No The paper mentions standard games like 'perturbed rock-paper-scissors' and 'Kuhn poker' as experimental settings, but does not provide specific links, DOIs, repositories, or formal citations with authors/year for publicly available datasets or problem instances used.
Dataset Splits No The paper describes experiments on perturbed rock-paper-scissors and Kuhn poker, but does not provide specific details on how data was split for training, validation, or testing, or any methodology for data partitioning.
Hardware Specification No The paper does not provide specific hardware details (e.g., exact GPU/CPU models, memory amounts) used for running its experiments.
Software Dependencies No The paper mentions using 'magnetic mirror descent (Sokota et al., 2023)' as a game solver, but it does not specify any software dependencies with version numbers (e.g., Python, PyTorch, CUDA versions).
Experiment Setup No The paper describes aspects of the solving procedures for the experiments, such as using 't iterations of MMD' for Kuhn Poker, but it does not provide concrete hyperparameter values or detailed system-level training configurations (e.g., learning rates, batch sizes, specific number of epochs, optimizer settings) that would be needed for reproducibility.