Exploiting hidden structures in non-convex games for convergence to Nash equilibrium

Authors: Iosif Sakos, Emmanouil-Vasileios Vlatakis-Gkaragkounis, Panayotis Mertikopoulos, Georgios Piliouras

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

Reproducibility Variable Result LLM Response
Research Type Experimental 5 Experiments This section demonstrates our method s applicability in a couple of different and insightful setups. Technical details of those setups, as well as additional experimental results are deferred to the supplementary material. We start with a regularized version of Matching Pennies zero-sum game where the players strategies are controlled by two individual preconfigured differentiable MLPs.
Researcher Affiliation Academia Iosif Sakos Engineering Systems and Design (SUTD) iosif_sakos@mymail.sutd.edu.sg Emmanouil V. Vlatakis-Gkaragkounis UC Berkeley emvlatakis@cs.columbia.edu Panayotis Mertikopoulos Univ. Grenoble Alpes, CNRS, Inria, Grenoble INP LIG & Archimedes RU, NKUA panayotis.mertikopoulos@imag.fr Georgios Piliouras Engineering Systems and Design (SUTD) georgios@sutd.edu.sg
Pseudocode No The paper defines the preconditioned hidden gradient descent (PHGD) as a "stochastic first-order recursion θi,t+1 = θi,t γt Pi,t Vi,t (PHGD)" with numbered points explaining its components, but it does not present this in a structured pseudocode or algorithm block.
Open Source Code No The paper does not provide any explicit statements or links indicating that the source code for the methodology is openly available.
Open Datasets No The paper describes game setups like "Matching Pennies zero-sum game" and "El Farol Bar congestion game" for its experiments. It constructs specific scenarios based on these games (e.g., using MLPs) but does not refer to or provide access information for any external, publicly available datasets.
Dataset Splits No The paper describes various game setups for its experiments and evaluates the performance of its algorithm within these simulated environments. However, it does not provide specific training, validation, and test dataset splits (e.g., percentages, sample counts, or references to predefined splits) in the context of data partitioning for reproducibility.
Hardware Specification No The paper does not provide specific hardware details (e.g., exact GPU/CPU models, memory amounts, or detailed computer specifications) used for running its experiments.
Software Dependencies No The paper mentions mathematical constructs like "multi-layer perceptrons (MLPs)" and "sigmoid activation functions" but does not list specific software libraries (e.g., PyTorch, TensorFlow) or their version numbers, which are necessary for reproducible software dependencies.
Experiment Setup Yes The algorithm employs a constant step-size of 0.01 and is initialized at the arbitrary state (1.25, 2.25) in the control variables space.