Learning Nash Equilibria in Rank-1 Games

Authors: Nikolas Patris, Ioannis Panageas

ICLR 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental C EXPERIMENTS In this section, we provide experimental evaluation supporting our theoretical finding. Here we restrict to a single example, while we also provide an anonymous repository containing multiple rank-1 games of multiple sizes
Researcher Affiliation Academia Nikolas Patris University of California, Irvine npatris@uci.edu Ioannis Panageas University of California, Irvine ipanagea@ics.uci.edu
Pseudocode Yes Algorithm 1: Rank-1 Game Solver [...] Algorithm 2: OMWU(x, y, λ)
Open Source Code Yes C EXPERIMENTS In this section, we provide experimental evaluation supporting our theoretical finding. Here we restrict to a single example, while we also provide an anonymous repository containing multiple rank-1 games of multiple sizes
Open Datasets No We constructed a random 6 6 game defined by the matrix A, and the vectors a, b. The paper constructs a specific example for experiments but does not provide access details for a publicly available or open dataset.
Dataset Splits No The paper constructs a specific 6x6 game for its experiments but does not provide any information about training, validation, or test splits.
Hardware Specification No The paper does not provide any specific hardware details (e.g., GPU/CPU models, memory) used for running its experiments.
Software Dependencies No The paper mentions using 'optimistic gradient ascent (OGA)' but does not provide specific software names with version numbers.
Experiment Setup No The paper provides theoretical bounds for learning rates and iterations but does not specify concrete hyperparameter values or detailed training configurations (e.g., specific learning rates used in experiments, batch sizes, number of epochs) for the experimental setup.