Linear Lower Bounds and Conditioning of Differentiable Games

Authors: Adam Ibrahim, Waı̈ss Azizian, Gauthier Gidel, Ioannis Mitliagkas

ICML 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical In this work, we approach the question of fundamental iteration complexity by providing lower bounds to complement the linear (i.e. geometric) upper bounds observed in the literature on a wide class of problems. We cast saddle-point and min-max problems as 2-player games. We leverage tools from single-objective convex optimisation to propose new linear lower bounds for convex-concave games. Notably, we give a linear lower bound for n-player differentiable games, by using the spectral properties of the update operator. We then propose a new definition of the condition number arising from our lower bound analysis.
Researcher Affiliation Academia 1Mila, University of Montreal 2Ecole Normale Supérieure, Paris.
Pseudocode No The paper does not contain structured pseudocode or algorithm blocks.
Open Source Code No The paper does not provide concrete access to source code for the methodology described.
Open Datasets No The paper is theoretical, focusing on mathematical lower bounds and condition numbers, and does not use or reference any publicly available datasets for training or empirical evaluation.
Dataset Splits No The paper is theoretical and does not describe experimental validation or dataset splits.
Hardware Specification No The paper is theoretical and does not mention any hardware specifications used for experiments.
Software Dependencies No The paper is theoretical and does not mention any specific software dependencies with version numbers.
Experiment Setup No The paper is theoretical and does not describe specific experimental setup details, hyperparameters, or training configurations.