Convergence of mean-field Langevin dynamics: time-space discretization, stochastic gradient, and variance reduction

Authors: Taiji Suzuki, Denny Wu, Atsushi Nitanda

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

Reproducibility Variable Result LLM Response
Research Type Theoretical We provide a general framework to prove a uniform-in-time propagation of chaos for MFLD that takes into account the errors due to finiteparticle approximation, time-discretization, and stochastic gradient. To demonstrate the wide applicability of our framework, we establish quantitative convergence rate guarantees to the regularized global optimal solution for (i) a wide range of learning problems such as mean-field neural network and MMD minimization, and (ii) different gradient estimators including SGD and SVRG.
Researcher Affiliation Academia Taiji Suzuki1,2, Denny Wu3,4, Atsushi Nitanda2,5 1University of Tokyo, 2RIKEN AIP, 3New York University, 4Flatiron Institute, 5Kyushu Institute of Technology
Pseudocode No The paper describes algorithms (F-MFLD, SGD-MFLD, SVRG-MFLD) in text but does not provide a formal pseudocode block or algorithm listing.
Open Source Code No The paper does not provide any statement or link regarding the availability of open-source code for the described methodology.
Open Datasets No The paper is theoretical and focuses on convergence analysis; it does not describe experiments performed on specific public datasets with access information. It mentions examples like "training data (zi, yi)" but this is a general reference within the problem setting, not an actual dataset used for empirical evaluation.
Dataset Splits No The paper is theoretical and focuses on convergence analysis; it does not present empirical experiments that would require training/validation/test splits.
Hardware Specification No The paper is theoretical and does not describe any experiments that would require specific hardware. Therefore, no hardware specifications are provided.
Software Dependencies No The paper is theoretical and does not describe any experiments that would require specific software dependencies with version numbers.
Experiment Setup No The paper is theoretical and focuses on convergence analysis; it does not describe specific experimental setups with hyperparameters or training settings.