Skip to content

Add new baseline: FedAMP #7375

Description

@kir-7

Type

Feature

Description

Paper

Personalized Cross-Silo Federated Learning on Non-IID Data

Link

https://doi.org/10.1609/aaai.v35i9.16960

Motivation about why the paper should be implemented as a baseline.

FedAMP is a similarity based personalized FL algorithm that aims to improve performance under heterogeneity. The paper shows improved performance when compared against standard algorithms like FedAVG or FedProx.

Plan is to implement FedAMP and evaluate its performance on CIFAR-10 dataset with pathological sampling and compare against FedAvg and FedProx. If the community is interested in this baseline, I would like to work on it and submit a PR.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Fields

    No fields configured for Feature.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions