Target informed client recruitment for efficient federated learning in healthcare
By expressing the representativeness of clients in function of the deviation in the local target distribution, the sample size and efficiency of the underlying hardware, we are able to define a recruitment approach that yields a subset of clients for the federation resulting in significantly reduced training time, without harming predictive performance, whilst improving the privacy preserving characteristics compared to the standard FL and central approaches.
Authors
- Bart De Moor (ORCID: https://orcid.org/0000-0002-1154-5028)
- Vincent Scheltjens (ORCID: https://orcid.org/0000-0002-6382-5750)
- Wouter Verbeke (ORCID: https://orcid.org/0000-0002-8438-0535)
- Lyse Naomi Wamba Momo (ORCID: https://orcid.org/0000-0002-9019-0236)
Institutions
- VIB-KU Leuven Center for Microbiology (BE)
- KU Leuven (BE)
Publication Details
- Journal
- BMC Medical Informatics and Decision Making
- Published
- 2024-12-18
- DOI
- https://doi.org/10.1186/s12911-024-02798-4
- Primary Topic
- Privacy-Preserving Technologies in Data
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- European Commission
- KU Leuven