A robust sampling technique for realistic distribution simulation in federated learning
Realistically client-biased label distribution can notably harm the training in a federated context. Our sampling algorithm for simulating realistic data distributions opens up an efficient way for prior analysis of this effect. The technique is agnostic to the chosen network architecture and target scenario and can be adapted to any feature or label problem with non-IID subpopulations.
Authors
- Andreas Wimmer (ORCID: https://orcid.org/0000-0003-4991-4391)
- Leonhard Rist (ORCID: https://orcid.org/0000-0002-9589-3774)
- Alexander Katzmann (ORCID: https://orcid.org/0000-0003-1958-1549)
- Michael Sühling (ORCID: https://orcid.org/0000-0002-1060-3759)
- Ashok Ravichandran
- Andreas Maier (ORCID: https://orcid.org/0000-0002-9550-5284)
- Robin Hoepp
Institutions
- Friedrich-Alexander-Universität Erlangen-Nürnberg (DE)
- Siemens Healthcare (United States) (US)
- Siemens Healthineers (Germany) (DE)
Publication Details
- Journal
- International Journal of Computer Assisted Radiology and Surgery
- Published
- 2025-09-02
- DOI
- https://doi.org/10.1007/s11548-025-03504-z
- Primary Topic
- Privacy-Preserving Technologies in Data
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Friedrich-Alexander-Universität Erlangen-Nürnberg