Design of an Explainable Input Control for Cross-Company Federated Learning
ECIS 19.06.2024We propose and demonstate a novel IT-artifact used for explainable input control in cross-company federated learning. It relies on privacy-preserving measurements of client imbalance to decide about a client's admission to the platform.
Monitoring Concept Drift in Continuous Federated Learning Platforms
IDA 24.04.2024Concept drift has negative effect on continuous FL performance. In this work, we therefore compare different drift detection approaches for FL with dynamic client participation. In particular, we focus on the point in time when different drift detection approaches intervene.
Recent Presentations
Predictive Diagnostics for Federated Learning: A Privacy-Preserving Toolbox Towards Successful Federated Learning
ECML PKDD PhD Forum 18.09.2023In order to increase both fairness and success of federated learning prior to actual participation and computation, I aim towards providing privacy-preserving tools in my dissertion, tentatively titled "Predictive Diagnostics in Federated Learning".
All Publications
Teaching
Although I am currently not involved in any lectures or seminars, I actively participate in maintaining academic operations in the following capacities:
- Academic Advisor: As such, I am responsible for student consultation and academic advisory for the Master's degree programme in Data Science at Bielefeld University
- Committee Work: I am a representative of the academic middle class in committee work of the Faculty of Technology