We are super happy to share that both paper submissions of our research group have been accepted for presentation at ESANN 2025 conference:

1. Our first paper, "Resource-Aware Cooperation in Federated Learning", authored by Manuel Röder, Fabian Geiger and Frank-Michael Schleif, introduces a novel framework for federated learning that uses game theory to model client interactions. In distributed learning, clients often face resource constraints and non-cooperative behaviour. The proposed framework allows clients to adapt their strategies based on past interactions and available resources, striking a balance between individual utility and collective contribution. This approach promotes cooperation and efficiency in real-world decentralised AI systems.
   
   The research is kindly supported by THWS ProPere and the European Regional Development Fund (ERDF).
   
   
2. Our second paper, titled "Multiclass Adaptive Subspace Learning", authored by Peter Preinesberger, Maximilian Münch and Frank-Michael Schleif, is about adaptive Subspace Kernel Fusion, a proximity-based learning approach, able to integrate heterogeneous data sources. Unlike Deep Learning, it has the benefit of being applicable in data-limited scenarios. In this paper ASKF is extended by a proper multi-classification strategy, detaching training cost from class count in the process.

Congratulations to the authors for their great work!

Both papers will be presented at the 33rd European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2025), taking place from April 23 to April 25 in Bruges, Belgium.