Core Concepts
MetaFedCBT improves CBT learning by addressing non-IID data with metadata-driven FL.
Abstract
The article introduces MetaFedCBT, a novel federated learning framework for connectional brain template (CBT) learning. It addresses the non-independent and identically distributed (non-IID) issue in multi-domain brain connectivity datasets. MetaFedCBT uses metadata to guide connectivity generation, improving CBT centeredness, discriminativeness, and topological soundness. The method outperforms existing models in terms of CBT quality and effectiveness.
Stats
"Extensive experiments on multi-view morphological brain networks of normal and patient subjects demonstrate that our MetaFedCBT is a superior federated CBT learning model."
"Our MetaFedCBT significantly advances the state-of-the-art performance in CBT learning."
"The results show that our MetaFedCBT outperforms state-of-the-art models remarkably."