Lukas Kirchdorfer

Enabling Mixed Effects Neural Networks for Diverse, Clustered Data Using Monte Carlo Methods

Neural networks often assume independence among input data samples, disregarding correlations arising from inherent clustering patterns in real-world datasets (e.g., due to different sites or repeated measurements). Recently, mixed effects neural networks (MENNs) which separate cluster-specific 'ran…