A Heterogeneous Mixture of Experts Framework for Interpretable Machine Learning

Soham Chatterjee, Rwitobroto Dey, Smarajit Bose

Mixture-of-Experts (MoE) models provide a flexible framework for partitioning complex prediction problems into simpler local learning tasks through an input-dependent gating mechan…