Integrating Random Effects in Deep Neural Networks

Giora Simchoni, Saharon Rosset

Modern approaches to supervised learning like deep neural networks (DNNs) typically implicitly assume that observed responses are statistically independent. In contrast, correlated…