A Comparison of Architectural Constraints for Feedforward Neural Diversity Machines

This paper is a follow-up study on an earlier introductory paper on Neural Diversity Machines (NDM). NDMs are a subclass of Hybrid Artificial Neural Networks (HANNs), which are digital representations of biological neural networks present in the human brain. As opposed to traditional artificial neural networks (ANNs) which tend to be focused around uniform neurons, HANNs and NDMs tend to adopt heterogeneous types of neurons, partly with the aim of exploring the potential benefits of neural diversity in ANNs. This paper demonstrates and analyzes the performance of three architectural variants of a subclass of NDM (coined Mini-NDMs) in solving classification problems when tested on real life data sets.