NEWRON: A New Generalization of the Artificial Neuron to Enhance the Interpretability of Neural Networks

TL;DR

This paper introduces NEWRON, a generalized neuron model that enhances neural network interpretability without sacrificing expressiveness, and demonstrates its effectiveness through extensive experiments.

Contribution

NEWRON provides a novel neuron generalization that improves interpretability while maintaining or exceeding traditional neural network performance.

Findings

NEWRON-based networks are more interpretable without losing accuracy.

Models generated by NEWRON outperform traditional interpretable models.

NEWRON achieves comparable or better results than standard neural networks.

Abstract

In this work, we formulate NEWRON: a generalization of the McCulloch-Pitts neuron structure. This new framework aims to explore additional desirable properties of artificial neurons. We show that some specializations of NEWRON allow the network to be interpretable with no change in their expressiveness. By just inspecting the models produced by our NEWRON-based networks, we can understand the rules governing the task. Extensive experiments show that the quality of the generated models is better than traditional interpretable models and in line or better than standard neural networks.