Biological connectomes as a representation for the architecture of artificial neural networks

TL;DR

This paper explores how biological connectomes, specifically the C. Elegans motor circuit, can inform artificial neural network design, showing that architectural statistics can serve as valuable priors even without high biophysical realism.

Contribution

It demonstrates that simplified connectome-inspired architectures can benefit machine learning, highlighting the importance of architectural statistics over biophysical detail.

Findings

Biophysical realism is not necessary for connectome benefits.

Architectural statistics serve as valuable priors.

Connectome structure aids locomotion tasks but may hinder others.

Abstract

Grand efforts in neuroscience are working toward mapping the connectomes of many new species, including the near completion of the Drosophila melanogaster. It is important to ask whether these models could benefit artificial intelligence. In this work we ask two fundamental questions: (1) where and when biological connectomes can provide use in machine learning, (2) which design principles are necessary for extracting a good representation of the connectome. Toward this end, we translate the motor circuit of the C. Elegans nematode into artificial neural networks at varying levels of biophysical realism and evaluate the outcome of training these networks on motor and non-motor behavioral tasks. We demonstrate that biophysical realism need not be upheld to attain the advantages of using biological circuits. We also establish that, even if the exact wiring diagram is not retained, the architectural statistics provide a valuable prior. Finally, we show that while the C. Elegans locomotion circuit provides a powerful inductive bias on locomotion problems, its structure may hinder performance on tasks unrelated to locomotion such as visual classification problems.