Can a Compact Neuronal Circuit Policy be Re-purposed to Learn Simple Robotic Control?

TL;DR

This paper introduces Neuronal Circuit Policies (NCPs), a biologically inspired neural system re-purposed for robotic control, demonstrating competitive performance with fewer parameters and interpretable dynamics.

Contribution

It presents a novel approach of re-purposing biological neural circuits, specifically C. elegans, for robotic control tasks using search-based optimization.

Findings

NCPs perform comparably or better than deep learning models.

NCPs use significantly fewer learnable parameters.

NCPs exhibit interpretable cell-level dynamics.

Abstract

We propose a neural information processing system which is obtained by re-purposing the function of a biological neural circuit model, to govern simulated and real-world control tasks. Inspired by the structure of the nervous system of the soil-worm, C. elegans, we introduce Neuronal Circuit Policies (NCPs), defined as the model of biological neural circuits reparameterized for the control of an alternative task. We learn instances of NCPs to control a series of robotic tasks, including the autonomous parking of a real-world rover robot. For reconfiguration of the purpose of the neural circuit, we adopt a search-based optimization algorithm. Neuronal circuit policies perform on par and in some cases surpass the performance of contemporary deep learning models with the advantage leveraging significantly fewer learnable parameters and realizing interpretable dynamics at the cell-level.