Benefits of Low-Cost Bio-Inspiration in the Age of Overparametrization

2026-04-22

AI Summary

This work addresses the challenge of inefficient learning in robot control tasks constrained by limited input–output spaces and inherent performance ceilings, where over-parameterization may hinder optimization. By integrating evolutionary algorithms with Actor-Critic reinforcement learning, the study systematically evaluates the control performance of central pattern generators (CPGs) and multilayer perceptrons (MLPs) of varying scales on robots with restricted proprioception. Introducing a novel Parameter Impact metric, the authors demonstrate that increased parameter count does not necessarily correlate with improved task performance. Empirical results across multiple reward formulations consistently show that shallow MLPs and densely connected CPGs outperform deeper MLPs and standard Actor-Critic architectures, thereby validating the efficacy and superiority of low-parameter, biologically inspired controllers in such constrained settings.

Technology Category

• Intelligent Robots: Learning & Optimization for ROB
• Machine Learning: Evolutionary Learning
• Search and Optimization: Learning to Search

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Social networks and social learning
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

While Central Pattern Generators (CPGs) and Multi-Layer Perceptrons (MLP) are widely used paradigms in robot control, few systematic studies have been performed on the relative merits of large parameter spaces. In contexts where input and output spaces are small and performance is bounded, having more parameters to optimize may actively hinder the learning process instead of empowering it. To empirically measure this, we submit a given robot morphology, with limited proprioceptive capabilities, to controller optimization under two bio-inspired paradigms (CPGs and MLPs) with evolutionary- and reinforcement- trainer protocols. By varying parameter spaces across multiple reward functions, we observe that shallow MLPs and densely connected CPGs result in better performance when compared to deeper MLPs or Actor-Critic architectures. To account for the relationship between said performance and the number of parameters, we introduce a Parameter Impact metric which demonstrates that the additional parameters required by the reinforcement technique do not translate into better performance, thus favouring evolutionary strategies.

Problem

• overparametrization
• robot control
• parameter efficiency
• bio-inspired paradigms
• performance optimization

Innovation

• overparametrization
• Central Pattern Generators
• Parameter Impact
• evolutionary strategies
• robot control