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

arXiv: https://arxiv.org/abs/2604.20365 · HTML: https://mlsanity.com/paper/2604.20365/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 30 of its 61 references that are in the map.
- First posted on arXiv 22 Apr 2026.

## Builds on (most cited first, 30 of 30)

- [High-Dimensional Continuous Control Using Generalized Advantage Estimation](https://mlsanity.com/paper/1506.02438/) (2015)
- [DARTS: Differentiable Architecture Search](https://mlsanity.com/paper/1806.09055/) (2018)
- [Understanding deep learning requires rethinking generalization](https://mlsanity.com/paper/1611.03530/) (2016)
- [Neural Architecture Search with Reinforcement Learning](https://mlsanity.com/paper/1611.01578/) (2016)
- [Dota 2 with Large Scale Deep Reinforcement Learning](https://mlsanity.com/paper/1912.06680/) (2019)
- [Fantastic Generalization Measures and Where to Find Them](https://mlsanity.com/paper/1912.02178/) (2019)
- [Learning agile and dynamic motor skills for legged robots](https://mlsanity.com/paper/1901.08652/) (2019)
- [Sim-to-Real: Learning Agile Locomotion For Quadruped Robots](https://mlsanity.com/paper/1804.10332/) (2018)
- [SMASH: One-Shot Model Architecture Search through HyperNetworks](https://mlsanity.com/paper/1708.05344/) (2017)
- [Towards Understanding the Role of Over-Parametrization in Generalization of Neural Networks](https://mlsanity.com/paper/1805.12076/) (2018)
- [Towards Generalization and Simplicity in Continuous Control](https://mlsanity.com/paper/1703.02660/) (2017)
- [DeepGait: Planning and Control of Quadrupedal Gaits using Deep Reinforcement Learning](https://mlsanity.com/paper/1909.08399/) (2019)
- [CPG-RL: Learning Central Pattern Generators for Quadruped Locomotion](https://mlsanity.com/paper/2211.00458/) (2022)
- [Fantastic Generalization Measures are Nowhere to be Found](https://mlsanity.com/paper/2309.13658/) (2023)
- [Learning Locomotion Skills in Evolvable Robots](https://mlsanity.com/paper/2010.09531/) (2020)
- [Optimizers Qualitatively Alter Solutions And We Should Leverage This](https://mlsanity.com/paper/2507.12224/) (2025)
- [Visual CPG-RL: Learning Central Pattern Generators for Visually-Guided Quadruped Locomotion](https://mlsanity.com/paper/2212.14400/) (2022)
- [The Effects of Learning in Morphologically Evolving Robot Systems](https://mlsanity.com/paper/2111.09851/) (2021)
- [Structural bias in population-based algorithms](https://mlsanity.com/paper/1408.5350/) (2014)
- [Comparing lifetime learning methods for morphologically evolving robots](https://mlsanity.com/paper/2203.03967/) (2022)
- [Reinforcement Learning of CPG-regulated Locomotion Controller for a Soft Snake Robot](https://mlsanity.com/paper/2207.04899/) (2022)
- [Efficiently Learning Small Policies for Locomotion and Manipulation](https://mlsanity.com/paper/2210.00140/) (2022)
- [HyperPPO: A scalable method for finding small policies for robotic control](https://mlsanity.com/paper/2309.16663/) (2023)
- [An Open-Loop Baseline for Reinforcement Learning Locomotion Tasks](https://mlsanity.com/paper/2310.05808/) (2023)
- [CPG-ACTOR: Reinforcement Learning for Central Pattern Generators](https://mlsanity.com/paper/2102.12891/) (2021)
- [A comparison of controller architectures and learning mechanisms for arbitrary robot morphologies](https://mlsanity.com/paper/2309.13908/) (2023)
- [Solving Deep Reinforcement Learning Tasks with Evolution Strategies and Linear Policy Networks](https://mlsanity.com/paper/2402.06912/) (2024)
- [Neural Circuit Architectural Priors for Quadruped Locomotion](https://mlsanity.com/paper/2410.07174/) (2024)
- [Qualitative Differences Between Evolutionary Strategies and Reinforcement Learning Methods for Control of Autonomous Agents](https://mlsanity.com/paper/2205.07592/) (2022)
- [Emergence of specialized Collective Behaviors in Evolving Heterogeneous Swarms](https://mlsanity.com/paper/2402.04763/) (2024)

## Abstract (arXiv)

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 highly constrained settings. As opposed to traditional Machine Learning contexts, our input and output spaces are small and performance is bounded thus having more parameters 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 optimisation under two bio-inspired paradigms (CPGs and MLPs) with evolutionary- and reinforcement- trainer protocols. By varying parameter spaces across multiple reward functions, we demonstrate 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 showcases diminishing returns for MLPs but not for CPGs. Taken together these results demonstrate, on a fixed quadrupedal morphology, the benefits of integrating prior bias when considering locomotion tasks with simple hinge actuators.