# Towards Multi-Morphology Controllers with Diversity and Knowledge Distillation

arXiv: https://arxiv.org/abs/2404.14625 · HTML: https://mlsanity.com/paper/2404.14625/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Multimodal distillation, one of 12 directions in Knowledge distillation.
- Builds on 23 of its 52 references that are in the map.
- First posted on arXiv 22 Apr 2024.

## Placed in

- Knowledge distillation / [Multimodal distillation](https://mlsanity.com/distill/multimodal-distillation/)

## Builds on (most cited first, 23 of 23)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks](https://mlsanity.com/paper/1703.03400/) (2017)
- [Addressing Function Approximation Error in Actor-Critic Methods](https://mlsanity.com/paper/1802.09477/) (2018)
- [Policy Distillation](https://mlsanity.com/paper/1511.06295/) (2015)
- [Meta-Transfer Learning for Few-Shot Learning](https://mlsanity.com/paper/1812.02391/) (2018)
- [Distral: Robust Multitask Reinforcement Learning](https://mlsanity.com/paper/1707.04175/) (2017)
- [Robots that can adapt like animals](https://mlsanity.com/paper/1407.3501/) (2014)
- [Actor-Mimic: Deep Multitask and Transfer Reinforcement Learning](https://mlsanity.com/paper/1511.06342/) (2015)
- [One Policy to Control Them All: Shared Modular Policies for Agent-Agnostic Control](https://mlsanity.com/paper/2007.04976/) (2020)
- [MetaMorph: Learning Universal Controllers with Transformers](https://mlsanity.com/paper/2203.11931/) (2022)
- [Learning to Control Self-Assembling Morphologies: A Study of Generalization via Modularity](https://mlsanity.com/paper/1902.05546/) (2019)
- [My Body is a Cage: the Role of Morphology in Graph-Based Incompatible Control](https://mlsanity.com/paper/2010.01856/) (2020)
- [Evolution Gym: A Large-Scale Benchmark for Evolving Soft Robots](https://mlsanity.com/paper/2201.09863/) (2022)
- [Scalable Co-Optimization of Morphology and Control in Embodied Machines](https://mlsanity.com/paper/1706.06133/) (2017)
- [MAP-Elites with Descriptor-Conditioned Gradients and Archive Distillation into a Single Policy](https://mlsanity.com/paper/2303.03832/) (2023)
- [How morphological development can guide evolution](https://mlsanity.com/paper/1711.07387/) (2017)
- [MAP-Elites enables Powerful Stepping Stones and Diversity for Modular Robotics](https://mlsanity.com/paper/2012.04375/) (2020)
- [The Quality-Diversity Transformer: Generating Behavior-Conditioned Trajectories with Decision Transformers](https://mlsanity.com/paper/2303.16207/) (2023)
- [Investigating Premature Convergence in Co-optimization of Morphology and Control in Evolved Virtual Soft Robots](https://mlsanity.com/paper/2402.09231/) (2024)
- [Evolving Modular Soft Robots without Explicit Inter-Module Communication using Local Self-Attention](https://mlsanity.com/paper/2204.06481/) (2022)
- [Modular Controllers Facilitate the Co-Optimization of Morphology and Control in Soft Robots](https://mlsanity.com/paper/2306.09358/) (2023)
- [Co-evolving morphology and control of soft robots using a single genome](https://mlsanity.com/paper/2212.11517/) (2022)

## Built on by (most cited first, 2 of 2)

- [Embodied Co-Design for Rapidly Evolving Agents: Taxonomy, Frontiers, and Challenges](https://mlsanity.com/paper/2512.04770/) (2025)
- [Social Learning Strategies for Evolved Virtual Soft Robots](https://mlsanity.com/paper/2604.12482/) (2026)

## Abstract (arXiv)

Finding controllers that perform well across multiple morphologies is an important milestone for large-scale robotics, in line with recent advances via foundation models in other areas of machine learning. However, the challenges of learning a single controller to control multiple morphologies make the `one robot one task' paradigm dominant in the field. To alleviate these challenges, we present a pipeline that: (1) leverages Quality Diversity algorithms like MAP-Elites to create a dataset of many single-task/single-morphology teacher controllers, then (2) distills those diverse controllers into a single multi-morphology controller that performs well across many different body plans by mimicking the sensory-action patterns of the teacher controllers via supervised learning. The distilled controller scales well with the number of teachers/morphologies and shows emergent properties. It generalizes to unseen morphologies in a zero-shot manner, providing robustness to morphological perturbations and instant damage recovery. Lastly, the distilled controller is also independent of the teacher controllers -- we can distill the teacher's knowledge into any controller model, making our approach synergistic with architectural improvements and existing training algorithms for teacher controllers.