mlsanityData to 7 Oct 2026

When Search Becomes Memory: Accelerating Robot Design Discovery with Self-Evolving Skills

arXiv 2605.25832 ↗First posted 25 May 2026cs.RO cs.AI cs.CLcited by 0

On the map: Synthetic data and self-improvement for LLMs → Self-evolving LLM agents

Abstract (arXiv)

Large language models (LLMs) are increasingly used as proposal generators for evolutionary robot design, yet most loops remain memoryless: simulator results shape the next population but are not preserved as reusable design knowledge. We present Auto-Robotist, a self-evolving LLM agent that distills morphology-search traces into an explicit natural-language skill library. Each skill stores a structural archetype, evidence-grounded positive and negative rules, and the evaluated designs that support them, making design memory inspectable rather than implicit in a population. During search, the agent retrieves skills to condition LLM edits of elite bodies while retaining a Genetic Algorithm (GA) mutation path for exploration; after evaluation, it updates the library through Add, Diagnose, and Merge. Across seven EvoGym tasks spanning locomotion, traversal, and object interaction, Auto-Robotist improves cold-start 5x5 search and transfers learned skills to 10x10 design spaces, where reference-conditioned transfer outperforms GA on every task. These results suggest that LLM agents can convert expensive physical evaluations into reusable, auditable design principles. Our code is publicly available at https://github.com/wangyf9/Auto-Robotist .

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Self-evolving LLM agents, one of 13 directions in Synthetic data and self-improvement for LLMs. Builds on 18 of its 27 references that are in the map. First posted on arXiv 25 May 2026.

Builds on

18 of 27 references.

References · 18

  1. 2017Proximal Policy Optimization Algorithmscited by 14,441
  2. 2025A-MEM: Agentic Memory for LLM Agentscited by 1,000
  3. 2023Large Language Models as Optimizerscited by 830
  4. 2015Illuminating search spaces by mapping elitescited by 406
  5. 2020Emergent Complexity and Zero-shot Transfer via Unsupervised Environment Designcited by 260
  6. 2020DreamCoder: Growing generalizable, interpretable knowledge with wake-sleep Bayesian program learningcited by 159
  7. 2021Embodied Intelligence via Learning and Evolutioncited by 154
  8. 2020Automatic Curriculum Learning For Deep RL: A Short Surveycited by 141
  9. 2025MemInsight: Autonomous Memory Augmentation for LLM Agentscited by 84
  10. 2024DrEureka: Language Model Guided Sim-To-Real Transfercited by 78
  11. 2018Reinforcement Learning for Improving Agent Designcited by 55
  12. 2018Modular meta-learningcited by 51
6 more
  1. 2022Evolution Gym: A Large-Scale Benchmark for Evolving Soft Robotscited by 44
  2. 2017Scalable Co-Optimization of Morphology and Control in Embodied Machinescited by 38
  3. 2019Data-efficient Co-Adaptation of Morphology and Behaviour with Deep Reinforcement Learningcited by 34
  4. 2021Transform2Act: Learning a Transform-and-Control Policy for Efficient Agent Designcited by 34
  5. 2024Text2Robot: Evolutionary Robot Design from Text Descriptionscited by 8
  6. 2025Natural Selection via Foundation Models for Soft Robot Evolutioncited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025MemEvolve: Meta-Evolution of Agent Memory Systems
  2. 2026AlgoEvo: Self-Evolving Agentic Search for Automated Algorithm Discovery
  3. 2026MemSkill: Learning and Evolving Memory Skills for Self-Evolving Agents
  4. 2026MUSE-Autoskill: Self-Evolving Agents via Skill Creation, Memory, Management, and Evaluation
  5. 2026EmbodiedRSI: Active Continual Robot Learning Through Hypothesis-Guided Co-Evolution
  6. 2026FreeEvolve: Learning to Evolve Beyond Fixed Loops
  7. 2026MetaSkill-Evolve: Recursive Self-Improvement of LLM Agents via Two-Timescale Meta-Skill Evolution
  8. 2026Explore, Execute, Evolve: A Skill Acquisition and Reuse Loop for Embodied Agents