Playground for Compositional Machine Design

Agentic Design Of Compositional Machines

Wenqian Zhang¹ Weiyang Liu^2,* Zhen Liu^1,*,†

¹The Chinese University of Hong Kong (Shenzhen)

²The Chinese University of Hong Kong

arXiv Code

^* Equal advising. ^† Corresponding Author.

Gallery of Tasks

Moving on rough terrains

Throwing stones far away

Throwing stones through a ring

Delivering boulders

Moving on a curved track

Picking objects from the bottom of a well

LLM-Agent-Generated Machines

Generated by agentic systems with different LLMs.

Results from RL-finetuned LLMs

Reinforcement learning with verifiable rewards (RLVR) improves Qwen-2.5-14B model in compositional machine design.

BibTeX

@article{zhang2025besiegefield,
  title={Agentic Design of Compositional Machines},
  author={Zhang, Wenqian and Liu, Weiyang and Liu, Zhen},
  journal={arXiv preprint arXiv:2510.14980},
  year={2025},
}