Reproduce “TopoMAS: Large Language Model Driven Topological Materials Multiagent System” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Baohua Zhang, Xin Li, Huangchao Xu, Zhong Jin, Quansheng Wu, Ce Li
TopoMAS introduces a human-AI multi-agent system that automates the entire topological materials discovery pipeline, successfully identifying a novel topological phase, SrSbO₃, which was confirmed by first-principles calculations. The framework demonstrates that smaller, resource-efficient Large Language Models can achieve competitive performance in complex scientific tasks compared to larger models.
Key insights
- Smaller Large Language Models (LLMs) can achieve comparable or superior performance to larger models in specialized scientific domains when integrated within an optimized multi-agent framework, leading to significant cost and resource efficiency.
- A comprehensive, end-to-end multi-agent framework can orchestrate an entire materials discovery pipeline, from natural language query to first-principles validation and autonomous knowledge integration.
- Autonomous integration of computational outcomes into a dynamic knowledge graph creates a self-enhancing knowledge base, reducing redundancy and accelerating future scientific inquiries.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.