An autonomous AI agent is now deriving predictive physical laws and writing its own code to discover new materials.
April 25, 2026
Original Paper
From Data to Theory: Autonomous Large Language Model Agents for Materials Science
arXiv · 2604.19789
The Takeaway
Autonomous agents are moving beyond summarizing existing science to performing original research. This system independently chooses mathematical forms and runs simulations to find new physical relationships. It can derive predictive laws from raw data without any human guidance or intervention. This marks the beginning of AI-driven discovery in fields like materials science and chemistry. Scientists can now delegate the entire process of hypothesis testing and derivation to an autonomous machine.
From the abstract
We present an autonomous large language model (LLM) agent for end-to-end, data-driven materials theory development. The model can choose an equation form, generate and run its own code, and test how well the theory matches the data without human intervention. The framework combines step-by-step reasoning with expert-supplied tools, allowing the agent to adjust its approach as needed while keeping a clear record of its decisions. For well-established materials relationships such as the Hall-Petch