Grant recipient
The project aims to make advanced computer simulations of new materials much more accessible while promoting efficient and environmentally friendly use of computing resources. To achieve this, a new framework will be developed that uses large language models (LLMs), like ChatGPT, to automatically set up and manage complex materials simulations. Today, running these simulations is complicated, expensive, and usually limited to expert researchers. Integrating LLMs as both user assistants and active decision-makers will enable more scientists, students, and engineers to use such tools without needing many years of specialized training. By integrating database knowledge of thousands of materials, the project will also explore how well the framework can directly predict material properties without needing to run calculations. The goal is to advance scientific discovery in renewable energy, electronics, and health technologies, helping to drive sustainable solutions in science and industry.
Line Jelver
ADaM: Autonomous workflows for Data-driven first-principles Modeling
Grant amount: DKK 13.297.000