Building AI systems for autonomous scientific discovery
From machine learning models to closed-loop systems that decide what to simulate, design, or measure next.
AI for Materials Lab @ UAM · Led by Jorge Bravo Abad, Professor of Physics at Universidad Autónoma de Madrid.
Welcome to the AI for Materials Lab at Universidad Autónoma de Madrid.
Scientific AI is moving beyond prediction. The next generation of systems will actively participate in the discovery process: choosing what to simulate or measure next, learning from each result, and iterating toward scientific objectives under limited time and resources.
At the AI for Materials Lab at Universidad Autónoma de Madrid, we develop the machine learning methods and computational infrastructure needed to build these autonomous discovery loops.
Our research connects AI agents, Bayesian optimization, active learning and reinforcement learning with scientific simulations and experiments. We apply these approaches to problems in materials science, condensed matter physics and chemistry, from the exploration of quantum systems to the discovery and optimization of functional materials.
Our goal is to move from AI models that predict to AI systems that discover.
Core activities
-
We develop closed-loop AI systems that choose what to simulate, design, or measure next, learn from each result, and navigate scientific search spaces efficiently.
-
We apply machine learning to the discovery and design of materials and to the study of complex quantum systems, combining data-driven methods with physical knowledge.
-
We teach, communicate, and build communities around AI-enabled science through courses, books, talks, workshops, and collaborative initiatives.
