Scientific need
Advanced magnetic, superconducting, catalytic, and quantum materials often live in vast multicomponent composition spaces. Exhaustive DFT and experimental search are not practical at this scale.
DOE BES Computational Materials Science Center
Machine learning accelerated materials discovery at exascale
We build open-source, exascale-capable workflows that connect AI/ML, first-principles theory, atomistic simulation, materials databases, and experimental validation to accelerate the discovery and synthesis of functional materials.
Why this matters
From materials discovery to synthesis guidance
Center strategy
MLAMD combines machine-learning screening, ML interatomic potentials, adaptive search, DFT validation, phase-stability analysis, molecular dynamics, and CALPHAD-style thermodynamic modeling.
DOE value
The center turns leadership-class computing into reusable community capability: open-source workflows, benchmarked scalability, versioned structure and thermodynamic data, and transparent validation gates.
Explore