An open-access database is designed to facilitate machine learning (ML) assisted discovery of new magnets. A feedback loop between computational and experimental efforts leads to the discovery of several new rare-earth-free magnets.

In addition to structures from Material Project (MP), database includes many low-energy structures from adaptive genetic algorithm (AGA) search (see (a)).

Structures from AGA exhibit favorable magnetic properties (see (b & (c)).

Promising compositions identified from database provides useful guidance to experimental discovery of novel magnetic materials (see (c ) and (d)).