Direct discovery · Materials science · Chemistry

Universal machine-learned potentials show systematic softening

Universal machine-learned potentials show systematic softening.

Summary

The npj Computational Materials paper evaluates universal machine-learning interatomic potentials including M3GNet, CHGNet and MACE-MP-0. It reports a systematic potential-energy-surface softening effect, with energy and force underprediction in surfaces, defects, ion migration, phonons and high-energy states.

AI role

AI contributes to generating, ranking, predicting, proving, designing, or validating a concrete research output in materials science, chemistry.

Narrative role

This belongs in the timeline because it records a concrete AI-assisted scientific output rather than only a workflow, product, or adoption signal.

Caveat

The event should be read as a bounded research result unless later validation, independent replication, adoption, or field-level impact is tracked separately.