From Evaluation to Design: Using Potential Energy Surface Smoothness Metrics to Guide Machine Learning Interatomic Potential Architectures

Ryan Liu 1,2, Eric Qu 2, Tobias Kreiman 2, Samuel M. Blau 3, Aditi S. Krishnapriyan 2,3 · 2026

Read From Evaluation to Design: Using Potential Energy Surface Smoothness Metrics to Guide Machine Learning Interatomic Potential Architectures by Ryan Liu 1,2, Eric Qu 2, Tobias Kreiman 2, Samuel M. Blau 3, Aditi S. Krishnapriyan 2,3 (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-04861.txt.

Description

Ryan Liu and 4 co-authors argue that machine-learned atomistic models should be judged by physical smoothness, not just accuracy on energies and forces, in a 2026 paper. A model's energy surface can be jagged in ways standard tests miss, causing later simulations to misbehave.

Their Bond Smoothness Characterization Test probes the surface through controlled bond stretches. It detects breaks, fake minima, and spurious forces near and far from balance, at a fraction of the cost of full dynamics runs. The test matched simulation stability closely. Using it as a design guide, the team refined a transformer model with a new neighbor method and temperature-controlled attention. The result combined low error, stable runs, and robust trait predictions.

Catalog record

CREATORRyan Liu 1,2, Eric Qu 2, Tobias Kreiman 2, Samuel M. Blau 3, Aditi S. Krishnapriyan 2,3
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE52 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-04861

Browse the full library · About the Tower of Babel