Using large language models to probe the limits of atom-centered structural descriptors
Abstract
Mapping an atomic structure to a compact set of geometric descriptors is an essential step in any machine-learning application to atomic-scale modeling. A powerful and widely-used approach can be understood as a discretization of the histogram of pair distances, triangles, etc., that results in a hierarchy of symmetry-invariant atom-centered descriptors. Unfortunately, the lower rungs on this hierarchy (two, three, four-neighbor clusters) were found to be incomplete, with symmetry-unrelated pairs of structures having exactly the same descriptors. However, all the ``degeneracies'' reported so far are resolved by considering larger clusters of neighbors to build the descriptors. We report examples of 3D structures that are indistinguishable even if one considers clusters of up to seven neighbors, and to arbitrary order when considering a practical level of discretization of the descriptors, discovered with the assistance of large language models. The key ingredients in their construction can be traced to results that have been known for decades in different communities: the model was able to find the references and recognize their significance for the problem at hand. We believe this experiment exposes an extremely fruitful usage pattern for AI in science: translating results between different communities and application domains, accelerating the process by which serendipitous discoveries in a field become breakthroughs in another.
Related
- Limits of geometric packing descriptors in molecular crystal energy ranking
- What Limits Lithium-Ion Battery State of Charge Estimation with a Hybrid Large Language Model? Layer Type Aware Adaptation and the Initial Condition Bottleneck
- How Molecular Weight Shapes the Relative Importance of Chemical and Structural Descriptors for Aqueous Solubility
- Cross-Dataset Assessment of Cathode Thermal Descriptors for Whole-Cell Self-Heating Kinetics in Lithium-Ion Batteries: Onset, Reaction Magnitude, and Limits of Transferability
- Machine Learning Prediction of Transport Properties in Amorphous Polymer Electrolytes Using Chemically Informed Structural Descriptors
BibTeX
@misc{domina2026using,
title = {{Using large language models to probe the limits of atom-centered structural descriptors}},
author = {Michelangelo Domina and Michele Ceriotti},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2607.26984},
eprint = {2607.26984},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2607.26984},
}
RIS
TY - UNPB
TI - Using large language models to probe the limits of atom-centered structural descriptors
AU - Michelangelo Domina
AU - Michele Ceriotti
PY - 2026
DA - 2026/10/02
PB - arXiv
DO - 10.48550/arxiv.2607.26984
UR - https://arxiv.org/abs/2607.26984
AB - Mapping an atomic structure to a compact set of geometric descriptors is an essential step in any machine-learning application to atomic-scale modeling. A powerful and widely-used approach can be understood as a discretization of the histogram of pair distances, triangles, etc., that results in a hierarchy of symmetry-invariant atom-centered descriptors. Unfortunately, the lower rungs on this hierarchy (two, three, four-neighbor clusters) were found to be incomplete, with symmetry-unrelated pairs of structures having exactly the same descriptors. However, all the ``degeneracies'' reported so far are resolved by considering larger clusters of neighbors to build the descriptors. We report examples of 3D structures that are indistinguishable even if one considers clusters of up to seven neighbors, and to arbitrary order when considering a practical level of discretization of the descriptors, discovered with the assistance of large language models. The key ingredients in their construction can be traced to results that have been known for decades in different communities: the model was able to find the references and recognize their significance for the problem at hand. We believe this experiment exposes an extremely fruitful usage pattern for AI in science: translating results between different communities and application domains, accelerating the process by which serendipitous discoveries in a field become breakthroughs in another.
ER -