LGAICEMay 9

Benchmarking Compositional Generalisation for Machine Learning Interatomic Potentials

arXiv:2605.0898867.6
Predicted impact top 28% in LG · last 90 daysOriginality Incremental advance
AI Analysis

It identifies a critical limitation in current MLIPs for computational chemistry and materials science, highlighting the need for models that truly learn compositional structure.

The paper proposes a benchmark to test compositional generalisation in machine learning interatomic potentials, finding that state-of-the-art models, including foundation models, show errors an order of magnitude higher on out-of-distribution molecules compared to in-distribution ones.

Machine Learning Interatomic Potentials play a fundamental role in computational chemistry and materials science, enabling applications from molecular dynamics simulations to drug design and materials discovery. While recent approaches can estimate inter-atomic forces with high precision, it remains unclear to what extent they can generalise to previously unseen molecules. Do they learn the compositional structure of chemistry, capturing how molecular fragments and their combinations determine properties, or do they primarily learn to interpolate patterns that are specific to the training examples? To address this question, we propose a benchmark consisting of four tasks that require some form of compositional generalisation. In each task, models are tested on molecules that were unseen during training, but the training data is chosen such that generalisation to the test examples should be feasible for models that learn the underlying physical principles. Our empirical analysis shows that the considered tasks are highly challenging for state-of-the-art models, with errors on out-of-distribution examples often an order of magnitude higher than on in-distribution examples, even when using foundation models that have been pre-trained on millions of molecules.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes