Response Matching for generating materials and molecules
This work addresses the challenge of generating stable materials and molecules for researchers in computational chemistry and materials science, offering a unified framework that is not incremental but introduces a new approach.
The paper tackles the problem of generating new molecular and material structures by introducing Response Matching (RM), a novel generative method that leverages energy and stress responses to drive structures to equilibrium, and demonstrates its efficiency and generalization across small organic molecules, stable crystals, and one-shot learning on a diamond configuration.
Machine learning has recently emerged as a powerful tool for generating new molecular and material structures. The success of state-of-the-art models stems from their ability to incorporate physical symmetries, such as translation, rotation, and periodicity. Here, we present a novel generative method called Response Matching (RM), which leverages the fact that each stable material or molecule exists at the minimum of its potential energy surface. Consequently, any perturbation induces a response in energy and stress, driving the structure back to equilibrium. Matching to such response is closely related to score matching in diffusion models. By employing the combination of a machine learning interatomic potential and random structure search as the denoising model, RM exploits the locality of atomic interactions, and inherently respects permutation, translation, rotation, and periodic invariances. RM is the first model to handle both molecules and bulk materials under the same framework. We demonstrate the efficiency and generalization of RM across three systems: a small organic molecular dataset, stable crystals from the Materials Project, and one-shot learning on a single diamond configuration.