Multimodal Language and Graph Learning of Adsorption Configuration in Catalysis

TL;DR

This paper introduces a multi-modal learning approach combining graph neural networks and language models to improve adsorption energy prediction in catalysis, reducing error and enhancing transferability without relying solely on atomic coordinates.

Contribution

It presents a novel graph-assisted pretraining method that integrates GNNs with language models, improving energy prediction accuracy and transferability in catalyst screening.

Findings

Reduces MAE of energy prediction by about 10%

Enhances fine-tuning transferability across datasets

Uses language models to generate inputs from chemical composition and surface orientation

Abstract

Adsorption energy is a reactivity descriptor that must be accurately predicted for effective machine learning (ML) application in catalyst screening. This process involves determining the lowest energy across various adsorption configurations on a catalytic surface, which can exhibit very similar energy values. While graph neural networks (GNNs) have shown great success in computing the energy of catalyst systems, they rely heavily on atomic spatial coordinates. In contrast, transformer-based language models can directly use human-readable text inputs, potentially bypassing the need for detailed atomic positions. However, these language models often struggle with accurately predicting the energy of adsorption configurations. Our study addresses this limitation by introducing a self-supervised multi-modal learning approach called graph-assisted pretraining, which connects well-established GNNs with emerging language model applications. This method reduces the MAE of energy prediction for adsorption configurations by about 10%. Furthermore, our findings demonstrate that graph-assisted pretraining enhances fine-tuning with different datasets, indicating the transferability of this approach. This method also redirects the model's attention toward adsorption configuration, rather than individual adsorbate and catalyst information, similar to common domain knowledge. Building on this, we propose using generative large language models to create text inputs for the predictive model, based solely on chemical composition and surface orientation, without relying on exact atomic positions. This demonstrates a potential use case of language models in energy prediction without geometric information.