XxaCT-NN: Structure Agnostic Multimodal Learning for Materials Science

TL;DR

This paper introduces a scalable, structure-agnostic multimodal learning framework for materials science that leverages elemental composition and X-ray diffraction data, bypassing the need for crystal structures.

Contribution

It presents a novel multimodal architecture with self-supervised pretraining strategies, enabling faster convergence and improved accuracy in materials property prediction without crystal structure input.

Findings

Pretraining accelerates convergence by up to 4.2x.

Multimodal models outperform unimodal baselines, especially with larger datasets.

The approach enables structure-free, experimentally grounded materials modeling.

Abstract

Recent advances in materials discovery have been driven by structure-based models, particularly those using crystal graphs. While effective for computational datasets, these models are impractical for real-world applications where atomic structures are often unknown or difficult to obtain. We propose a scalable multimodal framework that learns directly from elemental composition and X-ray diffraction (XRD) -- two of the more available modalities in experimental workflows without requiring crystal structure input. Our architecture integrates modality-specific encoders with a cross-attention fusion module and is trained on the 5-million-sample Alexandria dataset. We present masked XRD modeling (MXM), and apply MXM and contrastive alignment as self-supervised pretraining strategies. Pretraining yields faster convergence (up to 4.2x speedup) and improves both accuracy and representation quality. We further demonstrate that multimodal performance scales more favorably with dataset size than unimodal baselines, with gains compounding at larger data regimes. Our results establish a path toward structure-free, experimentally grounded foundation models for materials science.