Self supervised learning improves crystal pattern analysis accuracy
Exploring Joint-Embedding Predictive Architectures for Self-Supervised Representation Learning of Powder X-ray Diffraction Patterns
Artificial IntelligenceMachine Learning
Summary
Analyzing patterns from X-ray diffraction helps scientists understand crystal structures, but current methods struggle with complex mixtures and variations. The authors use a new self-supervised learning method called JEPA that teaches a computer to find meaningful features in these patterns without labeled data. This method can accurately classify crystal types and estimate their characteristics nearly as well as much larger, fully supervised models. It also naturally organizes how physical changes appear in data and can even generate realistic mixtures without being explicitly trained on them.
What this means in practice
- •For materials science laboratories: Automatically analyze XRD patterns to classify crystals and estimate lattice parameters without extensive labeled data.
- •For industrial quality control teams: Detect and characterize multi-phase crystalline mixtures in materials production using learned representations that model complex structure variations.
Authors
Ming-Chiang Chang, Brian DeCost
Abstract
Powder X-ray diffraction (XRD) is the primary high-throughput method for resolving crystal structures, a cornerstone of materials science. As high-throughput synthesis and self-driving laboratories advance, automated XRD pattern analysis has become vital for accelerating materials discovery. However, traditional matrix factorization methods fail when lattice variations cause non-linear peak shifts. At the same time, supervised learning approaches face a combinatorial explosion when handling multi-phase mixtures. To overcome these challenges, we adopt a self-supervised Joint Embedding Predictive Architecture (JEPA) to learn representation embeddings from 1D diffraction patterns. By optimizing a masked self-prediction loss directly within the latent space, the framework captures high-level crystallographic semantics. Downstream evaluations show that lightweight attentive probes trained on frozen JEPA representations achieve space group classification and lattice parameter estimation accuracy close to much larger supervised vision transformers. Furthermore, latent probing reveals that the learned representation space naturally organizes physical transformations, isolating smooth trajectories for peak broadening. Interestingly, decoding along latent paths between single-phase crystal representations yields realistic multi-phase mixture patterns, demonstrating an emergent continuous structure learned without explicit multi-phase supervision.