aiqbench analysis
ED-CSP advances crystal structure prediction by treating sparse electron diffraction data as a generative inverse problem, utilizing flow-based models to reconstruct 3D lattices directly from detector spot patterns. This approach demonstrates superior performance over existing X-ray conditioned methods and proves the model's ability to generalize to unseen chemical compositions without relying on simple formula retrieval. The work establishes a new benchmark for materials discovery, bridging the gap between simulated datasets and experimental diffraction observations.
ED-CSP: Crystal Structure Prediction from Electron Diffraction
arXiv:2608.06448v3 Announce Type: replace-cross Abstract: Recovering a periodic 3D crystal structure from sparse, unindexed electron diffraction (ED) observations is a challenging generative inverse problem. Existing ED-based learning methods mainly predict crystallographic labels, reconstruct structures from indexed reflections, or retrieve candidates from finite structure libraries. Here, we introduce ED-CSP, a machine learning framework that predicts crystal structures from chemical composition, atom count, and multiple detector-plane ED spot sets. ED-CSP combines a relational set encoder, permutation-invariant multi-view aggregation, and a periodic flow generator to jointly predict lattice parameters and fractional atomic coordinates. To train the model, we construct ED-CS, a dataset of 4.85 million simulated multi-view ED crystal structures, deduplicated across seven materials repositories and filtered to exclude CHILI-100K overlaps. On 2,075 held-out CHILI-100K materials, ED-CSP trained only on CHILI achieves a structural match rate of 57.49% MR@5, outperforming PXRDGen (52.92%), a state-of-the-art crystal structure prediction model conditioned on powder X-ray diffraction. Scaling training data further improves performance: initializing from a one-million-structure precursor raises MR@5 to 66.27%. On 1,024 compositions absent from the training retrieval library, the model still achieves 53.52% MR@5, demonstrating true generative capability beyond exact-formula retrieval. Replacing target ED observations with diffraction from non-isomorphic structures of identical composition decreases MR@5 by 22.09 percentage points, confirming that predictions depend on the input diffraction patterns rather than composition alone. ED-CSP and ED-CS establish a benchmark for generative crystal structure prediction from sparse ED observations and provide a foundation for future transfer to experimental data.
Abstract:Recovering a periodic 3D crystal structure from sparse, unindexed electron diffraction (ED) observations is a challenging generative inverse problem. Existing ED-based learning methods mainly predict crystallographic labels, reconstruct structures from indexed reflections, or retrieve candidates from finite structure libraries. Here, we introduce ED-CSP, a machine learning framework that predicts crystal structures from chemical composition, atom count, and multiple detector-plane ED spot sets. ED-CSP combines a relational set encoder, permutation-invariant multi-view aggregation, and a periodic flow generator to jointly predict lattice parameters and fractional atomic coordinates. To train the model, we construct ED-CS, a dataset of 4.85 million simulated multi-view ED crystal structures, deduplicated across seven materials repositories and filtered to exclude CHILI-100K overlaps. On 2,075 held-out CHILI-100K materials, ED-CSP trained only on CHILI achieves a structural match rate of 57.49% MR@5, outperforming PXRDGen (52.92%), a state-of-the-art crystal structure prediction model conditioned on powder X-ray diffraction. Scaling training data further improves performance: initializing from a one-million-structure precursor raises MR@5 to 66.27%. On 1,024 compositions absent from the training retrieval library, the model still achieves 53.52% MR@5, demonstrating true generative capability beyond exact-formula retrieval. Replacing target ED observations with diffraction from non-isomorphic structures of identical composition decreases MR@5 by 22.09 percentage points, confirming that predictions depend on the input diffraction patterns rather than composition alone. ED-CSP and ED-CS establish a benchmark for generative crystal structure prediction from sparse ED observations and provide a foundation for future transfer to experimental data.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2608.06448 [cs.LG]
(or arXiv:2608.06448v3 [cs.LG] for this version)
From: Arnaud Demortiere Dr. [view email] [v1] Thu, 6 Aug 2026 16:55:20 UTC (2,282 KB) [v2] Wed, 12 Aug 2026 05:03:20 UTC (1,056 KB) [v3] Thu, 27 Aug 2026 21:47:38 UTC (2,282 KB)
Full text under the publisher's open licence · 285 words
Related coverage
Synthesis Superintelligence: from Semiconductors to Superconductors — Periodic Labs’ Liam Fedus and Ekin Dogus Cubuk
Latent Space · Oct 8, 2026
Third Form of Magnetism Earns Europe Top Physics Prize : Could Reshape AI Hardware
techtimes.com · Jul 26, 2026
X-ray scans reveal broad domain walls in cuprate superconductors
Quantum Zeitgeist · Sep 18, 2026
MatCreatioNN: Machine learning-guided computational discovery of photocatalysts for environmental applications
arXiv · Jul 31, 2026